REPORT COVER PAGE - DFKIaabecker/Postscript/D2_1.doc · Web viewthe project directors and project...

‘WP2’ ‘System Requirements’

Know-Net

Project number: EP28928

Project Title: KnowNet

Title of deliverable: ‘System Requirements Specification ‘

Availability: ‘Confidential’

Number: ‘D2.1’

Version: ‘Preliminary’

Contractual date of delivery: ‘31/3/99’

Actual date of delivery: ‘n.a.’

Workpackage: ‘WP 02’

Nature of document: ‘Report’

‘KNOWNET/WP2/PLN/D2/F’ 1


Document Change Control Table

Previous version Chapter / Section ChangesDraft; version 1 ‘5.3 The Know-Net

Performance Evaluation Method’

‘CHANGES PENDING’

Draft; version 1…



System Requirements Specification

Table of Contents

1. OVERVIEW......................................................................................6

2. REQUIREMENTS ENGINEERING FOR KNOW-NET.....................82.1. Particularities of the Know-Net RE Process...........................................82.2. Linking Requirements to the Know-Net Framework............................10

3. REQUIRED SERVICES AND FUNCTIONALITIES.......................16

4. PROCESS-SPECIFIC REQUIREMENTS......................................28

5. PRODUCT-SPECIFIC REQUIREMENTS......................................34

6. REQUIREMENTS FOR THE INDIVIDUAL KNOWLEDGE PORTFOLIO...........................................................................................36

6.1. General considerations........................................................................366.2. Knowledge relevant features of our activity.........................................37

7. SUGGESTED INTEGRATION APPROACH..................................40

8. SUMMARY.....................................................................................44

APPENDIX A: REQUIREMENTS ACQUISITION QUESTIONNAIRE....45A.1. Task / Activity Level.............................................................................45A.2. Document Level...................................................................................46A.3. Knowledge Management Level...........................................................47A.4. Technical Requirements......................................................................50

APPENDIX B: PLANET’S REQUIREMENTS.........................................51“Unbiased” Use Case....................................................................................51Structured Account of Requirements.............................................................57

Task / Activity Level.............................................................................................57Document Level...................................................................................................62Knowledge Management Level............................................................................63Technical Requirements.......................................................................................67

APPENDIX C: GOOCH WEBSTER’S REQUIREMENTS.......................75“Unbiased” Use Case....................................................................................75Structured Account of Requirements.............................................................78



Task/Activity Level...............................................................................................78Document Level...................................................................................................81Knowledge Management Level............................................................................82Technical Requirements.......................................................................................87

APPENDIX D: UBS’S REQUIREMENTS................................................92KM goals of Credit Risk Valuation.................................................................92Some major characteristics of the daily work of CRV....................................92Requirements................................................................................................93Qualification of the requirements.................................................................100Conclusions.................................................................................................103

APPENDIX E: FHBB’S CONTRIBUTIONS..........................................106

REFERENCES......................................................................................111



1 Overview

According to the Know-Net project plan the following objectives were addressed within the work of workpackage 2:

to capture the KNOWNET user requirements for the technological platform;

to explicitly identify possible coverage of some user requirements with existing systems and tools of provider companies;

to examine the feasibility of translating user requirements into general user requirements for knowledge management technological platforms with the additional input from members of the Observers Companies.

Since the main technological innovation to be achieved in the Know-Net project is to provide process-centred as well as product-centred technological support within one coherent software tool, the following two main parts of requirements analysis had to be done:

1. capturing the requirements for the process-centred elements of the technological infrastructure. The process-centred elements involve all aspects of groupware support that aim to improve the communication, co-operation and co-ordination processes.

2. capturing the requirements for the product-centred perspective of the technological infrastructure that focuses on knowledge objects, their creation, storage, and reuse in computer-based organisational memories.

Moreover, workpackage 2 attempts to generate a set of general user requirements for a technological platform to assist knowledge management in industry. To this end, extensive discussions and interviews between Know-Net user partners and developer partners were made. Additionally, commercial market surveys and tool comparisons were reviewed to see what is available on the market and which criteria are usually employed to compare and evaluate KM tools. The detailed user requirements can be found in Appendix B, C, and D. A structured, condensed summary of the services and functionalities wished by the users is given in Section 3.

For finding out the requirements specific to the process and the product view, respectively, each developer-partner focussed on the user requirements that relate mostly to their existing products, i.e.:

KNOWLEDGE ASSOCIATES captured the requirements for the Knowledger™ suite of products; the outcome is documented in section 4.

DFKI captured the requirements for intelligent assistance dealing with document content; results are documented in section 5.


Objectives of work package 2 ...

... address the process-centred and the product-centred view on Knowledge Management

A general set of technological requirements is derived ...

... and the possible contributions of the three developer partners are identified.


FHBB captured the requirements of the individual knowledge portfolio tool, which involves the definition of a restricted subtask domain within the main task of assisting individuals in knowledge portfolio management; the results are described in section 6.

Because the several functionalities required by the user partners are already described in much detail in the general set of requirements and in the appendices, those three sections merely summarize, from the developer’s point of view, what can be done and what should be done within the Know-Net project with respect to those identified requirements.

Based upon these concrete ideas of what should be realized as the functionality of the Know-Net tool, the workpackage examined specific issues for the integration between the Knowledger™ suite of products of KNOWLEDGE ASSOCIATES and the intelligent agents approach of DFKI. As a first result, a refined integration architecture is sketched in Section 7.


An integration architecture is sketeched.


2 Requirements Engineering for Know-Net

3 Particularities of the Know-Net RE Process

Before presenting the results of workpackage 2 in detail, we should mention that one cannot expect a quite conventional software requirements specification here. Require-ments engineering had to be unusual for several reasons:

Users were not typical end users here with a strong focus on their application task and few or no IT knowledge at all. Rather, some of them have a strong IT focus, e.g., UBS where a strong research and development team has many experiences both in IT support for KM and in the product-centred view on KM. So, some user requirements could already be biased towards possible IT realizations the authors had in mind when describing the requirements.

The application task under consideration, Knowledge Management, is also a quite typical topic for the developers such that they might want to project their own ideas and wishes what should be done for KM support onto the user requirements. On the other hand, it is even necessary to do that because some users might have no clear idea what is possible with new IT solutions for KM support.

In a research project on a quite new topic like KM, even the application task to be supported is still evolving, such that it cannot be expected to have an ultimate result, but we will rather make first steps in an iterative improvement process.

Since the Know-Net project is launched in an application-oriented research programme with quite limited resources for new development and functionalities, the focus is to get a new quality of services by integrating existing, complementary system modules, and not by inventing fundamentally new things. For requirements engineering this means, for instance, that we can let out some parts typically contained in requirements documents (e.g., many technical details about platforms, interfaces, etc.) because we have to settle upon our existing tools, anyway.

On the RE process these characteristics imposed conflicting goals:

1. Users should be as unbiased as possible to describe really user-driven requirements, and not developer-driven ones.

2. But, users must be as „informed“ by the developers as possible in order to see all possibilities and to know all restrictions.

In order to meet these goals requirements engineering in Know-Net roughly followed the strategy sketched in Figure 1:

First the users were asked to describe an ideal scenario for an optimal IT support for their KM activities. This should be an unbiased basis for further thinking. The results can be used as a baseline for evaluating project outcomes in the later evaluation phases.



Then the users were asked to regard the requirements questionnaire attached in Appendix A and to fill in as concrete as possible, detailed specifications of what they need in the several KM processes. This creation of a structured account of requirements was supported by discussions with developers and other users via Knowledger and by awareness creation through the developers showing what technolgy can offer. The detailed results are attached as appendices. It turned out that it makes sense to organize such discussions around business processes, or company knowledge management processes, and not directly around technical or architectural terms in order to be understood by the users.

In order to come to a technical list of requirements, developers took the first filled-in questionnaires, thought about what they could do for solving the described problems and came up with concrete service and functionality offers in the Know-Net Brussels meeting. Here it was decided to compile the users’ answers to the questionnaire into a table of services and functionalities ideally provided by a comprehensive KM tool. This compilation table now was organized around classes of services already quite close to layers of a tool architecture.

Figure 1: Requirements Engineering Steps for Know-Net


Users Developers

unbiased use case

structured account of requirements – organized around business and knowledge processes

supportfor filling in

the questionnaire

awarenesscreation

requirements compilation table – organized around technological services and functionalities

user’s requirements table

developer’s coverage table


Finally the users assessed the respective importance of the offered services and functionalities with respect to their business in a user requirements table, and the developers described what they can already offer or plan to offer within the scope of the Know-Net project in their coverage table.

In the next workpackage 3, designing the Know-Net tool will build upon the analysis of which requirements were rated the highest in the user requirements tables, and which offers given by the developers can be realized with a reasonable effort with respect to the resources available. They will be further compiled into an “escalation strategy” identifying necessary first steps, optional later extensions, and interesting open problems for possible further projects in this context.

4 Linking Requirements to the Know-Net Framework

Requirements engineering was already the first benachmark for the usability and useful-ness of the Know-Net holistic framework developed in workpackage 1. Although there were some small problems using the framework (mostly coming from the necessity to have very clear and distinctive definitions of terms and categories when working towards a tool architecture which are not yet always clear, or clearly known to the involved parties, or clearly understood by all involved parties) which might give reason to minor modifications in later project phases, it turned out to be, at least, a good starting point for organizing work and for getting a comprehensive perspective.

Organisation

TeamInter-organisation

Individual

KMInfrastructure

Figure 2: Levels of Knowledge Networking

Having a look at Figure 2 and having also in mind that the technical goal of the Know-Net tool is to integrate the process-oriented view on KM – which is very much about collaboration levels and networking – with the product-oriented view – which is very much about the Knowledge Assets at the heart of the KM infrastructure in the middle of the picture – it is already clear that many simpler frameworks focussing on just one aspect cannot cope with this ambitious goal. On the other hand, requirements engineering (see the appendices at the end of this document) clearly showed that both views are essential for the required technical support, and that also most networking



levels described in Figure 2 must be considered still in very first solutions (except for the interorganisation level which has not been considered in our first prototype scenarios).

To say it again, at the highest level in the holistic framework, the process-centred approach is deeply connected to the social interaction and different communities of knowledge workers. Therefore, it is linked strongly to the four levels – the individual, the team, the organisation and the Inter-organisation.

Within the KM Infrastructure, the Knowledger suite of tools is strongly linked to the processes, but also strongly linked to the strategy, structure and systems too (see Figure 3 for the next more detailed view on the Know-Net framework zooming deeper into the KM infrastructure). You might say that through strategy it is strongly linked to identifying the critical Knowledge Assets and developing new structures, systems and processes to capture them. You might say the product-centred view better manages the Knowledge Assets for the individual, team, organisation, and inter-organisation.

Organisation

TeamInter-organisation

Individual

Strategy

Structure

Systems

ProcessesAssets

Figure 3: Overview of the Know-Net Framework

In essence, the process-centred view of knowledge management is concerned with redesigning the key business processes in an organisation, with a new and special emphasis on using the process itself, as a key enabler and facilitator to capture knowledge, retain, share knowledge and, through the social interaction and collaboration of the group, through performing this process, create new knowledge as a result. The process-centred view is concerned with the mega processes involved in selling, serving, developing people and developing new products and services through innovation.

Thus, the process-centred view, by design, ensures that only highly relevant to the business objectives, and high quality information is input into the system. On the other hand, the product-centred view to knowledge management is concerned more with software products to enable people to find information, and manage information through



push/pull technology, agentware, searching, indexing, data mining, pattern recognition, user profiling etc. This ensures better connection of people to relevant information, information to information, and information to people.

By combining the process-centred approach with the product-centred approach, we achieve the total knowledge management solution, as each part alone only deals with one aspect of knowledge management.

Zooming another step deeper into the Know-Net holistic framework, we come to the Knowledge Management Processes shown in Figure 4. which can be classified within the following five groups (see also Figure 4): acquisition, organisation, dissemination, use and creation.

To quote some paragraphs from deliverable D1:

Knowledge acquisition processes include, e.g., the identification of knowledge needs, the capture and collection and/or import of knowledge, etc. Before investing heavily in the development of new capabilities, companies should know what knowledge and expertise exist both inside and outside themselves. One way to increase internal knowledge transparency is by creating knowledge maps, which support systematic access to parts of the organisational knowledge base.

Knowledgeacquisition

Knowledgeorganisation

Knowledgesharing

Knowledgeuse

Knowledgecreation

Figure 4: Knowledge Management Processes

Knowledge organisation includes the interpretation, analysis codification, indexing, aggregation, filtering, synthesising, packaging, archiving, and link of knowledge to its context. Of course critical tasks include the maintenance and knowledge “purging” functions.

After knowledge has been acquired or created, it must be carefully organised and preserved. Many companies complain that in the process of reorganisation they have lost part of their corporate memory. This collective amnesia is often the result of the destruction of informal networks, which steer important but little-observed processes. To avoid the loss of valuable expertise, companies must shape the processes of selecting valuable knowledge for preservation, ensuring its suitable storage, and regularly incorporating it into the knowledge base.


Knowledge acquisition

Knowledge organisation


Knowledge dissemination includes mechanisms for knowledge distribution (such as pro-active distribution, event-based distribution, subscription-based distribution, etc), schemes such as “targeted” push and/or pull sharing, the commercialisation of available knowledge, the development of trust in corporate knowledge, etc.

In making knowledge available and usable across the whole organisation, the critical questions are: who should know what, to what level of detail, and how can the organisation support these processes of knowledge distribution?

Knowledge creation consists of all activities intended to produce new knowledge both at the individual and the collective level.

The main processes for individual knowledge creation rely on creativity and on systematic problem solving. Creativity may be called the chaotic component of the knowledge development process and the capability of problem solving the systematic component. The knowledge management system must support both components, for example through traditional tools such as corporate proposal systems that may be revitalised or reused.

Collective knowledge creation involves the learning dynamics of teams. Management must ensure that team members have complementary skills and that each group as a whole has defined realistic goals. Moreover, cultural issues like an atmosphere of openness and trust, play significant role and allow the intensity of communication that makes collective learning results superior to individual ones.

This group of processes includes the application of knowledge in corporate services or products and in supporting the delivery of value to customer.

The productive deployment of organisational knowledge in the actual delivery of products and/or services is the heart of knowledge management. Successful identification and distribution of critical knowledge does not ensure its daily use. Without consistent use, there is a high probability that any knowledge management systems will decay in quality, and the investment will be wasted. The potential user of knowledge has to see a real advantage in order to change his or her behaviour and “adopt” the knowledge.

Since our user partners think in terms of their business processes and their (planned, ongoing) Knowledge Management processes, these KM processes turned out to be the appropriate level of abstraction for organizing our requirements questionnaire (besides a bit rearrangement and renaming). More general categories would be too vague in order to provoke specific requirements, and more technical ones (like the categories in the compilation table in the next section) would be too difficult to understand and interpret by normal users. Of course, in order to come from these still quite general categories to concrete and detailed statements about necessary software support, we backed them up with quite a big amount of supplementing questions showing up interesting dimensions to think about within the several categories. Since we deal with technical support for handling explicit knowledge and information, the categories and supplementing questions were mainly about the document side of knowledge management, but not exclusively. Further, we had to start with the “unbiased use case”


Knowledge dissemination

Knowledge creation Knowledge use


to have a comprehensive picture of the ideal scenarios for the users, asked for knowledge assets used and created, elaborated a bit on the business tasks to be performed by the employees (since you need to understand to which purpose people want to use and reuse their knowledge), and ended with questions about technical requirements which have to regarded.

To sum up, the organization of the questionnaire for gathering structured lists of user requirements mainly reflected the issues of the holistic framework shadowed in Figure 5

Figure 5: Know-Net Framework and RE Questionnaire

As a last remark about the holistic Know-Net framework and the results of Know-Net requirements engineering one can observe that – again, except for the inter-organisation networking level – all collaboration levels of our framework are needed to describe the particularities of our three users, as well as FHBB’s development approach:

Gooch Webster is very much focussed on individual, highly document-centred knowledge workers which need frequent, fast, and precise access to a large amount, of possibly dispersed and unstructured information and knowledge. Team work is not the usual case, but specific links to certain individuals dealing with similar cases can be of utmost importance.

Planet is very much oriented towards project teams in daily work and is aiming at a higher support for company-wide reuse of knowledge and information. They need a strong link to existing work processes and have a strong structure in the kind they are doing their work.



UBS is also quite focussed on heavy team work, but with much more ad hoc work processes and structures, also less structure in team archives. On the other hand, they need a strong link to the company-wide network.

FHBB starts with very simple sorts of documents and completely neglects overall work processes. However, they focus on very flexible individual organization schemata for working within highly interactive teams.

Now that we have acquired the individual users’ requirements as documented in Appendix B, C, D, the next section will try to compile them into a more architecture oriented structure giving us clear hints for the development of the Know-Net tool.



5 Required Services and Functionalities

As a starting point for organizing technical requirements in Know-Net, we used the structure proposed in workpackage 1 for describing technical functionalities for Knowledge Management as it is shown in Figure 6.

Distribution Collaboration

Indexing Storage

Integration Search & Retrieval

Figure 6: ICT Services According to Deliverable D1

It turned out, that – at first hand – a number of problems occured because it seemed not too clear why certain services should belong to which category and how a so-organized account of requirements would lead to a helpful means for tool specification. Further, according to the project plan, we wanted to take into consideration also requirements from external sources, not specific to our Know-Net user partners. So, we reviewed a number of available KM technology surveys, also commercial ones, in particular the ones by Doculab and by the Ovum, respectively, in order to find out what requirements were considered there, and how they were organized.

It happened to be quite helpful to look at Ovum’s abstract KM system architecture shown in Figure 7. Essentially, the drawing shows nearly the same categories as ours, but arranging it in sort of an architecture makes it much clearer to see what belongs where, and why.

Having this Figure in mind, we shortly repeat the respective definitions of the six classes of ICT services introduced in Know-Net deliverable D1:

An information map defines the channels available for use by individual employees or the enterprise at large and describes the mechanisms available for information processing and knowledge formulation. When put to work, the information map will provide a representation of available knowledge (knowledge bases, topics, sources, narrative summaries, higher-level descriptions, etc.). This will help ensure that


Indexing, mapping and classification services.


employees know "what they know". Automated indexing routines can be facilitated in

order to ensure complete synchronisation of indices and data sources and thereby ensure that employees know "where is what they want".

Search and retrieval services should provide transparent access to multi-platform, heterogeneous sources, including Internet/intranet sites, file servers/databases, popular proprietary formats, legacy IS. Various types of search services should be accommodated, i.e. hierarchical (e.g. traversing hyperlinks), attribute (query-type searching), and content (e.g. crawler-type searching of popular WWW search engines).

Distribution and publication services can include: subscription-based approaches on internal (such as bulletin boards) and external (such as WWW sites) information sources; and push and “smart-pull” approaches coupled with intelligent, selective


Search, retrieval and navigation services

Distribution, Publication, and Filtering Services

Applicationlayer

Interface

Knowledge managementservices

Corporate taxonomy

Information & process management

Information & knowledge sources

Infra-structure

CompetitiveIntelligence

Best PracticeSystem

Product Development

Project Spaces

Knowledge Portal

Knowledge Map

Knowledge Repository

WP DBMS EDM E-Mail WWW People

E-Mail, file servers, Internet/intranet services

Discovery Services Collaboration Services

Figure 7: The Ovum Abstract KM Tool Architecture


mechanisms of content relevance assessment that will provide useful knowledge while preventing information overload.

Integration with: application and tools that are currently used within the organisation (e.g., word-processors, spreadsheets, databases, etc.) This way, the infrastructure will tap into the flow of information that is already happening in the organisation, and therefore will improve acceptability by the users.

This is a core service, that is being facilitated by the use of technologies that span from relational database management systems and document management systems to AI-based corporate memories. Such services should allow access to existing applications, services and databases, have layered storage capabilities that provide different abstraction levels, and provide intelligent assistance to users for the execution of knowledge organisation processes.

This is a group of services offered by technologies providing rich, shared, virtual workspaces in which interactions occur between people who share a common goal. Indicative collaboration services include email, messaging, on-line discussions, electronic scheduling and meeting, video and audio conferencing, virtual workshops, just-in-time workgroup alerting, etc.


IntegrationStorage and Metadataservices

Collaboration services


Now, the following table gives an overview of all ICT services we found potentially useful in our application scenarios.

Distribution, Publishing, and Filtering notification / push mechanisms

1. mandatory readings2. changes/modifications3. topic oriented filtered delivery

user participation / publishing support

Collaboration e-mail communication real-time communication document-sharing, white-boarding net-based meetings, conferencing threaded discussions availability (who is accessible, how)


Gateway to user and applications

Corporate taxonomy

Information & process management


IntranetInfrastructure


Best PracticeSystem

Product Development

Project Spaces

Knowledge Portal

Collaboration Service

Knowledge Map


WP DBMS EDM E-Mail WWW WP


Knowledge manage-ment services

Integration Services

Search, Retr., Navigation

Collboration

Indexing

Storage & Metadata

Distrib., Publish., Filtering

Figure 8: Know-Net ICT Services in the Ovum Architecture


company yellow pages

Search and Retrieval browser interface to KM services basic: keyword and boolean search basic: metadata based search advanced: combine criteria advanced: semantic analysis search over major document types restrict search area flexible ordering of search results smart pull mechanisms

Indexing basic: corporate taxonomy / ontology basic: metadata support, incl. versioning basic: short content description flexible knowledge organization & reorganization corporate thesaurus advanced: pattern detection / text mining advanced: document clustering user profiling map sources / contents to business views

Storage storage support for all major document types easy addition / modification of documents smooth integration of existing documents convert printed into electronic documents export facilities flexible and powerful access control mechanisms archiving

Integration homogeneous env. for person, group, and company project management integrated workflow and document management collaboration with other tools

Administration Support access mining link maintenance support retrievability checks broker for external information and services

Technical Requirements scalability reliability security, integrity, and authenticity



Figure 9: Detailed List of Potentially Useful ICT Services

In the following we will elaborate a bit on the more unusual functionalities and explain what we mean concretely.

1. Distribution, Publishing, and Filtering

1.1. notification & push mechanisms:

1.1.1. mandatory readings: More and more it happens that information / knowledge people should be aware of is not distributed individually but is posted on intranets, in Lotus Notes databases, etc. An example is the IPR proposal for Know-Net prepared by Ron Young for the Fontainebleau meeting. In addition to reading the proposal, the consortium members were expected to comment it, but apparently nobody did so. Personalized KM tools could make the user aware that there are readings the user is expected to have a look at, and remind the user that a response is due. No doubt, such facilities would be extremely helpful and could relieve humans from writing notification mails, asking for replies, sending reminders, etc. However, there is a certain risk that such tools could be perceived as intruding. Thus, a careful design, implementation and introduction seems extremely important in order to gain the user's acceptance.

1.1.2. changes / modifications: Taking the UBS Bank Web as an example, the user can ask to be informed when new documents are added or changes occurred. This is done by sending the user a notification e-mail. Unfortunately, the user easily can get overwhelmed by such e-mails (which typically arrive when one is busy with a different matter). A more user-friendly way of notification could be to provide a specific view which shows all the documents of a site (or part thereof) which have changed or have been added since the last visit of the user (one could call this a personalized pull notification). This view could be provided by a personal information assistant which knows what topics and changes the user is interested in and what modifications she is already aware of.

1.1.3. topic oriented filtered delivery: corporate Intranets use to be hierarchically organized. This is, for instance, helpful when one wants to find out more about a department. However if one is interested in a particular topic which is important within different departments (for example, one would like to know what is going in the area of Java programming within the different IT departments of the company), a topic-centred organization would be more helpful. Of course, one could perform a search. However, a list of documents dealing about the topic in some way is clearly not very helpful neither. A more suited form would be to organize the documents along an ontology or a thesaurus, this way constructing some sort of knowledge map. Having such a knowledge map as a browsing and retrieval interface, typical queries could be stored as



„scripts“ performed by personal information agents depending on given trigger events (typically from time to time).

1.2. user participation / publishing support:

The KM tool shall allow changes (comments, the adding new documents etc.) by the user wherever possible as this promotes the sharing of knowledge. In the UBS case, for example, ithe Bank Web s of static nature - the regular user does not have the possibility to add information / knowledge to the Bank Web. Taking an organizational / management point of view, UBS practices an in-house culture where the employees are passive consumers of information / knowledge. This definitively does not promote the sharing of knowledge.

2. Collaboration

2.1. e-mail communication

2.2. real-time communication

2.3. document-sharing, whiteboarding

2.4. net-based meetings, conferencing

2.5. threaded discussions

2.6. availability (who is accesible, how)

2.7. company yellow pages

3. Search and Retrieval

3.1. browser interface to KM services

3.2. basic: keyword and boolean search

3.3. basic: metadata based search

3.4. advanced: combine criteria

3.5. advanced: semantic analysis

3.6. search over major document types:

Lotus Notes has severe shortcomings as certain document types are not indexed, for instance PDF documents, which means that information/knowledge stored in a PDF document can not be searched for. To overcome these shortcomings, mainly third-party tools available on the market should be considered.

3.7. restrict search area:

to search in a specified area of a knowledge pool or the entire pool, as well as to search in external sources. This will be a main aim of the envisioned system approach sketched in Section 7.

3.8. flexible ordering of search results



3.9. smart pull mechanisms

4. Indexing

4.1. basic: corporate taxonomy / ontology:

The classification of document contents best works with a formal conceptual model of the domain in form of a thesaurus or an ontology. Such conceptual models have to be built up first and then to be maintained continuously. Both tasks require manual work but could be supported by text mining techniques. The DFKI KnowMore project worked out some steps towards integrated semi-automatic content-characterization plus ontology evolution on the basis of text mining techniques.

4.2. basic: meta data support, including versioning:

Meta data at UBS, are for instance::

who is the author/owner,

when was the document created,

when was it modified, as well as

who is the intended audience (a particular group of persons, any staff member, outside world, etc.).

what is the status of the document (draft/official),

what are previous/later versions,

what is the lifetime of a document (for instance, a proposal for a meeting agenda could be archived after a short time in order to keep the document collection slim and up-to-date)

4.3. basic: short content description: There are many occasions where short description of a document´s content would be extremely helpful. For instance, a user could better judge the relevance of a document before clicking on a URL and waiting for the application to be launched and the document to be loaded. Although much research efforts have been spent, textual content descriptions are out of the reach for broader applications, at least for the near future. However, content deceptions in form of a set of concepts/terms seem feasible. In addition to give the user a better idea of a document´s contents, such descriptions could also be used to provide alternative, content-oriented views of a document collection. Further, the same underlying technique also could be useful to indicate changes between two versions of a document. The DFKI Karat project extensively used this technique for information modeling.

4.4. flexible knowledge organization and reorganization

4.5. corporate thesaurus

4.6. advanced: pattern detection / text mining:

Text mining aims at deriving patterns from documents of a collection. Such



patterns typically are collection-specific phrases or multi-word terms. Text mining tools could learn, for example, from analyzing the Bank Web of UBS that "service level agreement" is an organisation-specific phrase and SLA is its abbreviation. Such collection-specific phrases could be used, for instance, to augment a domain-specific thesaurus. Further, they could improve the automatic derivation of content descriptions.

4.7. advanced: document clustering

4.8. user profiling:

I would propose quite a general definition here: Both spontanously querying an information system and defining an information agent for active delivery and notification require some symbolic representation of user interest. In the case of a spontanous query, this is typed by the user ad hoc and by hand. For active delivery services, like personal information agents, or Internet search services, such a representation of an information need is an important part of the agent's persistent knowledge base. It can be seen as a predefined query stored persistently. We use a comprehensive definition of "user profiling" extending the above basic notion along three dimensions:

Ownership of profiles: represented interests are not necessarily bound to single persons, but can also desribe the information needs of groups, teams, or departments, or may describe clusters of users with similar interests like they are grouped together implicitly for technical reasons in collaborative filtering approaches. Specific information needs may even be attached to certain business tasks or organizational roles.

Expressiveness of profiling: specific profiles, especially individual ones, may not only comprise a simple representation of query terms, but also additional information helpful for efficiently processing information system queries. In particular, this could be typically useful query expansions or query reformulations.

Maintenance: a further important advantage of explicit, persistent profiles which shall be subsumed under the "user profiling" issue is the possibility of adapting and evolving profiles over time with machine learning techniques based upon user observation and explicit user feedback.

4.9. map knowledge sources / repository contents to business views

5. Storage

5.1. storage support for all major document types:

The most frequent document types at our user sites are Word, Excel, PowerPoint and Adobe PDF. A complete list of document types, however, cannot be specified. The KM tool should support the major document types and should be extensible such that additional document types can be handled. For this functionality, it is nearby to have a look at third-party software providing this functionality.



5.2. easy addition / modification of documents

5.3. convert printed into electronic documents:

It should be possible to integrate printed documents into a document collection (by scanning the documents and applying OCR). The handling of such documents should be (almost) the same as for electronic documents, e.g., search on such documents should be possible. DFKI has a lot of experience in the areas of document analysis and understanding.

5.4. smooth integration of existing documents:

UBS-CRV as well as Planet and G&W have already put together a knowledge pool consisting of a quite large set of documents such that a smooth transition from the current environment to the new KM solution is extremely important. Further, one has to keep in mind that many of the documents (at least at UBS and many other companies experimenting with Intranet solutions) are HTML documents with links to other documents. It is not clear yet what happens with these links when the documents are integrated into a Lotus Notes database. Leaving the documents where they are and using links to the original documents within the KM tool might not be a good solution neither, as this may weaken access control.

5.5. export facilities

5.6. flexible and powerful access control mechanisms

5.7. archiving:

Due to new forms of distributing information and sharing knowledge, it may not only be necessary to archive documents of general importance (e.g., directives) but also documents which gained the interest of a single reader. Let us consider a small example: a user bases a decision on a piece of information found on the companies intranet. Some time later the user is asked to explain the decision. This may be difficult, when the piece of information has been removed or overwritten in the meantime! Individual storage would burden the user and would rise authenticity and integrity problems (HTML files and Word documents can be easily changed). Thus, the reader of a document should be able to request archiving. Of course, it should still be possible to request archiving on creation time.

6. Integration

6.1. homogeneous environment for person, group, and company

6.2. project management

6.3. integrated workflow and document management

6.4. collaboration with other tools

7. Administration Support



7.1. access mining:

This idea came up at UBS. There, a knowledge manager relies on statistics about accesses to his KM pages on the intranet. This helps him to improve the organization of the information presented, e.g., what should be included in a particular view. More sophisticated analysis tools could be employed in order to find out typical access patterns of different types of users or even single users.

7.2. link maintenance support:

Tools should help the knowledge manager to find dangling links, outdated documents (i.e. that were not modified after a certain date), cyclic links, etc.

7.3. retrievability checks:

Typically, many documents on a corporate Intranet are not found by the official search engine, even if an appropriate query is given. The search engine relies on a list of locations to be indexed. Apparently, this list is not complete. To avoid write-only information/knowledge, there should be tools that make sure that a document is retrievable by the corresponding search engine.

7.4. knowledge broker for external information and services

8. Technical requirements

8.1. scalability

8.2. reliability

8.3. security, integrity, and authenticity

In the next table, we summarize the ratings of the three user partners. Each entry in a user compilation table is indicated by an ‚x‘. The sum of all weights is evaluated as follows: #nice*2points + #important*3points + #essential*4points


1. mandatory readings 10 pts 2. changes/modifications 11 pts

3. topic oriented filtered delivery 11 pts user participation / publishing support 11 pts

x xxx xxx xx

xxx

Collaboration e-mail communication 8 pts real-time communication 6 pts document-sharing, white-boarding 10 pts net-based meetings, conferencing 8 pts

x xxxxxx xxxx x



threaded discussions 12 pts availability (who is accessible, how) 10 pts company yellow pages 11 pts

xxxxx xx xx

Search and Retrieval browser interface to KM services 11 pts basic: keyword and boolean search 12 pts basic: metadata based search 11 pts advanced: combine criteria 10 pts advanced: semantic analysis 7 pts search over major document types 12 pts restrict search area 11 pts flexible ordering of search results 10 pts smart pull mechanisms 10 pts

x xxxxx

x xxxx x

xx xxxx

x xxxx xxx x

Indexing basic: corporate taxonomy / ontology 12

pts basic: metadata support, incl. versioning 12

pts basic: short content description 11

pts flexible knowl. organization & reorganization 11

pts corporate thesaurus 9

pts

xxxxxx

x xxx xxxxx

x xxxx xx xxx xx

Storage storage support for all major document types 12 pts

easy addition / modification of documents 12 pts

smooth integration of existing documents 12 pts

convert printed into electronic documents 6 pts

export facilities 9 pts

flexible / powerful access control 8 pts

archiving 9 pts

xxxxxxxxx

x xxx x xx xxx x x

Integration



homogeneous environment 8 pts project management 9 pts integrated workflow and document mgt. 7 pts collaboration with other tools 11 pts

x xxxxx

xx xx xx

Administration Support access mining 9 pts

link maintenance support 11 pts

retrievability checks 10 pts

broker for external information and services 11 pts

x x xx xx

x xxx xx

Technical Requirements scalability 11 pts

reliability 10 pts

security, integrity, and authenticity 11 pts

x xxx xx

x xx

Figure 11: The Identified ICT Services Rated by the Users

One can see that almost all identified services were rated quite important except for some “exotic” ones the users have possibly no clear idea about possibilities and usefulness of them. So, finding these services was the more worthful activities than rating them. In order to come to really distinctive statements about what should be done for a user and what not, one would nee a more differentiated rating. Nevertheless, one can see that powerful searching on the basis of a flexible indexing machinery has the first priority, and that technical aspects like scalability etc. are also quite important for the users.



6 Process-Specific Requirements

In the next table, Knowledge Associates shows what can and should be done within Know-Net tool devleopment, from their point of view. The following ratings are used:

KGR 2.0: Knowledger 2.0 can do this today

KA KnowNet: KA have the technical capability to develop this user requirement from Knowledger 2.0 for KnowNet, subject to sufficient budget resources being available. Once we have agreed what DFKI, FHBB and KA can do, we could in WP3 prioritise the development sequence, estimate work durations and finalise specifications and first prototypes. Having now studied the user requirements and expectations of an ideal KM system for them, it is quite clear that we cannot possibly fulfill all their requirements within the scope of Know-Net , but we should certainly attempt to meet all the 'essential' requirements, if possible, and seek common agreement to a prioritised list for the remainder, in WP3.

No KA : KA cannot do within Knownet.




KA Know-NetKA Know-NetNo KAKGR 2.0 & KA Know-Net

Collaboration e-mail communication real-time communication document-sharing, white-boarding net-based meetings, conferencing threaded discussions availability (who is accessible, how) company yellow pages

KGR 2.0KA Know-Net1

KA Know-NetKA Know-NetKGR 2.0KA Know-NetKA Know-Net

Search and Retrieval browser interface to KM services basic: keyword and boolean search basic: metadata based search advanced: combine criteria advanced: semantic analysis search over different document types restrict search area flexible ordering of search results smart pull mechanisms

KGR 2.0KA Know-NetNo KANo KANo KAKA Know-NetKA Know-NetNo KANo KA

Indexing1 Through integrating Lotus Sametime technology.



basic: corporate taxonomy / ontology basic: metadata support, incl. versioning basic: short content description flexible knowledge organization & reorganization corporate thesaurus advanced: pattern detection / text mining advanced: document clustering user profiling map sources / contents to business views

KGR 2.0 & KA Know-NetKA Know-NetKA Know-Netpartially KA Know-NetNo KANo KANo KANo KApartially KA Know-Net


No KApartially KGR & KA Know-NetNo KANo KApartially KA Know-Netsignificantly KA Know-NetKA Know-Net

Integration homogeneous env. for person, group & company project management integrated workflow and document management collaboration with other tools

KGR 2.0 & KA Know-NetKGR 2.0 & KA Know-NetKGR 2.0 & KA Know-Netpartially KA Know-Net


No KApartially KA Know-Netpartially KA Know-Netpartially KA Know-Net


KA Know-NetKA Know-Netsignificantly KA Know-Net

Figure 12: Knowledge Associates’ Coverage Table

Since Knowledge Associates‘ point of view is very much oriented towards business and knowledge management processes, it makes sense to additionally show their possible contributions to the wished process support identified by Planet and Gooch Webster’s:

Title Importance1. Competence Development1.1 Map employee competencies to a pre-defined structure of company-wide competencies

KGR 2.0

1.2 Grade competences using predefined criteria KGR 2.01.3 Provide space for annotations, preferences, descriptions and comments to competences

KA Know-Net



1.4 Develop personal plans for further development of competences. KGR 2.01.5 Link former and current consulting assignments, training seminars and learning experiences of consultants to their competences

KGR 2.0 & KA Know-Net

1.6 Provide space for company training material and link it to competences KGR 2.0 & KA Know-Net

1.7 Search for consultants based on competencies. Results whould include: KGR 2.01.7.1 experiene from previous and current assignments, KA Know-Net1.7.2 seminars attented, KA Know-Net1.7.3 personal qualificatoins, KA Know-Net1.7.4 personal development plan, KA Know-Net1.7.5 personal preferences KA Know-Net

1.8 Link to time and resource management system to supply info about current and foresseable availability of consultants

KA Know-Net

1.9 Provide an electronic space where the advisor and the “advisee“ could collaborate in an unstructured way, i.e.:

KGR 2.0

1.9.1 arrange meetings, KGR 2.01.9.2 discuss on-line but not real-time concerning the consultant’s development plan,

KGR 2.0

1.9.3 monitor periodically the agreements they have reached concerning the consultant’s core competences and critical skills,

KGR 2.0

1.9.4 make changes and suggestions for revising the personal development plan

KGR 2.0

1.10 At the 2 business-units and company level monitor the overall current status of competences and skills and the expected organisational competence “map” - based on the plans of each individual consultant.


1.11 At the company level provide support for scheduling and monitoring the training requirements, providing on-line availability of the training material (electronic and printed documents and presentations), identifying requirements for further training, etc.

KA Know-Net

1.12 Link targets from business plans and foreseeable consulting assignments to the competence map of the company and be able to even make changes and modifications to the overall competence map, in order to define the “future” company-wide competence map


1.13 Evaluate the missing competences of current business-unit competence maps and perform organisational “gap analysis” between the current state of competences and future requirements (based on the results of the business plan).


2. Requirements for Supporting Service Delivery (Consulting Assignments)2.1 Analysing previous company experience for a similar assignment; hence what is needed is a repository facility that includes the project plans (i.e. time and resource plans) for all consulting assignments (PLANET is in the process of establishing such a company-wide system in MS Project). Normally the project director with the help of the project manager would refine the time, activity and resource plan for the new assignment based on the contract with the client and the previous experience within the company (if any). Hence facilities are required to:

KGR 2.0

2.1.1 link to time and resource management system (MS Project application)

KA Know-Net

2.1.2 link to project and proposal database (MS Access) KGR 2.02.1.3 find similar (based in customer, service) past assignments, KGR 2.0 & KA

Know-Net2.1.4 locate their project plans, KGR 2.0 & KA

Know-Net2.1.5 locate the people responsible for such plans, KGR 2.0 & KA

Know-Net2.1.6 “talking” to these people - e.g. in a collaborative space or on the phone - in order to come up with the new plan.

KA Know-Net



2.1.7 which were the major problems (if any?) KA Know-Net2.1.8 which were the major results? KA Know-Net2.1.9 which are the “traps” a project director may fall in when defining the plan? Etc.

KA Know-Net

2.2 Functionalities for the project director and project manager to monitor and update the project plan (i.e. tasks, their duration, resource allocation and usage - i.e. people and man-months allocated per task, etc.)


2.3 Facilities to search, find and retrieve related material during the execution of a project, from:


2.3.1 Previous assignments that have been carried out within the company,


2.3.2 Material that the company has access to from other consulting firms or

KA Know-Net

2.3.3 Material that can be found on the Internet (e.g., general news, news from consulting companies, business and management sites, industry-specific sites, etc.)

KA Know-Net

2.4 Project-level collaboration tools KGR 2.0 & KA Know-Net

2.5 Support the Knowledge Input - Output process, as described in ‘ideal use case scenario’ and D9.1

KA Know-Net

2.6 Assist consultants in the classification of new knowledge KA Know-Net2.7 Keep track of the use of the system (e.g. number of hits, most asked items, etc.)

KA Know-Net

3. Requirements for Supporting Business Development3.1 Space for “account managers” to manage clients’ relationships to PLANET KGR 2.03.2 Capture and organise the experience that all PLANET’s consultants may have with each client (such experience may arise both during assignments, i.e. within the service delivery process, and during business development)

KGR 2.0

3.3 facilities to capture, organise and distribute information about new and prospective client contacts, contact information per customer, etc.

KGR 2.0

4. Requirements for Supporting Bid Preparation4.1 The facilitation of handling such “administrative” and document-intensive work as:

4.1.2 Preparing letters of collaboration (e.g. in the case of a consortium)

KA Know-Net

4.1.3 Using company profiles (in Greek or English, in various formats - short or long, etc);

KA Know-Net

4.1.4 Managing CV’s KGR 2.0

4.2 Locating possible contact points for partners or subcontractors (e.g., from a repository of companies which is organised according to the company’s competences and services provided)

KA Know-Net

4.3 Electronically supporting the communication and documents exchanged with possible partners and subcontractors


4.4 Locating and (partly) reusing material from the documents of previous bids KA Know-Net

5. General Requirements

5.1 Security, Integrity and Authenticity (Best Internet possible is sufficient) see above



5.2 Interoperability with MS Access and MS Project on specific tasks (to be clearly defined during the application of the software on site)

KA Know-Net

6 Basic KM requirements

6.1 Support for assigning roles with specific rights on documents (e.g. read only, write, update, delete)


6.2 Knowledge organisation and re-organisation (ability to modify the ‘knowledge organisation’ or ‘knowledge schema’)

see above

7. Document handling and management

7.1 Smooth integration of existing documents. Currently there exists 10-year experience in electronic documents.

KA Know-Net

7.1.2 At least, the system should provide support for full-text searching within these documents.


7.1.2 Better, the system should facilitate the (automated??) transition of links (meta-information, selected documents?) from the old repository on the file server to the new system

No KA

7.2 Full support for all major documents types (PDFs, HTML, MS Office) No KA

7.3 Metainformation partially KGR 2.0 & KA Know-Net

7.4 Flexible access control mechanisms partially KGR 2.0 & KA Know-Net

7.5 Archiving partially KGR 2.0 & KA Know-Net

8. Information access

8.1 Search facilities see above

8.1.1 Search based on meta-information and full-text see above

8.1.2 Ability to index usual document types No KA

8.1.2 Search in a specified area or the entire knowledge base see above

8.1.3 Seamlessly incorporate external sources (e.g., specified WWW sites)

No KA

8.1.4 Order search results according to different criteria see above

8.1.5 Preview results (e.g., show summary or abstract) see above

8.2 Intuitive Navigation facilities (e.g., site maps, etc.) KGR 2.0 & KA Know-Net

8.3 Improved access to information and knowledge KGR 2.0 & KA Know-Net

8.3.1 Mandatory readings see above

8.3.2 Push mechanisms available to selected users and to specified audience of which the originating user has control.

No KA

8.3.3 new or modified documents see above



9. Project Management and support for team work (as described in Requirements for service delivery, e.g. support for the K-I-O process

see above

10. Better support for discussions KGR 2.0 & KA Know-Net

11. Ontologies/Thesauri and Classification No KA

11.1 Automatically generated content description No KA

12. Tools for the knowledge manager

12.1 Access mining (as discussed in Requirements for Supporting Service Delivery)

No KA

12.2 Document and linik maintenance tools see above

12.3 Retrievability of documents see above

Figure 13: Knowledge Associates’ Coverage Table for Process Support



7 Product-Specific Requirements

In this section, we mark possible contributions to be done by the DFKI. We use the following categories:

significantly DFKI: can and must be done within the product-centric view

partially DFKI: can and should be done, maybe together with or by other developer partners

optional: could be done, but seems not preferable at first hand due to a high research part, or very much development effort, or because it seems not to be central to the Know-Net goals

long-term: interesting point, but cannot appropriately be dealt with in the Know-Net project; could be an interesting point for future work

No DFKI: not a part of the product-centric view




significantly DFKIpartially DFKIoptionalpartially DFKI


No DFKINo DFKINo DFKINo DFKINo DFKINo DFKIoptional

Search and Retrieval browser interface to KM services basic: keyword and boolean search basic: metadata based search advanced: combine criteria advanced: semantic analysis search over different document types restrict search area flexible ordering of search results smart pull mechanisms

partially DFKIsignificantly DFKIsignificantly DFKIsignificantly DFKIoptionalpartially DFKIsignificantly DFKIsignificantly DFKIpartially DFKI




significantly DFKIsignificantly DFKIpartially DFKIsignificantly DFKIoptionallong-termoptionaloptionaloptional


partially DFKIpartially DFKIoptionalsignificantly DFKIoptionalNo DFKIpartially DFKI

Integration homogeneous env. for person, group, and company project management integrated workflow and document management collaboration with other tools

partially DFKINo DFKIoptionaloptional


long-termpartially DFKIoptionaloptional


partially DFKINo DFKINo DFKI

Figure 14: DFKI’s Coverage Table



8 Requirements for the Individual Knowledge Portfolio

KnowPort investigates the tier of INDIVIDUAL knowledge management with the aim to develop a tool "KnowPort" (SW tool and means) that allows knowledge workers to manage knowledge (Knowledge Portfolio) so that learning and knowledge sharing be supported and improved.

KnowPort consists of 4 subprojects MailTack, FileTack, TaskTack, WordTack), each with own user requirements specifications. Therefore only general considerations are made here (more detailed information about the overall KnowPort scenario can be found in Appendix E). Our general requirements (non tool-specific) are very near at the requirements of our partner UBS, the essential difference concerns the hierarchy of access which is in a bank of essential importance, but is of course not relevant at individual level.

6.1. General considerations

The identification of knowledge

We understand our project as individual or personal knowledge management. We are are concerned with how we acquire knowledge, use it and share it.. From our investigations into these questions we expect insight and therefore a more appropriate management of knowledge. Therefore, we understand by knowledge "Management" in the first placethe task of identifying knowledge, namely in a double sense: we want to develop a concept of Knowledge and at the same time our own knowing. The means and tools developed and used in this process may act as concrete results (artifacts) of our project.

Since we understand our project as individual knowledge management, we can ask us, which requirements we ourselves make on our product. In this case, we give up to be representative, however we win a lot more specific and more thorough knowledge than with surveys of course. Particularly, in this procedure the so-called hidden or tacit knowledge becomes available to us. In addition, we can regularly update our requirements, if that is suggested by new insights.

Since we develop tools and techniques, we can understand the resulting knowledge as technology. In this sense, we perform the development of the knowledge tools in order to develop our understanding of knowledge. This methodology has an autopoietic character in that the very results of developing (the software tools) are used as means for doing that same developing work. A consequence of this selfreferentiality consists in one basic requirement to the tools we develop: that they should serve us during our progress in the process of developing them as well as in the process of understanding



what we mean by "knowledge". When we start we imply to know what knowledge is. Then the tools and their use make explicit, what we mean by knowledge.

What ever knowledge is, at least partially it appears in form of data, which may act as external memories. We distinguish different data related actions as speaking, making a phone call, writing, sending e-mails, and so forth. However here we are concerned primarily with data-related operations, which lead to electronically written texts, i.e.to texts that can be processed on computers (Inter- and Intranet).

6.2. Knowledge relevant features of our activity

The holistic engineering processAt the beginning we speak about what we construct, that is which one of our activities we intend to systematise and realise as tools and how we plan to do this. We collect ideas in form of tools and texts, that seem useful to us. In order to keep survey we arrange the texts, we manage them. We reflect on how to do that best (best practice). We specify a wish list, whichsays what we would like to have and therefore want to construct.

We document what we collect:

there are books and articles about knowledge and knowledge management

there are experts, that one can ask

we have our own texts to these or other topics

we have electronic and conventional correspondence to the topic

we remember conversations, lectures etc.

there are software tools and methods

there are psychological theories and methods

Everything, what we collect from the viewpoint of the development of knowledge, can be understood as "Artefact". A partial set of the artefacts we are interested in, are texts.

A task - one the we unconsciously already perform everytime - is ordering these artefacts, bringing them into relations and so forth. Therefore, we can observe, what we do with that artefacts. We manage the texts on a server which can be accessed by all of us. We associate the texts via links and build a structure.

Of course the texts that we collect are not arbitrarily scattered in the world, they are already embedded into structured contexts. Therefore the way we handle the texts heavily depends on the source and on the text type.

O Inside of these documents we distinguish different categories that contain different types of knowledge and must therefore be handled differently from the viewpoint of knowledge management. E-mails (letters) for example contain other kinds of knowledge than reports. E-mails are usually fragmentary because they document a current communication process, whereas reports are built up



systematically and integrated. Therefore we have distinguished the four subprojects mentioned above in which we deal with different viewpoints on data.

Our tools and means must support us in this. More precise requirements are explained in the subprojects.

The organisation of knowledgeTypically knowledge is stored in an encyclopedia. However, the encyclopedia is necessarily context neutral and contains no knowledge about current cases. So no encyclopedia says for example what the stock trader recommended yesterday to me at phone or which conditions were linked to a specific offer which I received about four weeks ago. In practice however we want very frequently to look up such context binded, current knowledge.

In order to solve this problem we will pursue an action oriented approach in the KnowPort project which is based on our Trace-your-Tack-Principle: Trace your knowledge in action (on tack), in order to sail on the right tack in the knowledge ocean! KnowPort therefore, will be used for recording knowledge relevant events during the application of personal knowledge in the context of a current activity.

We therefore concentrate on the input side of the document administration in order to reduce retrieval complexity. We want to support the process in which information is linked into the knowledge base. Automatable processes will be automated. Interactive processes will be supported by visualisation.

Tool specific requirementsGeneral requirements

Hardware and software platform: Our tool shall be usable on the usual computers/operating systems, first of all on Personal Computers and workstation with Windows and Internet browsers.

Safety, reliability: Safety and reliability are defined in the field of the available hardware/software. The personal Knowledge management does not specify additional requirements.

Basic requirements

Integration with usual tools: Knowledge management contains many different fields as information retrieval, document management, workflow management, time and project management. Therefore a smooth collaboration between the used tools and sources is really important.

Knowledge organisation and reorganisation: Our tools should not force specific organisational structures. It is important that the tool allows the reorganisation of the data and knowledge base. This includes the definition of new (sub-)categories, renaming and deleting of categories, moving documents to new places, etc. Such activities should be possible at any time, also for running applications.

Document handling and management

Integration of existing documents: The integration of existing documents is a relevant goal. In particular it should be possible to import HTML documents with



their links. Updating local files after processing them on external platforms should be supported like in MS Windows. Integration of printed documents (from scanning and OCR readers) should be supported.

Integration of document type and format: All the most common document types and formats should be supported, in particular the formats of the most used text and graphics software packages (Word, Exel, Adobe, Corel and so forth)

Meta-information: The documents should be visible as objects with qualities like author, date, version and so forth.

Sequence formation between documents: Contributions to discussions which are stored in various places should easily be brought into sequences.

Ontologies, thesauri and classification: As the automatic derivation of content descriptions is still experimental, a mixed mode with short, rather general, manually assigned content descriptions in addition to automatically derived ones is possibly the best solution. The classification of document contents best works with a formal conceptual model of the domain in form of a thesaurus or an ontology. Such conceptual models have to be built up first and then to be maintained continuously. Both tasks require manual work but could be supported by text mining techniques.

Cluster and pattern recognition: Pattern recognition is usable on different tiers. It can be used in order to enlarge domain-specific thesauruses. In this way the automatic derivation of contents descriptions could be improved. More refined analysis tools can recognise and support typical profiles of different users.

Document and link maintenance tools: Tools should help the knowledge manager to find dangling links, outdated documents (i.e. that were not modified after a certain date), cyclic links, etc.

Retrievability of documents: One should be able feature the documents for particular search machines, to make sure, they are really found, if they correspond to the search inquiry.

Ergonomic requirements

The tools should be adjusted to the most used GUIs so that the users do not have to change their habits, where this is not combined with massive advantages. The tools shall give clear feedbacks and avoid asking cryptic questions.



9 Suggested Integration Approach

Regarding Figure 15 and Figure 16 we see that combining our two technological approaches achieves already a considerable coverage of the ideal Ovum KM scenario.

Requirements engineering clearly showed that the process view is very valuable in order to get useful knowledge into the system and to be linked to the corporate business and knowledge management mega processes. Understanding the continuous collaboration via team and company networks as the basic knowledge creating process this continuous filling of the corporate knowledge assets is even an important extension of Ovum’s framework which did not lay too much emphasis on knowledge creation and rather describes a comprehensive information sharing and reuse approach.

On the other hand, awareness creation in workpackage 9 as well as the ongoing use of Knowledger 2.0 for the Know-Net project coordination and collaboration strongly pointed out that powerful knowledge and information organization, search, and distribution facilities, as provided by the product-view on KM is a seriously missing link for an appropriate exploitation of the knowledge and information stored in the IT systems. Also the restrictedness to the specific Knowledger databases and formats was regarded a considerable drawback.





Best PracticeSystem

Product Development

Project Spaces

Knowledge Portal

Discovery Services Collaboration Service

Knowledge Map




KM mega processes

NavigationInterface


Lotus Notes Infrastructure

R & D DevelopPeople

Serve Sell

Knowledger Navigator


Know-ledgerNotes Data-bases

Figure 15: Knowledge Associates’ Process View Compared to Ovum’s KM Architecture


To sum up, requirements engineering gave substantial reasons to believe that the intended integration endeavor will make sense for all of the three users which in turn gives rise for the assumption that this way, we can cover the requirements of a whole range of knowledge-intensive businesses and companies.

Detailed discussions between the three developers as well as with some of the other Know-Net partners have already begun and led to a rough integration architecture. This rough architecture is currently being consolidated with respect to technical feasibility what directly leads to the work in workpackage 3 about tool design.

Figure 17 gives a rough idea of what we are currently investigating. On the client side, search applets similar to the Karat user interface (Tschaitschian, Abecker et al. 1997; Abecker, Aitken et al. 1998) provide comfortable access to the product-centred services. Internally, queries are processed by a community of agents, coordinated by a central coordinator, all of them having access to a central set of ontologies. The basic agent platform and ontology representation mechanisms were developed in the DFKI KnowMore project (Abecker, Bernardi et al. 1998 & Abecker, Bernardi et al. 1999) and are currently being tested in a prototype development for Deutsche Telekom AG. Ontology storage could be done with Lotus Notes. The Knowledger still acts as the main


Search & Browse Interface, Agents

Corporate taxonomy

Information management


IntranetInfrastructure


Best PracticeSystem

Product Development

Project Spaces

Knowledge Portal


Knowledge Map


WP DBMS EDM E-Mail WWW WP



Knowledge Portal

Discovery ServicesColl. Serv.

Knowledge Map



E-Mail, file servers, Internet/intranet

Figure 16: DFKI / FHBB Product View Compared with Ovum’s Generic KM Architecture


support for filling a structured set of databases with high-quality knowledge about the company’s important business and knowledge processes. The Knowledger will be extended by additional collaboration services to enhance the knowledge sharing and creation facilities. Additional knowledge sources will be made available by the several serach and retrieval agents provided at the knowledge server side.




..

.

Domino Server + Knowledger

http demonDB’s

Ontology

DB XMLDB

WWWsearchengine

Domino wrapper

Ontologywrapper

DB wrapper XML-DBwrapper

WWW ...wrapper

internalontologyrepresentation

CoordinatorAgent

A1

A2

An

search +retrieval specialistagents

GUI / Appletwrapper

searchapplet

WWWbroser

..

.

searchapplet

WWWbroser

Figure 17: Planned Know-Net Tool Architecture

KnowMore Agent Platform


10 Summary

We shortly reviewed the results of requirements engineering in Know-Net. More details about concrete user requirements can be found in the appendices.

A first outcome of workpackage 2 was the requirements elicitation questionnaire which together with the Know-Net holistic framework can further be developed to an even more comfortable and better structured tool for requirements acquisition when designing IT support for KM.

The next results were the unbiased use cases of the users which can serve as a baseline for the evaluation of our first Know-Net software protoype.

The very requirements resulting from workpackage 2 are to a large extent very specific details (like content of a corporate taxonomy, or meta data required by a certain user) which will go into the specific characteristics of tool development and tool deployment at the user sites. At a more general level, users and developers together defined the services and functionalities to be provided by a comprehensive KM tool, rated the importance of these services and assessed whether and how those services could be implemented in the Know-Net project.

It turned out that there are excellent chances for developing a really useful tool addressing many of the specific needs of our users and bringing some real value-added by the novel integration of process-oriented aspects of KM and product-oriented aspects of KM in a single tool.

We are already in the process of checking integration ideas for their feasibility and will probably make good progress with tool design in workpackage 3. It seems to be appropriate to have a close look at third-party software, in particular for dealing with numerous types of document formats.



Appendix A: Requirements Acquisition Questionnaire

How to use this questionnaire? Some short remarks:

I think, you should go into detail here only if you have already written your first version of the unbiased use case, because elsewhere it is not unbiased. Afterwards, you can come back, think about the use case again and find out whether your point of view has been changed by these questions.

Do not try to answer each and every of these questions. The goal is to have some more structure in your requirements description (i.e., it should be organized according to the first level and second level headings of this document) and some more detail.

A.1. Task / Activity Level

The goal of this set of questions is to find out what knowledge-intensive object-level work processes are currently performed. What are the knowledge-centered characteristics of these work processes? What should knowledge management activities focus on? What kind of KM support could be appropriate?

Which knowledge-intensive tasks do you perform personally, or in your group? Why do you think these tasks are knowledge-intensive?

How often do you perform the respective tasks / activities / processes? How long do they take? What about the process duration and other time related aspects (e.g., is it possible that it is not clear when a certain business process is finished, or that a finished process must be considered again after some time)?

How important are the respective tasks / activities / processes / process steps? Which ones are „mission-critical“?

What improvement potential is there in the different tasks and processes? Which things could be done better? Could be done easier? Could be done safer? Are often done wrong, insufficient, or inefficient? Are there obvious errors because certain documents, rules, information sources, old experiences, experienced colleagues, etc. were not regarded?

How repetitive are the tasks under consideration? Are there central working steps which are performed in a similar way quite often? Could formal business process models be built? Or at least more or less structured task lists saying what things must be done, and partly, in which order? Would such a structure be useful (e.g., for tracing open business processes)?

How „knowledge-intensive“ are the tasks under consideration? Why? Do they require much personal knowledge? Tacit knowledge, experience, social or communication skills, factual expert knowledge? Is the processing of internal or external knowledge sources (persons, databases, documents, archives, ...)



important and central? Is it useful to have access to knowledge, artifacts, or documentation from older, similar business cases in order to reuse or adapt older decisions and experience?

How „document-intensive“ are the tasks under consideration? Why? Are many paper-based or electronic inputs needed or outputs created? Are formal documents in use to support the business process, communication, or documentation? Is it usual to have personal archives, memos, notes, annotations to perform and document the process?

How „communication-oriented“ are the tasks under consideration? Why? How many people do work together? Where are they working? How often do they communicate? Always the same people? Do they know together? Why is communication required? Via which media do they communicate? Would it be possible and useful to have an archive of such communication?

How „collaboration-oriented“ are the tasks under consideration? Why? How many people with which specific expertise and competences do collaborate? How important are the different contributions for the end result? Is the collaboration process in some way structured or organized? Are there specific ways for collaboration required or would be useful, like shared document editing, shared design editing, shared artifact repositories, video conferences, telephone conferences, live meetings?

What about change management? How stable are the processes, their input/output, constraints (market situation, competitors, legal restrictions, suppliers), results? How are changes noticed and what effects do they have on the tasks and the knowledge required? Are there specific knowledge and information sources to be constantly watched for noticing changes? Are there specific procedures to deal with change?

Would it be possible to draw a picture of the tasks in quest linking together tasks to be done, people which perform them, documents created and consumed, and knowledge required, used, and generated/changed?

A.2. Document Level

The goal of these questions is to get a clear idea of what explicit knowledge and information sources have to be considered and how they are related.

Which documents and explicit knowledge sources occur in the processes under consideration? Which media? Already electronic ones? Where stored? Which formats? How many? How large?

Are there paper-based documents or artifacts which would be better in electronic form?

What relationships are there between documents? Are there discourse structures which should/could be represented (e.g., links between documents, hyperlinks from within a text, several documents have to be seen in a certain sequential order,



several documents are answers to other documents, one document is a later version of another one, several documents were the input to build another one, certain documents ask or answer questions, give pros or cons for certain decisions, several documents together form a larger/more abstract aggregate-like a case, etc.)? How important is it to represent such relationships? Is there a fixed set of such relationships? If not, could it be that this set can be fixed if one thinks about it? Or are there such dynamic processes that arbitrary links must be possible which can be annotated with comments?

Is it usual / would it be useful to have some personal/generally accessible/shared annotation mechanism? I.e., a document is posted and parts can be marked and marginal notes be attached to these passages. Would it also be required to annotate annotations etc.?

A.3. Knowledge Management Level

The goal of this list of issues is to find out detailed requirements for knowledge management support talking about how knowledge and information sources must be handled and processed. It is organized around some basic categories of knowledge management processes and tries to be as specific as possible.

Knowledge Organization Expressiveness of organization schema: if you would have to your disposal

arbitrary structuring mechanisms for organizing your documents, artifacts, and thoughts, what would be an ideal organization schema? Further question: can one single, fixed schema be determined, or must the number of categories continuously be extended / changed? By whom? Or, will also the structure of the organization schema be changed (delete categories, mix two categories together, build one new from several older ones, split one category in several new ones, insert a category at another place in the structure, ...)?

- a list: an ordered or unordered set of categories, maybe with some synonyms

- one hierarchy: a tree (like a filebrowser) or a directed-acyclic-graph (liks are directed, one node could have several parents)

- multiple hierarchies: a number of trees, one item can be classified at the same time with respect to several (or all? of them)

- networks: an arbitrary semantic network with nodes and named links where documents can be attached to nodes (or also to links)

- categories + links: a special case of an arbitrary network: given are some categories and a fixed set of links with special semantics, like in a formalism for describing argumentation structures: categories are questions, arguments, decisions, links connect pro and contra arguments with questions, and decisions with questions, or decisions with follow-up decisions. Another variant of categories and links: basically, documents are characterized by one category in each of a number of hierarchies (e.g., business=transportation,



process=sales, area=south-west-california) plus some links to other documents (e.g., the document with an executive summary of a consulting project for improving sales in a south-west-californian transportation company is directly linked to a document about a partner who was also involved in this project, or a method or tool which was used).

- fuzzy categorization: is it possible that attaching one document to a given category cannot be definitely made, but „more or less“, maybe a vague keyword attachment characterized by some quantitative value?

Contents of organization schema

What standard meta attributes are needed in any case to describe a document (author, creation date, ...)? Which concrete document attributes are useful or necessary for access and processing in your case? How would you serach typically?

Are there reusable (parts) of ontologies? I.e. generally applicable categories for document organization? This means somethings like KNOWLEDGER’s view of the world: all companies can organize their documents in the sections sales, services, ... Or, O’Leary’s review of consulting companies‘ KM systems: all of them had similar structures.

More specific: what categories occur in your case? What topics etc. Which document types (presentations, memos, ...).

What do you think is more appropriate for search and retrieval? "deep" (detailed, complex structures) vs. "superficial" (few basic categories ) organization models?

What about personal views? What about access rights? Do we need authentification mechanisms? How important is privacy? Do users hold their private documents? Is it also possible that users have private comments on public documents?

Who organizes the knowledge base (i.e., determines and publishes categories, categorizes documents) ? When?

Knowledge Search and Retrieval What about the typical usage of the archive system? Browsing about access

structures or searching via keywords etc.? Is there active support possible? Are there push or notification mechanisms possible? How would be specified what to push when to whom?

What about typical users? Are there classes of them, determined by the same job, department, interests, education, customers, etc.? Could they be characterized so exactly that this would help to determine their information needs, or at least to restrict (or expand) their search in the archive automatically? Are there any individual or group profiles of interest? How would they look like in practice (be as specific as possible, give examples).

What about typical usage problems and failures in your personal or group knowledge handling? Things are there but cannot be found? Things are there, but nobody is aware of it? Outdated versions are used? Confidential things go out?



Errors, because rules were not considered? Reuse possible, but not done because people do not want to search? Be specific, give examples! Are there some organizational repair mechanisms for such situations? Based on actual problems in information logistics and archiving, search, and retrieval, what would be the wished situation? How compares this to the situation in use?

Knowledge Creation and Acquisition Who fills the knowledge base? Who categorizes? Is it possible that categorizations

are not seen in the same way by all members of the group? What about „forgetting“ of documents?

Could things automatically be acquired, or at least their acquisition be supported / promoted? For instance, consider the automatic filtering of news groups, or notification mechanisms which watch certain document input streams, business processes, document pools, user actions? Are there routine publishing or document routing tasks which could be automated or at least reminded by an software agent? Is it possible that automatic Internet search webbots would find useful information? How could their search task be specified? How would a human searcher act in order to fulfill this task? Is it possible that software agents could automatically analyse certain information sources and extract interesting facts (which would be recognized by certain syntactic language pattern), like Price Waterhouses EdgarScan software does?

Can automatic filling / searching / filtering agents, e.g., analysing the Internet, or personal Notes databases, or documents, be envisioned?

Knowledge Maintenance and Change What about maintenance? How often how important, how drastic changes required?

Driven by which forces? Does change not only concern the documents, but also the organization structures?

Are there different (other words) or orthogonal (additional dimensions) / deeply different (completely other views) structures and vocabularies across departments or between user groups or between different users etc.? This concerns both words IN documents (if I do a fulltext search) and words ABOUT documents (names of categories).

Where come classifications from? User, knowledge manager? Could they be automatically attached? Or, are there meta data which can automatically be derived, e.g., when a document is created and the author is concerned with a certain business case or a certain tool.

What about outdated knowledge? Is there any? How is it detected? How marked or deleted? Is there a „life-cycle“ for certain kinds of documents such that it is clear when they are outdated (e.g., new version, fixed point of time, ...)?

What about identification and deletion of wrong knowledge? Who identifies? How? Who decides about deletion?



A.4. Technical Requirements

Are there already existing information systems? What is their functionality? What about known drawbacks, problems, or advantages and strengths? Are they still needed or should they be replaced? Is interoperability required with new or other system parts?

Which hardware / software environment must be taken into account?

What about efficiency and performance (number of accesses, response time) and size and number of documents?



Appendix B: PLANET’s Requirements

“Unbiased” Use Case

The “ideal” KM system for PLANET would normally include support for at least the following issues:

1. an environment for individual, as well as organisational, competence development.

2. an electronic space for supporting work during the definition, planning and implementation of consulting assignments, i.e. what is normally called within PLANET: the “service delivery process”.

3. an electronic space for supporting bid preparation;

4. an electronic space for supporting the less structured work during business development and client management.

The four above mentioned issues can be considered as the most critical for PLANET’s case. They do not cover the whole range of activities but put emphasis on the core processes of the company.

Requirements for Supporting Personal and Organisational Competence Development This part of the ideal KM system would normally provide a space where three “types” of users can work:

the individual consultants concerning their own competence development

the project directors and project managers when planning the allocation of staff to a new consulting assignment, since they have to find “who knows what” within the company and make the appropriate staffing decisions (given of course the resource availability of consultants)

the consultants involved in the human resource development function; within PLANET each consultant is related to an advisor (i.e. another consultant) who acts as a coach and facilitator / discussant of the consultant’s development and career path within the company.

the company’s directors, who need some kind of assistance when formulating and monitoring the overall business plan of the company, with respect to existing and future planned competences at the organisational level, possible mismatches between the current organisational competences and the ones required to implement alternative future business development routes, etc.



The system would provide the following facilities for the individual consultants:

map their competences to a pre-defined structure of company-wide competences

have the possibility of “grading” their competences using some predefined criteria that are applicable across the company

have a space where they can make annotations, descriptions and comments to their competences

identify these competences for which they need further work

somehow “link” their former and current consulting assignments, training seminars and learning experiences to their competences - in order to justify the grade they allocate to the competence

provide facilities for on-line availability of company training material, as well as information about other possibilities for enhancing their competences (e.g. from conferences, seminars, etc)

develop personal plans for further development of their competences.

The system would provide the following facilities for the project directors and project managers in order to support their staffing decisions:

provide facilities for finding consultants based on searches for competences; the result of the search would ideally provide information about the current and previous assignments the consultant has been involved, the training seminars he/she has attended, his/her personal development plan for enhancing competences, his/her preferences, etc.

provide information about the current and foreseeable availability of each consultant (hence the system would be normally somehow “linked” to such company systems such as the time-sheet system and the information coming from time and resource management of consulting assignments

The system would provide the following facilities for supporting the human resource management functions of the company:

an electronic space where the advisor and the “advisee“ could collaborate in an unstructured way, i.e. arrange for meetings, discuss on-line but not real-time concerning the consultant’s development plan, monitor in a periodic manner the agreements they have reached concerning the consultant’s core competences and critical skills, make changes and suggestions for revising the personal development plan, etc.

at the business-unit and company level (there are two business units in PLANET), support is provided for monitoring the overall current status of competences and skills and the expected organisational competence “map” - based on the plans of each individual consultant.


Requirements for individual competence development

Requirements for project directors

Requirements for human resource management


at the company level support is provided for scheduling and monitoring the training requirements, providing on-line availability of the training material (mainly documents and presentations), identifying requirements for further training, etc.

The system would provide the following facilities for supporting company-level directors:

link targets from business plans and foreseeable consulting assignments to the competence map of the company and be able to even make changes and modifications to the overall competence map, in order to define the “future” company-wide competence map

evaluate the missing competences of current business-unit competence maps and perform organisational “gap analysis” between the current state of competences and future requirements (based on the results of the business plan).

NOTES:

1. Currently (February 1999) PLANET has defined the company-wide competences and the grading system, it has developed a three-year business plan that also maps future directions to company competences, and is in the process of rolling out the human resource management implementation process (i.e. defining the advisor per consultant and starting the “coaching” sessions). However, no IT system has been developed. A number of MS-Excel spreadsheets are used for mapping individual competences - no company-wide system is available for overall competence tracking and monitoring.

1. Some (certainly not all) of the above functionalities are already provided by the Personal Development Planner (PDP) module of Knowledger. Since the structure and concepts behind PDP match extremely good the related business processes of PLANET, it would be certainly interesting to examine how the PDP could be further enhanced to cover all additionally required functionalities. Further description of how PDP is linked to PLANET’s requirements is given in deliverable “D9.1 Report on Awareness Creation at User Companies”.

Requirements for Supporting Service Delivery (Consulting Assignments)An “ideal“ KM system would provide support for the following phases of consulting assignments:

definition and planning of the assignment

implementation of the assignment

co-ordination and monitoring of all assignments

The support for the definition and planning of the assignment process would normally be available both:

when a consultant is preparing - with the support of the Tender Support Unit - a bid for an assignment


Requirements for company directors

Current Approach in PLANET

Current support provided by Knowledger


when a project director or project manager defines and plans an already assigned project (i.e. when a contract already exists with a client)

In the latter case the (accepted) proposal forms (at least) the basis for the contract with the client and outlines the basic elements of the work to be done such as duration, resources to be allocated (i.e. man-months), activities to be implemented, total financial figures, etc.

However, the negotiation with a client would normally change (at least) some of the elements of the initial proposal. In addition, since there is usually a considerable time from the proposal to the project initiation phase, the original people allocated may not be available, etc.

Hence the actual project definition and planning task is much more detailed and “restrictive” and may lead to results which are significantly different from the ones outlined in the original bid.

The following text focuses on the project definition and planning, while the requirements for supporting proposal (i.e. bid) preparation are examined below in the business development process.

When preparing the initial plan of the project the project director and project manager require support for:

analysing previous company experience for a similar assignment; hence what is needed is a repository facility that includes the project plans (i.e. time and resource plans) for all consulting assignments (PLANET is in the process of establishing such a company-wide system in MS Project). Normally the project director with the help of the project manager would refine the time, activity and resource plan for the new assignment based on the contract with the client and the previous experience within the company (if any). Hence facilities are required to find similar past assignments, locate their project plans, locate the people responsible for such plans, “talking” to these people - e.g. in a collaborative space or on the phone - in order to come up with the new plan.

examining the resource availability of people in order to make staff allocations and define the consulting assignment project team. This would mean defining the required competences for the specific assignment, searching across the people and competence “database” in order to find the consultants that more or less match the needed profiles and checking their current and planned availability.

analysing previous company “knowledge” experience, i.e.

how have similar assignments been carried out by the company in the past?

which were the major problems (if any?)

which were the major results?

which are the “traps” a project director may fall in when defining the plan? etc.


Definition and planning of the assignment


NOTES:

1. The availability of such information, of course, requires a “project repository” that keeps track of the various phases of each assignment, is filled-in by project directors and project managers, its content is organised and classified according to the services provided by PLANET and that there is a culture and discipline such that the project directors and managers periodically update this “repository” during and after the completion of assignments.

2. PLANET has already defined and is in the process of using some systems towards this end: an MS-Project-based system for project time and resource plans and an MS-Access database for “capturing” the basic information of each assignment (such as client, industry, financial information, project director, people involved in the project team, detailed project description, list of deliverables, etc). However, these systems are not tightly coupled and certainly do not form part of one integrated system.

3. The time duration of the definition and planning process is rather small. For example it would normally take two-to-three days for the preparation of a project plan.

During the consulting assignment (whose duration may be as short as one or two months, or as long as one or two years) the “ideal” system would provide the following:

functionalities for the project director and project manager to monitor and update the project plan (i.e. tasks, their duration, resource allocation and usage - i.e. people and man-months allocated per task, etc.)

functionalities, for all people involved in the project team, to facilitate the successful completion of the assignment. Such functionalities would include:

facilities to search, find and retrieve related material from

previous assignments that have been carried out within the company,

material that the company has access to from other consulting firms or

material that can be found on the Internet (e.g. general news, news from consulting companies, business and management sites, industry-specific sites, etc.)

facilities that support the members of the project team to collaborate electronically (since it may be the case that members of the project team are not always physically in the same place), as well as

facilities that support the collaboration and information exchange between the members of the project team and the other consultants of the company (such collaboration is currently supported by phone calls and e-mail).


Implementation of the consulting assignment


NOTE: In order to facilitate the capturing, organisation and further dissemination of project material PLANET has developed the Knowledge Input-Output sub-process (within the “Service Delivery” process), for which each consulting assignment is expected to define:

the “knowledge input” i.e. the knowledge that is considered essential as an input for completing the assignment and can be somehow found within the company (from previous assignments, from training material, from people’s knowledge, etc.)

the “knowledge output” i.e. what information and experience the assignment is expected to result to either at the various intermediate stages and/or at the end of the project.

the “knowledge review” i.e. a number of meetings of the project team with “knowledge reviewers” (people that are considered knowledgeable in the area of the project) who are expected to both: assist the team in the successful completion of the assignment and guarantee that the “knowledge output” is of a certain quality and is more generally applicable and useful for the company.

The process of defining and then monitoring the Knowledge Input-Output is currently based on elementary (paper) forms and managed company-wide by a newly formed “Knowledge Services Group”. Further description of the process is given in deliverable “D9.1 Report on Awareness Creation at User Companies”.

The “ideal” system would normally provide support both:

for the project teams - to input their experiences, new knowledge (or at least pointers to it), search for already existing items across the whole “knowledge base”, retrieve them (if available electronically) or just locate them (if available only in physical form), etc.

for the KM-related organisational roles that are currently developed within PLANET, i.e. the Knowledge Officer and the Knowledge Services Group.

The latter roles would be facilitated to:

track the completion of Knowledge Input-Output tables for all consulting assignments

monitor Knowledge Review meetings and send alarms when they are delayed;

assist consultants in the categorisation and classification of acquired information and knowledge;

help in the identification and classification of new information and knowledge;

keep track of the use of the system (e.g. number of hits, items mostly asked for, etc.)


Current situation in PLANET

Co-ordination and monitoring of all assignments


Requirements for Supporting Business DevelopmentThe ideal system would support business development and client management by providing the following functionalities:

support for people that take care of a specific client’s relationship to PLANET (internally called “account managers”) to keep track of previous, current and future contacts with the specific client

facilities to capture and organise the experience that all PLANET’s consultants may have with each client (such experience may arise both during assignments, i.e. within the service delivery process, and during business development)

facilities to capture, organise and distribute information about new and prospective client contacts, contact information per customer, etc.

Requirements for Supporting Bid PreparationThe ideal system for the bid preparation process (which normally takes place within ten to fifteen days for each bid) would support:

the facilitation of handling such “administrative” and document-intensive work as:

preparing letters of collaboration (e.g. in the case of a consortium);

using company profiles (in Greek or English, in various formats - short or long, etc);

the facilitation of such work as:

locating possible contact points for partners or subcontractors (e.g. from a repository of companies which is organised according to the company’s competences and services provided);

electronically supporting the communication and documents exchanged with possible partners and subcontractors;

identifying the consultants within PLANET whose competences match the ones required from the call for tender and preparing the CVs of the proposed project team;

locating and (partly) reusing material from the documents of previous bids; etc.

Structured Account of Requirements

Task / Activity LevelThe following text focuses mainly on the “Service Delivery” process of PLANET; however, it also includes requirements that are somehow connected to the support of other business processes of the company.



One could consider as “knowledge-intensive” at least all tasks carried out within the project team on a consulting assignment, when the members of the project team (either individually or as a group) communicate, co-operate, brainstorm or actually write assignment reports and client presentations.

Other tasks and activities within the company can also be considered “knowledge-intensive”, e.g. client contacts for discussing the possibility of an assignment, negotiating a contract, etc.

Finally, there tasks whose ““knowledge-intensity” is rather lower (i.e. they are rather more “information-intensive”).

Because they are based on the experience of people, the tacit and/or explicit knowledge they have available, the information and knowledge available either within the company or from outside (other companies, Internet, etc.), the ways tacit knowledge is converted to explicit and transferred to other people(e.g. by communicating - face-to-face, on the phone, by e-mail - by writing reports, etc.), made tacit again, made explicit again when preparing a report for a client, etc.

Continuously:

during consulting assignments with clients,

during the preparation of bids,

during business development with prospective clients.

A consulting assignment may have a duration of one to two months, to one to two years; an average assignment would be something like six months.

When an assignment is finished, normally its completion is associated to some to some formal approval process of the final result; hence practically there is no case for a finished assignment to be considered again after some time.

Extremely important, since they form the heart of the consulting business; hence could be “mission-critical” for the company.

If an assignment “fails”, that would normally mean that the client refuses to accept the deliverables submitted, which will result to delay in the payment, re-work, possibly involvement of additional people, etc, it may even imply the application of penalties, i.e. the decrease of agreed amount of money.

Various areas of improvement exist; the majority of them are associated with somehow supporting the consultants to be more “knowledgeable” in doing the work.

This would imply providing them the most appropriate information/knowledge (either the required item itself or a concrete pointer to it), at the most appropriate time, at the most appropriate form, etc.

When working for an assignment, a bid or when contacting a client, very often consultants “do not know what they know”, in the sense that there may be colleagues,


Which knowledge-intensive tasks do you perform personally, or in your group?

Why do you think these tasks are knowledge-intensive?

How often? How long do they take? How important are the respective tasks? What improvement potential is there? Which things are often done “wrong”?


project deliverables, presentations, books, reports, sites, etc. available and known to somebody else, that may be useful and have the specific knowledge required, without the consultants knowing it.

Some indicative results of something going “wrong” are that:

projects get delayed due to “re-inventing the wheel”;

bids may be over- or under-estimated;

clients may not be appropriately contacted; etc.

The most common “failure” is for projects to “re-invent the wheel” which leads to delays and increase of allocated resources.

Within consulting assignments there are some tasks (e.g. preparing a financial plan or developing an IT architecture) that are found in the provision of the same of similar assignments (e.g. a business plan or an IT strategy plan respectively).

Such tasks may be required for the completion of various assignments e.g. two to ten times a year. Hence the frequency is not high, however, the number of these tasks is significant and (most) can be well-framed into concrete knowledge “objects”.

Although PLANET’s business processes have already been “mapped”, since the overall work is knowledge and people-intensive there is no way that a formal and structured business process or workflow model could be of any use.

On the other hand, more or less structured task lists with the things that must be done, and partly, in which order, could be useful in supporting not the actual (and more creative) task of working on the assignment, but all associated business processes (i.e. planning and managing the assignment, handling the financial issues of the assignment, etc).

As mentioned in the beginning working on the assignment is “knowledge-intensive” because:

it involves both tacit and explicit knowledge,

it includes the associated conversions from one type to another,

it requires communication between people (within the project team, with other consultants, with the client, etc) and hence requires advanced communication skills,

it relies on locating and retrieving factual expert knowledge either from other people or from physical/electronic documents, databases, etc.

it requires knowledge that may be found both internally within PLANET, as well as externally (clients, competitors, news, etc.)


How repetitive are the tasks under consideration?

Could formal business process models be built?

How „knowledge-intensive“ are the tasks? Why? Do they require personal knowledge?


As is obvious from the above it is extremely useful to have access to knowledge, artifacts, or documentation from older, similar business cases in order to reuse / adapt decisions and experience.

The “document-intensity” of the work is rather high, since it involves mainly the preparation of reports, presentations, etc.

The input required for an assignment is either paper-based (e.g. when it is retrieved from the client or when taken from a partner or a book/report), or electronic (when available within the company or from the Internet).

The output created as the result of the assignment is electronic in a document (or presentation, or spreadsheet) form.

Not many- only the ones mentioned in the previous section and in D9.1 (i.e. the forms for the Project Plan, the Knowledge Input-Output, etc).

Of course, there are also the more administrative ones (e.g. the forms for expenses, travel, purchases, etc).

Yes, it is very usual for consultants to have their own private archives, memos, notes, etc. during the implementation of the assignment.

This material, is either not available to others, or available only to the other members of the project team with whom the consultant co-operates (this happens only through the co-operation process and not through any formal requirement).

On the other hand, each consulting assignment and each bid are allocated a unique number and have an electronic directory in the company’s shared disk server. Normally this electronic “space” is the shared repository of the team working on the specific assignment or bid.

However, the organisation of electronic material within this space does not follow a general organisation scheme, which results to major problems when searching for something from another project.

Implementing a consulting assignment is extremely “communication-oriented”, since consultants are expected to:

co-operate and communicate within their project team;

co-operate and communicate with other consultants of the company;

co-operate and communicate with the client.

An average consulting case team may have from three to five team members.


How “document-intensive“ are the tasks under consideration?

Are formal documents used to support the business process?

Is it usual to have personal archives, memos, notes, annotations to perform and document the process?

How „communication-oriented“ are the tasks?

How many people do work together?


However, being a member of a case team does not mean full-time (although there are times when it does). Normally, a consultant is involved in more than one projects (there is also a correlation with seniority - the more senior ones being involved in more projects).

The tendency of (at least the Greek) market the last years is for assignments with long duration (e.g. a year) that require the physical presence of consultants at the clients site.

Currently a percentage of 30-40% of PLANET’s consultants work full-time at clients’ sites.

The communication between project team members is continuous and usually phase-to-phase. Normally such communication will involve the people of the project team, while it may also involves people somehow related to the project (e.g. people doing business development for the same client, people working on similar assignments, etc.

Communication is not restricted within project team members. Informal face-to-face communication between people happens continuously for company (or even friendly) issues, while the company organisational structure (currently based on allocating people to “disciplines”, i.e. groups which share common “thematic” knowledge) foresees regular formal face-to-face meetings (within “disciplines”, meetings for company-related administration and management issues, etc).

Although e-mail is used extensively, the most common communication medium (when not face-to-face) is telephone.

Practically all tasks involved are “collaboration-intensive”, since the outcome of a bid, an assignment, a contract negotiation, etc. involved more than one people, each one contributing his/her part.

The number of people collaborating, as well as their specific expertise and competences depends on the specific assignment.

All different contributions are important for the end result.

Although not required, some ways for enhancing collaboration would be useful. Shared document editing would be useful when preparing assignment reports or bids (currently managed “manually” by dividing parts of documents and then re-assembling the final version).

Video conferences could be useful but seem a luxury at the current stage (e.g. they could be used in training seminars or in projects in which people from Athens collaborate with people from Thessaloniki, Patras or Brussels.

Since the company’s business processes do not generally follow a formal structure, managing their changes does not require any formal approach.


Where are they working?

How often do they communicate?

How „collaboration-oriented“ are the tasks?

Are there specific ways for collaboration required or useful?

How stable are the processes?


Since PLANET’s business processes are cannon be easily formalised and modelled, a picture would rather create misunderstandings than facilitate a mode deep understanding.

However, we include some indicative (UML-style) diagrams of the knowledge-related business process that refers to the preparation and monitoring of the “Knowledge Input-Output Plan” for a consulting assignment”.

Knowledge creation

review : Knowledge Identity : Project Director : Knowledge

Officer1: fill in

2: review

3: assign reviewer(s)

: Knowledge Plan

: Knowledge Review Report

: Knowledge Final Report : Professional : Project

Manager

4: fill in

5: review

6: enhance/modify

7: formulate

8: formulate

9: inspect quality

10: inform

Details of K-I-O


Would it be possible to draw a picture?

Project DirectorKnowledge Officer

Knowledge Input-Output cycle

KnowledgeReviewer


Proffesional

Knowledge Officer Service/Thematic Area Manager

feedback

findProposal and Project archive

Internet

Proposal and Project Database (STEP, MS Access db)

filter

Knowledge Use

Document LevelDocuments that are internally produced (mostly MS-Word documents, MS-Powerpoint presentations and MS-Excel spreadsheets) are either part of a bid or an assignment’s deliverables.

In addition, paper-based documents (reports, books, client information) are gathered for each assignment.

Moreover, other information (e.g. from competitors, from the Internet) may be available either in paper or electronic form.

Most electronic documents are stored in the headquarters file server. All consultants have access to documents stored in the “K” (projects) and “P” (proposals) disks on the server. Each projects or proposal has a unique directory on the server. Typical directories in the K disk are 50MB and include 50 to 100 sub-directories (organised in tree-like hierarchical structures) and 200-800 files. For the P disk the corresponding figures are 10MB, 5 to 30 sub-directories and 100 to 200 files.


Which documents and explicit knowledge sources occur in the processes?

How many? How large?


Documents that are internally produced are stored on this server in almost all cases. Additional material, that is available in electronic format (e.g. Acrobat files, html documents, etc.) and is directly related to a specific project or proposal, is also stored in the server, while material that is not directly related is usually stored on local disks. Non-electronic material is in all cases archived at the individual level.

Yes. A typical example is the Financial Times (or Greek newspapers).

The electronic version of the newspaper could allow for easier searching, linking, pushing, etc.

One can identify such relationships at two levels:

the project (or proposal, or team) level, where relations between documents are rather project (or proposal) -specific (e.g. for proposals, the submitted proposal consists of several documents that deal with specific parts of the call for proposals, such as CVs, company profile, previous experience, etc.). It is relatively important to represent these relationships in order to give an overview of the work done and to provide an information map within the specific assignment.

the organisational level. Here it should be meaningful to represent relationships between more highly ‘refined’ and ‘reviewed’ documents such as methodologies, training seminars, best practices, etc. It is obvious that the organisation and representation of links between such documents, corporate services, people, market, etc. is absolutely imperative.

In broad terms, one can say that documents within the organisational level fall in the first category of fixed relationships (more on this in the next paragraph “Knowledge Organisation”), while documents of the project level fall in the later category of dynamic linking through annotations, etc. Each project or proposal is quite unique, and although one could identify groups of similar projects (or proposals) and try to come up with a relation representation schema for each of these groups, it is doubtful whether there is significant value added to this representation.

Knowledge Management Level

Knowledge Organisation

The organisation schema could be based on the categorisation of the organisation’s knowledge assets. This is shown graphically below:

Practices / Business Areas

They represent the different target market segments of the company. The ‘Public Sector” business area, for example, should


Are there paper-based documents which would be better in electronic form?

What relationships are there between documents?

Is there a fixed set of such relationships?

Expressiveness of organisation schema?


contain knowledge, best practices, etc. about this market segment.

Customer Specific client knowledge (e.g. history of PLANET’s contact with the client, background information) as well as contact information, ‘intelligence’, etc.

Competitor / Collaborators

Similar for competitors / collaborators / partners

Project Knowledge related to specific assignments - see detailed description on Knowledge Plan, section 3.1.

Proposal Knowledge gained / used for the preparation of bids to tenders

People (currently) CV’s of employees and external consultants and partners, organised and linked to other knowledge assets such as projects.

(future) Detailed knowledge “owned” by company consulting staff (who knows what)

Service knowledge about key services that PLANET provides to clients (e.g. project management services). It should include the way these services have been provided in the past and/or should be provided in the future (know-how), learning histories from assignments, methodologies, etc.

Thematic Area Knowledge about key thematic areas that the company has experience in, or wants to develop, in order to incorporate them into services (example: specific knowledge about financial management - thematic area - that is useful in Business Plans - service) .

Table 1 Description of PLANET’s knowledge assets

It seems like a categories + links representation schema is the most appropriate for PLANET’s case, primarily because this combines a structured organisation of artifacts with more arbitrary (loose) relation building capability.

The example given by DFKI is rather representative of a typical case in PLANET.(e.g. the document with an executive summary of a consulting project for improving sales in a south-west-californian transportation company is directly linked to a document about a partner who was also involved in this project, or a method or tool which was used.)

The following could be applicable:

author

date

major sources (if any)

keywords

summary


Can one single, fixed schema be determined?

What standard meta attributes are needed?


Yes. At least the top view is relatively fixed and can be e.g. associated to the company’s business processes.

Indicative categories include: COMPANY, BUSINESS DEVELOPMENT, DELIVER SERVICES, DEVELOP SERVICES, DISCIPLINES, DEVELOP PEOPLE, etc.

Detailed, complex structures, appropriate or not, are certainly not required.

Intuitive, standard, ‘user-friendly’ searching mechanisms are needed.

NOTE:

Top-down navigating is equally required as bottom-up searching mechanisms.

Personal views is one of the ‘nice to have’ things, but is not an absolute necessity. We see it more as personal notification mechanisms (that of course should be customisable to meet the individual requirements).

Enforcing tight security to prevent external intruders from accessing the system is of paramount importance. Therefore the latest authentication mechanisms should be applied.

Within the company, in general everybody should have at least ‘reader’ access rights.

‘Write’ access rights will vary with respect to particular elements of the system.

A more detailed description of the current KM-related organisational structure in the company is described in D91.

This situation should be somehow reflected in the Know-Net system.

The following is taken from D9.1

--------------------------------------------------------------------------------

Knowledge Officer (KO)

Senior management position responsible for the:

supervision and co-ordination of Knowledge Management System (knowledge input, knowledge output);

supervision and co-ordination of service development;

supervision and co-ordination of discipline development;

supervision and co-ordination of training activities;

organisation and administration of library;

organisation and follow-up support of Service Development and Discipline Development meetings.

Knowledge Services Group (KSG)

A team of active consultants who will:


Are there reusable (parts) of ontologies?

What do you think is more appropriate for search and retrieval?

What about personal views?

What about access rights? Who organises the knowledge base? When?


develop, package and maintain PLANET’s KM system;

help ensure PLANET’s intellectual capital and solutions are organised and accessible during engagement pursuits;

assist SDM/TAMs to structure information for input to the KM System;

In the long-term the KSG could develop services that will provide to PLANET’s consultants in order to support them during proposal preparation, etc.

Service Development / Thematic Area Managers

These are active consultants that:

are regarded as PLANET’s subject matter experts (service/thematic area);

are responsible for collecting, storing, updating, and advancing knowledge in their subject matter;

will be responsible for managing related content in future KM system.

--------------------------------------------------------------------------------

Knowledge Search and Retrieval

Top-down navigation and bottom-up searching are required.

Currently this process is carried out with human intervention: senior stuff is asked to provide information and links to people or documents based on their experience and previous engagements.

‘Push’ notification mechanisms could be used both at the project (or team level) and at the organisational level.

In the first case the Project Director or Project Manager could specify the ‘who’ to push information while Consultants of the Project Team could specify the ‘what’ and ‘when’.

At the organisational level a ‘push’ mechanism could be utilised by the Knowledge Officer and the Service Development / Thematic Area Managers to push information to groups that are going to be maintained by them.

We think it is more meaningful to restrict search by providing the functionality to specify discrete parts of the knowledge base and subsequently restrict the search scope to the selected sites, rather than determining the search scope base on user position or interest.

There are no limits between the different “disciplines” and the job responsibilities of each consultant vary from one engagement to the other to an extend that it would not be appropriate to ‘characterise’ them.

There exist some informal communities of interest within the company.


What about the typical usage of the archive system?

Are there push or notification mechanisms possible?

Are there classes of users for search?

Are there any individual or group profiles of interest?


Typically these are small groups of consultants that share the same interests (e.g. project management concepts, Eastern Europe economic affairs, etc.) but all are highly informal.

Consultants store material in electronic format on both personal computers, company file server, and laptops.

Personal document management is really up to the individual. The case of someone not remembering where she/he stored a particular file is typical.

At the project (or team) level main problem is caused by the mobility of consultants. Files are in many cases kept in laptops, and although they are copied to the company server, consistency is by no means enforced.

Documents do get lost in because of this.

Obviously more senior people have a better overview of documents available in the company but up to now there was no formal process or system to support this process.

Currently PLANET is in the phase of applying an IT - system to support this process: an MS-Access database for “capturing” the basic information of each assignment (such as client, industry, financial information, project director, people involved in the project team, detailed project description, list of deliverables, etc)

An ideal situation would be for the KM-system to provide interfaces and access to all internal sources of information and selected external ones.

This implies that it should be for example possible to query the proposal and project database (STEP, developed on MS-Access) from within the KM-system

Knowledge Creation and Acquisition

Again, we could distinguish between more ‘refined’ and reviewed knowledge such as a methodology that should be provided by the a SDM and more general knowledge and information where anyone is free to contribute.

In a particular thematic area for example the TAM should be the one that, having collected and reviewed information from consultants, fills and categorises the knowledge base.

An agent could:

(external)

‘Spy’ selected Web sites (big-five consultants’ web sites, Cordis Web site )and notify when new material is added.

Scan selected, specified sites for specific keywords (standard search)

(internal)

Support the SDM/TAM by scanning existing and newly developed internal documents for information related to service / thematic area.


What about typical usage problems?

What would be the wished situation?

Who fills/categorises the knowledge base?

Could things automatically be acquired?


Support the K-I-O process ???

Automatically generate a ‘what’s new’ page for the additional to the knowledge base (file server, corporate intranet, etc.)

notification mechanisms when a project has been completed or a milestone has been reached, e.g. for monitoring the Knowledge Input-Output Plan, disseminating information to the Service Development Manager, etc

…

For instance, consider the automatic filtering of news groups, or notification mechanisms which watch certain document input streams, business processes, document pools, user actions? Are there routine publishing or document routing tasks which could be automated or at least reminded by an software agent?

Technical RequirementsPLANET’s telecommunications and IT equipment includes :

A LAN for Athens office

Office automation applications: Microsoft Windows 95, MS Word for Windows, MS Excel, MS PowerPoint.

Leased-line connection to Internet.

Access to INTERNET from all the workstations.

(Limited) capability of remote node connection with the company's network through a dial-up telephone line.

Currently there exist or are under-development the following systems:

Human Resource Management and Development System (not yet IT-supported system, focused on the development of people, career management, advising, etc. Also includes definition of appraisal method.)

Project Plan System (MS Project based application for designing and managing projects)

Business Plan System (the company business/master plan)

Business Development System (non-IT, looks at strategic, planning and operating activities of the company)

Projects and Proposals Documentation System (STEP) (for documenting projects and proposals, developed in MS Access, includes project descriptions, contractual info, etc.)

Financial System (IT application, commercial package used)

It should also be mentioned that on a daily - basis, all employees use the Internet and email and that the majority of electronic documents that are being created and used are MS Word, Excel And PowerPoint files. A simple archiving mechanism is being used for filing electronic documents:


ICT infrastructure Support systems


each proposal and each project is being assigned a number

there exist a drive on the file server for proposals (P:) and for projects (K:)

each project/proposal has its own folder on the drive

structure of subfolders is not rigid, but typical subfolders include: Correspondence, Administrative, Work, etc.

Standard Windows search functionality is being utilised.

All company offices should have access to the knowledge base.

Highest possible security should be enforced to prevent external intrusions.

Leased line and dial-up bandwidth limitations should be taken into consideration.

PLANET’s requirements Table

Distribution, Publishing, and Filtering n. appl. nice import. essent.

notification / push mechanisms1. mandatory readings2. changes/modifications3. topic oriented filtered delivery


********

****


*******

*******

****

Search and Retrieval browser interface to KM services basic: keyword and boolean search basic: metadata based search advanced: combine criteria advanced: semantic analysis search over different document types restrict search area

************

*****

********


What about efficiency and performance?


flexible ordering of search results smart pull mechanisms

******


****************

*****

***********


************

***

*****


*****

*****


**************


*****

****



Title Importance1. Competence Development1.1 Map employee competencies to a pre-defined structure of company-wide competencies

e

1.2 Grade competences using predefined criteria d1.3 Provide space for annotations, preferences, descriptions and comments to competences

d

1.4 Develop personal plans for further development of competences. o1.5 Link former and current consulting assignments, training seminars and learning experiences of consultants to their competences

e

1.6 Provide space for company training material and link it to competences e1.7 Search for consultants based on competencies. Results whould include: e

1.7.1 experiene from previous and current assignments, e1.7.2 seminars attented, e1.7.3 personal qualificatoins, d1.7.4 personal developemnt plan, d1.7.5 personal preferences d

1.8 Link to time and resource management system to supply info about current and foresseable availability of consultants

d

1.9 Provide an electronic space where the advisor and the “advisee“ could collaborate in an unstructured way, i.e.

1.9.1 arrange meetings, e1.9.2 discuss on-line but not real-time concerning the consultant’s development plan,

e

1.9.3 monitor periodically the agreements they have reached concerning the consultant’s core competences and critical skills,

d

1.9.4 make changes and suggestions for revising the personal development plan

d

1.10 At the 2 business-units and company level monitor the overall current status of competences and skills and the expected organisational competence “map” - based on the plans of each individual consultant.

d

1.11 At the company level provide support for scheduling and monitoring the training requirements, providing on-line availability of the training material (electronic and printed documents and presentations), identifying requirements for further training, etc.

e

1.12 Link targets from business plans and foreseeable consulting assignments to the competence map of the company and be able to even make changes and modifications to the overall competence map, in order to define the “future” company-wide competence map

d

1.13 Evaluate the missing competences of current business-unit competence maps and perform organisational “gap analysis” between the current state of competences and future requirements (based on the results of the business plan).

d

2. Requirements for Supporting Service Delivery (Consulting Assignments)2.1 Analysing previous company experience for a similar assignment; hence what is needed is a repository facility that includes the project plans (i.e. time and resource plans) for all consulting assignments (PLANET is in the process of establishing such a company-wide system in MS Project). Normally the project director with the help of the project manager would refine the time, activity and resource plan for the new assignment based on the contract with the client and the previous experience within the company (if any). Hence facilities are required to:

2.1.1 link to time and resource management system (MS Project application) d2.1.2 link to project and proposal database (MS Access) e2.1.3 find similar (based in customer, service) past assignments, e2.1.4 locate their project plans, e2.1.5 locate the people responsible for such plans, e



2.1.6 “talking” to these people - e.g. in a collaborative space or on the phone - in order to come up with the new plan.

o

2.1.7 which were the major problems (if any?) d2.1.8 which were the major results? d2.1.9 which are the “traps” a project director may fall in when defining the plan? Etc.

d

2.2 Functionalities for the project director and project manager to monitor and update the project plan (i.e. tasks, their duration, resource allocation and usage - i.e. people and man-months allocated per task, etc.)

o

2.3 Facilities to search, find and retrieve related material during the execution of a project, from:

2.3.1 Previous assignments that have been carried out within the company, e

2.3.2 Material that the company has access to from other consulting firms or d

2.3.3 Material that can be found on the Internet (e.g. general news, news from consulting companies, business and management sites, industry-specific sites, etc.)

d

2.4 Project-level collaboration tools e2.5 Support the Knowledge Input - Output process, as described in ‘ideal use case scenario’ and D9.1

d

2.6 Assist consultants in the classification of new knowledge e2.7 Keep track of the use of the system (e.g. number of hits, most asked items, etc.) e3. Requirements for Supporting Business Development3.1 Space for “account managers” to manage clients’ relationships to PLANET e3.2 Capture and organise the experience that all PLANET’s consultants may have with each client (such experience may arise both during assignments, i.e. within the service delivery process, and during business development)

d

3.3 facilities to capture, organise and distribute information about new and prospective client contacts, contact information per customer, etc.

d

4. Requirements for Supporting Bid Preparation4.1 The facilitation of handling such “administrative” and document-intensive work as:

4.1.2 Preparing letters of collaboration (e.g., in the case of a consortium) d

4.1.3 Using company profiles (in Greek or English, in various formats - short or long, etc);

e

4.1.4 Managing CV’s e

4.2 Locating possible contact points for partners or subcontractors (e.g., from a repository of companies which is organised according to the company’s competences and services provided)

o

4.3 Electronically supporting the communication and documents exchanged with possible partners and subcontractors

o

4.4 Locating and (partly) reusing material from the documents of previous bids e

5. General Requirements

5.1 Security, Integrity and Authenticity (Best Internet possible is sufficient) e

5.2 Interoperability with MS Access and MS Project on specific tasks (to be clearly defined during the application of the software on site)

e



6 Basic KM requirements

6.1 Support for assigning roles with specific rights on documents (e.g., read only, write, update, delete)

e

6.2 Knowledge organisation and re-organisation (ability to modify the ‘knowledge organisation’ or ‘knowledge schema’)

d

7. Document handling and management

7.2 Smooth integration of existing documents. Currently there exists 10-year experience in electronic documents.

7.1.2 At least, the system should provide support for full-text searching within these documents.

e

7.1.2 Better, the system should facilitate the (automated??) transition of links (meta-information, selected documents?) from the old repository on the file server to the new system

o

7.2 Full support for all major documents types (PDFs, HTML, MS Office) e

7.3 Metainformation e

7.4 Flexible access control mechanisms e

7.5 Archiving d

8. Information access

8.1 Search facilities e

3.1.1 Search based on meta-information and full-text e

3.1.2 Ability to index usual document types e

3.1.2 Search in a specified area or the entire knowledge base e

3.1.3 Seamlessly incorporate external sources (e.g., specified WWW sites) e

3.1.4 Order search results according to different criteria o

3.1.5 Preview results (e.g., show summary or abstract) d

3.2 Intuitive Navigation facilities (e.g., site maps, etc.) e

3.3 Improved access to information and knowledge

3.3.1 Mandatory readings d

3.3.2 Push mechanisms available to selected users and to specified audience of which the originating user has control.

e

3.3.3 new or modified documents e

9. Project Management and support for team work (as described in Requirements for service delivery, e.g. support for the K-I-O process

d

10. Better support for discussions o



11. Ontologies/Thesauri and Classification

11.1 Automatically generated content description d

12. Tools for the knowledge manager

12.1 Access mining (as discussed in Requirements for Supporting Service Delivery) e

12.2 Document and linik maintenance tools d

12.3 Retrievability of documents o



Appendix C: GOOCH WEBSTER’s Requirements

“Unbiased” Use Case

The “ideal” KM system for Gooch Webster would normally include support so as to satisfy the goals of the company. These goals are:

To find who knows what or who

To touch and use knowledge or information only once. This means that the usual mistake of reinventing the wheel could be avoided.

To leverage the existing processes and the already installed software.

To record and make available throughout the company unique knowledge and best practices.

To ensure that the company knowledge base is available to everyone in the company.

To provide a system that is simple, intuitive and which offers immediate benefit/ return.

Since Gooch Webster is a recently merged company, which will definitely grow in the years to follow, it would be useful to require that the system is scalable.

One of the first and key issues to any merger and to any company is know its capabilities and to know what it knows and more importantly to be able to find it. In creating any knowledge management system, there is a need to create a “killer application”. This is one which everyone can use and immediately derive benefit. For example, one such killer application is the Central Directory. This is a Lotus Notes database listing all of the personnel in the firm and their contact numbers - effectively a glorified telephone list. This system was to have been developed to include photographs, CVs, details of training courses attended and areas of special expertise – the generation of core competencies. Such basic information can be used on a daily basis for profit and enhanced service provision to clients.

A further source of assistance is the Champions Database which is a Lotus Notes library. This aims to list people and their particular expertise. Both the Champions Database and the Central Directory need further development and to be linked to the Personnel Database system before they will truly achieve results.


The system should provide support for the goals of the company.

GW need a “killer” application that will offer immediate benefits.


At the moment, there is a need for a reliable and consistent system, which would support the existing software and would work in the background. Too much overhead would definitely not be liked by users unless they could see immediate benefit.

Gooch Webster had to invest and needs to invest heavily in new IT infra-structure and training of people. Efforts are also continuing to redesign the information systems in the organisation so that all the people share a common platform. At the moment, all the offices are interconnected, using 64-bit kilostream lines. There is a mixture of Novel and NT servers at each office and some offices, mostly where people in the Asset Management group are located, have a UNIX network. They have Lotus Domino servers at every office, so that all people can use Lotus Notes but access to the Internet is still not universally accomplished, due to the fact that 25% of the employees still use Windows 3.1.1. The company still has differing versions of the main Microsoft desktop software which can does inhibit the sharing of knowledge and data.

An “ideal” KM system should provide support throughout all the phases of a project, undertaken by a surveyor.

When preparing for the carrying out of an assignment, the surveyor requires support for:

Analysing previous company experience for a similar assignment: hence what is needed is a repository facility that will include information about a particular property, a particular client, the type of the previous assignment (presentation, advice, valuation), and the connections relevant to this project. Also, facilities are required to locate the people responsible for similar projects, “talking” to these people – e.g. in a collaborative space or on the phone – in order to come up with as much information as possible.

Examining the resource availability of people in order to make staff allocations and define the assignment team of surveyors. This would mean defining the required competencies for the specific assignment, searching across the people and “competence” database in order to find the surveyors that more or less are suitable for the specific project.

During the assignment, the “ideal” system would provide the following:

facilities to search, find and retrieve related material from:

previous assignments that have been carried out within the company

material that can be found on the Internet and the databases (e.g. property news, FOCUS database, client contacts, etc.)

facilities that support the members of the team to collaborate electronically (since there are cases where the members are not physically in the same place. After all, Gooch Webster consists of ten regional offices), as well as

facilities that support the collaboration and information exchange between the surveyors within the company.


The current infra-structure of the company.

The system has to provide support for the surveyor.

The system should provide support throughout the assignment


The “ideal” system should provide the mechanisms to support and simplify the everyday duties of the surveyor. That would include:

monitor the process of the project and send alarms when assignments are delayed;

assist surveyors in the categorisation and classification of acquired information and knowledge;

help in the identification and classification of new information and knowledge;

keep track of the use of the system (e.g. number of hits, items mostly asked for, etc.)l;

learn from user interaction and offer suggestions to create short-paths, hence simplifying the process of search;

alert when work is being duplicated.

The “ideal” system would also support business development and client management by providing the following mechanisms:

support for surveyors to keep track of previous, current and future contacts with a specific client;

facilities to capture and organise the experience that Gooch Webster’s surveyors may have with each client and with each property;

facilities to capture, organise and distribute information about new and prospective client contacts, contact information per customer, new properties on the market, etc.

NOTE: The ideal system should always work in the background. Seamless collaboration with existing repositories and processes is more than essential. What the surveyor does in his/her daily work should be recorded and re-used without any (or very little) extra author input.

We think that the system should be flexible to permit changes and by that we mean that the system should allow re-mapping of links and tags. After all, Gooch Webster is a continuously growing company, in which the most important knowledge assets – the employees – are constantly “moving”. People leave, teams re-organise, and culture evolves. The ideal system has to be dynamic to cope with these changes.

Last but not least, we think that the “ideal” system should not impose organisation – the diary parable. Surveyors are currently using paper diaries, PDA’s, etc. but they will add to an electronic system to the benefit of all. They may even choose to use only the electronic version if the system permits users flexibility to work in a way that suits them. It should also permit their work to be added. Gooch Webster is an organisation of predators that work mostly alone but are more successful as a pack. The system must allow them to come together in their own way, as far as possible.

On the whole, the reader will realise that the system requirements for Gooch Webster (included in the unbiased case) do not differ significantly from the requirements included


Everyday duties should be simplified through the system

The system provides mechanisms for business development

A dynamic system is required

The system should not impose organisation


in the following paragraphs – actually they are completed by them. Gooch Webster have already been using Lotus Notes as a means of managing knowledge within the company. Therefore, the requirements are mostly based on building a system that would support the existing processes and would act as a helpful tool for even the most technophobic surveyor.

Structured Account of Requirements

Task/Activity Level

The following text focuses on the requirements that are somehow connected to the support of all the business processes of the company.

One could consider as “knowledge-intensive” at least all tasks that are carried out within a surveyor’s project, when he/she communicates, co-operates with other surveyors or actually writes reports and client presentations.

Other tasks and activities within the company that can also be considered “knowledge-intensive” are, for e.g. client contacts for discussing the possibility of an assignment, negotiating a contract.

Because they are mainly based on the experience of people, the tacit and/or explicit knowledge they have available, the information and knowledge available either within the company or from outside (Internet, off-shelf databases, etc.), the ways tacit knowledge is converted to explicit and transferred to other people by means of face-to-face communication, e-mails, phone calls, written reports, etc., made tacit again, made explicit again when preparing a presentation or a report for a client and so on.

These tasks are performed continuously during the work of the surveyor, when for example he/she is consulting clients, finding information about a certain property, or writing a report about a specific project he/she has undertaken.

A surveyor works on deals, assignments and projects. These may be as short as an hour – to discover a deal, an idea, a disposal or whether an advice will work or not – or they may take years, where on-going relationships are built when managing properties for a client.

Extremely important, since they form the heart of the surveying business; hence they are “mission-critical” for Gooch Webster. A possible failure in an assignment – a non-satisfactory or delayed retrieval of information about a property, or a general failure in providing the right service for a client – may result in the decrease of agreed amount of money, re-work on a contract, delay in payment or, in worst cases, in losing the client.


Which knowledge-intensive tasks do you perform personally, or in your group?

Why do you think these tasks are knowledge-intensive?

How often? How long do they take?How important are the respective tasks?


This in turn produces a greater need to share knowledge and experience in a multi-cultural and diverse network. The only way to retain clients, to build the business and to increase profits is to continuously improve the quality and delivery of the firm’s services

There is a wide potential of improvement in the way tasks are run so far. The effort is mainly based on supporting the surveyors to be more “knowledgeable” in doing their work.

This would imply offering them the means to have access to the appropriate information for their work, at the most appropriate time, at the most appropriate form, etc. Also, providing them with a powerful tool to search the “know-who” within the company. Managing time effectively is important for a surveyor.

As it happens in most cases of companies that work on a constant interaction with people, Gooch Webster do not have an “Intelligent know-who”. When working for an assignment, or when contacting a client, surveyors very often “do not know what they know”, in the sense that there may be colleagues, project reports, presentations, journals, sites, etc. available and known to someone else, that may be useful, without the surveyors knowing it.

Some results of this wrong management of knowledge are:

Projects get delayed due to “re-inventing the wheel”;

Clients may not be appropriately contacted, etc.

The mere existence of knowledge somewhere in the organisation is of little benefit. It becomes a valuable corporate asset only if it is accessible. Its value increases with the level of accessibility. This is something that the people at Gooch Webster are gradually realising and something they are trying to work on by building their IT infrastructure and a common culture.

There are some tasks that are repeated within an assignment of a surveyor, i.e. preparing a contract for a client or writing a report for a project. These tasks follow a specific structure according to the Quality Assurance Manual.

As we have already mentioned, working on an assignment is “knowledge-intensive” because:

it involves both tacit and explicit knowledge,

it includes the associated conversions from one type to another

it requires communication between people (with other surveyors, with the client, etc.) and therefore, requires advanced communication skills,

it relies on locating and retrieving knowledge either from other people or from electronic documents, databases, Internet sites, etc.


What improvement potential is there?

Which things are often done wrong?

How repetitive are the tasks under consideration?

How “knowledge-intensive” are the tasks? Why? Do they require personal knowledge?


it requires knowledge that may be found within the different regional offices, as well as externally (clients, contacts, news, etc.)

The “document-intensity” of the work is rather high, since it involves the preparation of reports, presentations and the storage of information about properties.

These documents can be either paper-based, or electronic (when available within the company or from the databases and the Internet).

They need to follow a Quality Management System, which consists of the Quality Manual, the Quality Procedures, and the Practice Notes.

The Quality Manual is a control document, issued on the company’s Lotus Notes for use as a training and guidance document in-house. It may also be run off in hard copy as a marketing document, for presentation to clients. Quality Procedures fall into two distinct categories; Professional procedures, which identify the specific working practices undertaken for a client and Administration procedures: which are applicable across the company, and control the administration of the business, handling instructions, the support functions, and administering the management system itself.

Finally, we should mention that all the instructions accepted by the firm are identified with a unique client or property file reference and will be allocated with an individual number for accounting purposes. The nature and the content of the professional service, which they provide varies considerably, but in, all cases, instructions are carried out in a controlled manner, and to an agreed framework. The method of working and the progress on the instruction are documented as procedures. All staff have access to these and to guidelines, work instructions, professional codes of practice and other information as required.

Yes, it is very usual for surveyors to have their own private archives, memos, notes, etc. during the undertaking of a project.

This material might not be available to others. Especially at the beginning, the upper echelons of the firm were more technophobic and resisted the use of the new technology. The result of this was a fear of the widespread availability of the firm’s jealously guarded corporate knowledge and data. Client contacts, accounting and billing information were held in a multiplicity of systems, the majority of which were on paper and under lock and key.

Gradually the culture of the newly-formed company is changing.

After the merger, the two sides have agreed on putting all the projects down on to the file server, they call it T drive and although each department has their own subdirectory, they all use this file server for their proposals. There is now no formal naming


How “document-intensive” are the tasks under consideration?

Are formal documents used to support the business process?

Is it usual to have personal archives, memos, notes, annotations to perform and document the process?


convention for documents, except that the name is being placed at the bottom of the document.

However, the organisation of electronic material is not sufficient and this results to major problems when searching for something from another project.

Implementing a surveying assignment is extremely “communication-oriented”, since surveyors are expected to:

co-operate and communicate with other surveyors

co-operate and communicate with the client

At Gooch & Wagstaff they concentrated very strongly on the team ethos. After the merger, the team ‘spirit’ disappeared significantly, although in certain parts of the firm it is beginning to flourish once again. Teams are assembled for particular clients and for particular projects. The number of participants varies according to the project.

The team networking is also supported by the sharing of ideas and information via the IT system. They all communicate (network) to a degree using the contacts, both internally and externally that are necessary to achieve objectives.

The communication between surveyors is continuous. Each division has developed some effective knowledge sharing processes such as a department meeting once a week. The primary agenda of such meetings is to exchange and swap information about deals that are happening or are about to happen in the market. The instigators of such meetings extend an invitation to people from other divisions to attend with the aim of cross-selling ideas and information.

Although e-mail is used extensively, as well as telephone (which is the most common medium of communication), the majority of the surveyors strongly advocate knowledge transfer through face-to-face meetings and through narratives in addition to more structured forms. The signals that convince people they can communicate effectively are best given in person. Knowledge is more often than not acquired or shared over lunch, dinner or in a pub while sharing a drink with other agents and/or colleagues.

There are some ways of enhancing collaboration. For example, the electronic space of the system should be capable of taking direct postings of multiple forms of knowledge – spreadsheets, texts, pictures, notes reminders and even ‘yellow post-it notes’

Video conferences could be useful but seem a luxury at the current stage, although we would try virtual space for projects – like team room – like the Know-Net use of Knowledger. At the moment, discussion databases are exclusively used. Gooch Webster would be interested to know what the possible overhead for a virtual space would be – say five video conference feeds onto an electronic space or six way keyboard discussion on a posted tender document with live interaction.


How “communication-oriented” are the tasks?

How many people do work together?

How often do they communicate?

Are there specific ways for collaboration required or useful?


Document Level

Documents that are internally produced (mostly MS-Word documents, MS-Powerpoint presentations and MS-Excel spreadsheets) are either part of a contract or an assignment’s reports.

In addition, paper-based documents (reports, books, journals, client and property information) are used through the implementation of the project.

Moreover, other information (e.g. from the Internet and the databases) are available in electronic form.

Gooch Webster has in excess of 350 databases. One of the favourite databases is FOCUS, an external database that provides information on properties. It actually contains all the information found in the top estate agency journals, like Estates Gazette and Property Week. This information is updated twice or thrice a day and through FOCUS it is easily looked up. Despite this, hard copies of the journals are still collected and collated. Although this duplicates the process, it is a reflection of the preference to read hard copy as compared to a screen.

After the merger, the two sides have agreed on putting all the projects down on to the file server, they call it T drive and although each department has their own subdirectory, they all use this file server for their proposals. There is now no formal naming convention for documents, except that the name is being placed at the bottom of the document.

Yes. A typical example is Estates Gazette journal. The electronic version of the journal allows easier searching and linking.

In broad terms, one can say that documents within the organisational level are related somehow through the Quality Management System, which consists of the Quality Manual, the Quality Procedures, and the Practice Notes. It describes the professional services provided by Gooch Webster. It defines the policy and commitment to quality by management, and details the way the company is organised. It also describes the documentation that supports the quality management system.

Knowledge Management Level

Knowledge Organisation


Which documents and explicit knowledge sources occur in the processes?

How many? How large?

Are there paper-based documents which would be better in electronic form?

Is there a fixed set of relationships?


The organisation schema could be based on the categorisation that Planet have presented. This is shown graphically below:

Practices / Business Areas

They represent the different market segments of the company and they comprise the different divisions. For example, the Retail Division should contain knowledge, best practices, etc. about this market segment.

Customer Specific client knowledge (history of Gooch Webster’s contact with the client, background information) as well as contact information, ‘intelligence’, etc.

Project Knowledge related to specific assignments (reports, presentations, etc.)

People (currently) The Central Directory. This is a Lotus Notes database listing all of the personnel in the firm and their contact numbers - effectively a glorified telephone list. This system was to have been developed to include photographs, CVs, details of training courses attended and areas of special expertise – the generation of core competencies.

The Champions Database which is a Lotus Notes library and aims to list people and their particular expertise.

(future) Detailed knowledge “owned” by company staff (who knows what)

Service Knowledge about key services that Gooch Webster provides to clients (Valuation, Rating, Project Management, etc.). It should provide the learning histories from assignments.

Thematic Area Knowledge about key thematic areas that the company has experience in, or wants to develop, in order to incorporate them into services (example: specific knowledge about Investment Property – thematic area – that is useful in the Professional Services department).

Table 1: Description of Gooch Webster’s knowledge assets

It seems like categories + links representation schema is the most appropriate for Gooch Webster’s case, for the same reasons that Planet have underlined.

The following could be applicable:

author

date


Expressiveness of organisation schema?

Can one single, fixed schema be determined?

What standard meta attributes are needed?


version

purpose

description

type (e.g. presentation/ advice/ valuation)

The more appropriate ontologies for search and retrieval are certainly the simple and not detailed structures. The searching mechanisms should be ‘user-friendly‘, intuitive and standard.

Personal views is one of the ‘nice to have’ things, and it would be good to have a personal notification mechanisms (that of course should be customisable to meet the individual requirements). For example, this could include:

An “open” system, i.e. a system with the ability to work with existing legacy systems. It should have “free” space for personal use – user customisable.

Project planning, that is ability to show in pictorial form a project (create milestones) and be reminded when these are due.

A short list compiled intelligently, of new events. So far Lotus Notes shows that there are new postings but these need to be opened and then read. There is a need of a synopsis that is relevant to the user.

Enforcing tight security to prevent external intruders from accessing the system is of paramount importance. Within the company, in general everybody should have at least ‘reader’ access rights. ‘Write’ access rights will vary with respect to the organisational role of the employee.

Ian Heritage holds the heavy duty of organising all this knowledge. He keeps his own databases, his annual report about the company and a special database with all the information sources. He admits that the last mentioned database used to be fairly complete when the firm was a lot smaller, but because of the merger he cannot guarantee how updated this database might be.

One of his tasks is to organise and keep a reminder of all the agents’ publicity, agents’ reports and annual reports. Therefore, he has a catalogue where he files reports according to subject. He records all the new publications that come through him and that includes books, journals, reports, legislation and databases. He also keeps the legislation books, directories and journals.

Gooch Webster have few people who specialise in research. One could possibly call them Thematic Area Managers, when using modern jargon. There is one employee in the Professional Services, who is an expert on IPD – that stands for Investment Property Database – and one in Leeds, who has a research role in Retail. They are both


What do you think is more appropriate ontologies for search and retrieval?

What about personal views?

What about access rights?

Who organises the knowledge base? When?


responsible for collecting, storing, updating, and advancing knowledge in their subject matter. This is very limited for a company of the size of Gooch Webster.

We think that the use of intelligent agents has a huge role to play here so that existing procedures and systems are leveraged intelligently. Diaries, systems and repositories must be watched, scanned and retrieved by agents working in the background, more efficiently than a mere search engine. If an intelligent agent scans an existing process, i.e. one which the knowledge worker does or needs to do as part of his/her daily tasks, then such knowledge is leveraged for the company benefit automatically.

Knowledge Search and Retrieval

Search facility over existing databases – we cannot give up existing Lotus Notes data. Gooch Webster have group calendaring, Lotus Notes libraries for QA (Quality Assurance) best practice, tender documents, and phone list with CVs, therefore such information must be incorporated into the system. Duplication will not work.

The system methodology should work in the background. It is essential that it does not add another process, because this will prove out to be too much overhead and will not be liked by users, unless they can see immediate benefit. For example, if they entered their personal diary details, which were then automatically entered into a group calendar and read by others, then they would be impressed, because people would ring or e-mail them in response. That would instantly mean that their business is done easier and the service quality is better. But above all, that would imply that they would not have to start a project from scratch. We think that a ‘Virtuous Circle’, where normal processes with no extra overhead leverages result, would be essential.

The system must have a very simple interface, which is easy to navigate with as few steps as possible.

The system should allow easy creation of links between relevant documents and future re-organisation. It should be compulsory to create a tag when a document is first received or first created. A pictorial interface will be very helpful.

There should be a search/ repair tool to uncover broken links.

The system should archive data to agreed criteria in order to avoid storage overhead.

At the moment, Lotus Notes shows that there are new postings but these need to be opened and then read. ‘Push’ notification mechanisms could be used both at the project (or team level) and at the organisational level. The ability of the system to show in pictorial form a project and to create milestones would remind the surveyors of deadlines.


What about the typical usage of the archive system?

Are there push or notification mechanisms possible?


Finally, we believe that an intelligently compiled short list of new events and a notification of new postings, relevant to the user, would push information to the individuals.

There are general profiles of interest. There is need for a check for existing knowledge and connections within the firm and this means that, by entering details of a client/ address/ project/ type, the system should prompt to advise of the connection and to inform that there is existing work/ previous work/ current work.

Also, the system should seek intelligently to match and map interests and knowledge. In Gooch Webster, we want to, firstly, find someone who knows or who can do and, secondly, find the actual knowledge, i.e. the last tender, the most recent address for a client, the last instruction carried out in a town, or the presentation with content which matches what we are doing at the moment.

One matter, which is apparent in the merged firm of Gooch Webster, is that it is no longer certain what it has, what it orders and where to find it. There is a common belief that records in information are duplicated. The will still exists to preserve personal libraries, cuttings and sources of data rather than to rely upon either a central resource or an IT based system.

A central filing system is required, common to all offices.

The whole agency system is in disarray at the moment, mostly due to the fact that information, useful to the firm, is being held into two different systems. These two systems need to be amalgamated as soon as possible.

An ideal situation would be to have a system that learns form users and offers suggestions, so as to create short-paths, searches. It should scan internal and external sources, i.e. the Internet and the on-line databases, and monitor and then report back with relevant knowledge/ information to a user. It would be ideal to have an adaptive system, which observes and learns from user interaction and can be programmed by the user to suit his/ her criteria (a sector, a company, a type of product, a location, or a person).

We need to be alerted when work is being duplicated, for example, work for the same client, in the same street, in the same town. The check for conflicts is vital for professionals and a match to alert users in the company to complementary work helps to avoid duplication and to encourage collaboration.

Ideally, the methodology/ system should match and reinforce the existing procedures, e.g. QA and group calendaring. A new process will fail unless there is an immediate win for users, the so-called killer application.


Are there any individual or group profiles of interest?

What about typical usage problems?

What would be the wished situation?


Knowledge Creation and Acquisition

So far, the knowledge base is filled and categorised manually with the help of the librarian. But as we have already mentioned, we think it is of importance that the system learns from users and offers suggestions to create short-paths.

Yes,the use of intelligent agents could play a huge role here so that existing procedures and systems are leveraged intelligently. Diaries, systems and repositories must be watched, scanned and retrieved by agents working in the background, more efficiently than a mere search engine. If an intelligent agent scans an existing process, i.e. one which the knowledge worker does or needs to do as part of his/her daily tasks, then such knowledge is leveraged for the company benefit automatically.

An agent could also:

‘spy’ selected Web sites (for example, other surveyors’ web sites, Estate Gazette’s site) and notify when new material is added;

scan selected, specified sites for specific keywords;

automatically generate a ‘what’s new’ page for the additional to the knowledge base (file server);

notification mechanisms when a project has been completed or a milestone has been reached;

offer the ability to measure use, i.e. the number of hits, the use of a site, and to follow a trail with the aim to put people in touch, to match the relevant interests or projects. We need an “Intelligent know-who” that is active.

Knowledge Maintenance and Change

Measurement must be essential so as to link to appraisal and competency. We must know who is using the system. This is a strategic company goal, linked to measurement of who is using the system, how often and for what. By monitoring the process of search, we will be able to answer questions on whether surveyors are working together in a particular project to understand, to develop, to work in, or to win business. Such measurement should be available for reports – ideally in a graph or a pictorial format. It would be available to a Knowledge Manager/ Director and to Group Heads/ Section Managers. The result of all these will show a return on investment and a business mapping that will answer whether we are pursuing our set goals and if so whether we are pursuing them sufficiently. If not management action can be taken to redress the situation.

We must be able to re-map quickly and easily. Gooch Webster is dynamic, i.e. people join, collaborate and then leave. There must be an automatic re-mapping to permit access to knowledge when authors leave.


Who fills/categorises the knowledge base?

Could things automatically be acquired?

Should we consider mechanisms to rate knowledge usability?

What about maintenance? Driven by which forces?


It must be easy to add and modify documents.

The system should be retroactive, so as to be able to link “old knowledge” into the system.

Tagging should permit actual storage anywhere. The system should not impose a structure for storage. Gooch Webster recognise that the choice is random data storage and an exceptional search agent/ tool or efficient data storage and an adequate search tool. We would prefer to compromise to permit more random storage as this reflects more the way that we work at Gooch Webster.

Finally, the system should intelligently offer choices to the author for the lifetime use – nagware reminder – for example, offer the option of storing a document for three months/ six months, etc., keep/ store indefinitely, delete, archive.

Technical RequirementsThe information mentioned below is an extract from Deliverable 9 (Awareness Creation):

Gooch Webster‘s telecommunications and IT equipment:

After the merger, Gooch Webster had to invest and needs to invest heavily in new IT infra-structure and training of people. Efforts are also continuing to redesign the information systems in the organisation so that all the people share a common platform. At the moment, all the offices are interconnected, using 64-bit kilostream lines. There is a mixture of Novel and NT servers at each office and some offices, mostly where people in the Asset Management group are located, have a UNIX network. They have Lotus Domino servers at every office, so that all people can use Lotus Notes but access to the Internet is still not universally accomplished, due to the fact that 25% of the employees still use Windows 3.1.1. The company still has differing versions of the main Microsoft desktop software which can does inhibit the sharing of knowledge and data.

They are currently going through a Telecoms project with Genesis. The outcome of it will be a combined and integrated Data and Telecoms solution for the company with the dual benefits of better voice and data handling.

In order to elaborate a bit more on the systems that Gooch Webster has been using after the merger, it would be quite useful to see how each of the six groups is run. Starting with Asset Management, we find that the access to the management accounting system is done through a programme system called Tramps. They have a brand new system, called Mantra that is currently being tested and is located at the London office. The whole system is a big database with full and rich details of the properties managed. The Building Consultancy Division use the same system but they


What about outdated knowledge? Who decides about deletion?

ICT infrastructure The Telecoms project.

The systems that the different divisions are using at the moment.


also have access to the Autocad for the digital design of buildings, alterations, projects and schemes. Moving on to Investment, the bulk of the investment team are located at the Grosvenor Street office. They are at the moment concentrating on Q&A databases and they keep their own information in-house. They have recently launched Lotus Notes databases and libraries to store information about market activity. Future plans involve a possible replacement of the Q&A databases. Finally, Retail agency is scattered around the country and they experience quite a few problems in organising their work. There is no central retail requirements database at the moment. Five different systems are in use and people are still keeping their reports on paper within folders. They too have got a Q&A database but they urgently need a comprehensive retail requirement system that can be distributed widely to all of the company’s offices.

Each group has developed its own specialised tools and mechanisms to capture and disseminate knowledge. Prior to the merger, all the databases, that the employees of Gooch & Wagstaff were using, were designed by users in conjunction with their in-house IT team, while in JT&W databases and systems were designed by an external company.

Gooch Webster has in excess of 350 databases. One of the favourite databases is FOCUS, an external database that provides information on properties. It actually contains all the information found in the top estate agency journals, like Estates Gazette and Property Week. This information is updated twice or thrice a day and through FOCUS it is easily looked up.

People are sometimes complaining that there are too many databases to check and quote that they would rather have all the information more centralised, so they would just open one big “box” and everything would be there.

Gooch Webster: Further system development

There are numerous proposals for further developing the existing systems. In no particular order these are:-

Work in Progress (WIP) - a further enhancement to this system has just been tested and the system goes live in February. This system monitors and tracks all work in progress in the firm. It might further be developed to act as a reference point or a logging system for the firm’s files. This system links into the company’s management accounts.

Accounting system - the company’s existing accounting system is not millennium compliant. It will be replaced during 1999 with a system that will link into work in progress.

Property database - this system will record information on properties for use as comparables in valuation and professional work and also to record details of property deals where the firm has acted in order to capture its knowledge. Such a system could and should be developed to provide a recall so that further work under lease contracts can be obtained at a appropriate dates in the future.


About Gooch Webster databases.


The client database has fallen into disrepair. It needs to be developed as a client and contact database to include all information about the company’s activities.

Calendar - the firm has just developed and introduced a shared electronic calendar system. This functions as an electronic diary both for individual and management use. The system allows groups of diaries to be read simultaneously in order to track where people are and more importantly to spread knowledge of whom they are meeting and where they are going. This enables employees to share knowledge, seek knowledge and to innovate.

Central Directory - this needs to be further developed into a “know who and know how system”. It should include photographs of personnel, CVs, courses attended and explicit knowledge with the aim of enabling teams to be assembled, knowledge and expertise to be shared, and core competencies registered.

Personnel Database - the company needs a good personnel database which links to other systems so that the valuable assets of the firm, its people, can be tracked monitored, appraised, updated and generally looked after.

Fault tracking - this database will enable people to enter details of administrative and house keeping matters, which require attention. These can be tracked by the appropriate administrative people to ensure that action is taken. It will cover such matters as the replacement of light bulbs, efficiency of cleaning services, availability of stationery, meetings rooms and such like.

Rating - the rating department needs a modern database to deal with the forthcoming rating revaluation. A proprietary software system will need to be purchased.

Web Site - the company has plans to develop both the content and the format of a web site.

Software management system - this will track and monitor faults and enable the remote fixing of IT problems. It will remove some of the load on the existing IT team.

Agency system - following the departure of some of the administrative staff a comprehensive solution is now needed for the creation, storing and categorising of agency information. This is information on properties, which are available for purchase or to let.

A central filing system is required, common to all offices.

Systems used on a daily basis

The best example of a broad knowledge repository is the Internet. As a source of outside knowledge, the Internet can overcome some of the disadvantages of the localness and asymmetry of knowledge, since a subject search will return results from the whole system. Lotus Notes and databases are two of the leading toolsets for


The main knowledge repositories are Lotus Notes and the databases.


managing knowledge repositories in the firm, while emails and telephones comprise the main media of exchanging knowledge on a daily-basis.

The archiving mechanism, which is being used for recording dead file storage electronically, is called FiFi (File Finder). Prior to the merger, all the information, reports, letters and documents, generated by surveyors, was stored locally on the secretaries’ PCs and then the PC was backed up. But this hampered the sharing of information, since one could not get into a certain document unless the secretary helped you to access it. After the merger, the two sides have agreed on putting all the projects down on to the file server, they call it T drive and although each department has their own subdirectory, they all use this file server for their proposals. There is now no formal naming convention for documents, except that the name is being placed at the bottom of the document.

Last but not least, all the procedures, no matter how different they might be from department to department, are co-ordinated by the Quality Assurance Master Procedure. It describes the professional services provided by Gooch Webster. It defines the policy and commitment to quality by management, and details the way the company is organised. It also describes the documentation that supports the quality management system.




****************


******

********************

Search and Retrieval browser interface to KM services basic: keyword and boolean search

*******


In which way are the projects stored and retrieved?


basic: metadata based search advanced: combine criteria advanced: semantic analysis search over different document types restrict search area flexible ordering of search results smart pull mechanisms

******

**********

******


************

*********

************


************

*******

*******


*********

****


****************

Technical Requirements scalability ****



reliability security, integrity, and authenticity

*******



Appendix D: UBS’s Requirements

The requirements described in this paper focus on the needs of Credit Risk Valuation (CRV), but are influenced also by experiences made at other places within UBS. In order to make the requirements product-independent, also requirements are listed which are covered by Lotus Notes, such as access control. The requirements do have different importance. For example, text mining clearly is less important for CRV than are search facilities. However, requirements which have low priority in the given context could be more important in other contexts and thus be the target of further research work.

Section 1 of this paper briefly describes the KM goals of CRV and Section 2 characterizes the daily work of CRV. Section 3 states in detail the requirements. The importance of the requirements is then qualified in Section 4. Section 5 summarizes and makes a few suggestions regarding further work.

KM goals of Credit Risk Valuation

The KM goals of Credit Risk Valuation (CRV) are essentially

to provide information about CRV on the Bank Web, e.g. its mission, goals, organization, contacts, etc.

to make use of knowledge available on the Bank Web, which includes

general information and knowledge, e.g. directives and phone books,

city maps, etc. and

specific information and knowledge, e.g. products suited for corporate clients

to support daily work better

within Corporate Clients as well as

within CRV, such that the team's knowledge is available for everybody in the team and the knowledge pool helps new team members to get started within a short time.

Some major characteristics of the daily work of CRV

CRV is responsible for estimating the credit risks of corporate clients. This includes ratings for different regions, industry sectors and single companies. The members of the CRV team

rely on information and knowledge from different internal and external sources,



work largely in groups; group composition may be temporary and groups may be locally distributed,

produce different kinds of documents2 for different audiences (for example, a colleague, the own team, a particular department of the bank, the management, etc.), and

need to trace back decisions in order to improve their rating tools and to make rating updates more efficient.

Further characteristics are

the current set of documents is already quite large;

apparently not all of the knowledge/information is stored in electronic form;

modifications and updates are frequent (e.g. new versions of documents; condensation of a paper in form of a presentation, etc.);

work at CRV is highly creative and lowly standardized;

access control is extremely important; it is crucial that documents are accessible only by the intended audience!

CRV already relies on some means for knowledge management, which are HTML based.

Requirements

General requirements

Support for major hardware and software platform

The KM tool should run ideally on any kind of workstation. Currently, most important are PCs with Windows NT. CRV prefers that the KM tool itself runs on a PC with Windows NT, because this way no additional hardware and software would be needed. The KM tool's knowledge base should be accessible to a bankwide audience by Internet browsers such as Netscape.

2 Examples are team lists, task lists, weekly progress reports, status reports for the management, mapping tables, discussion proposals, meeting protocols, definitions, descriptions of methodologies, presentations, explanation of task-specific terms, analyses, FAQs, rating tables and software documentations.



Security, integrity and authenticity

Security3, integrity4 and authenticity5 are crucial for a department like CRV and are clearly "a sine qua non".

Scalability

Currently, the data set of CRV is of medium size. However, increases are to be expected. Thus the KM tool should be able to handle larger data sets as well. In a similar way, the KM tool must be able to handle an increasing number of users.

Reliability

As the KM tool is supposed to support the daily work of CRV, reliability is extremely important.

Basic KM requirements

Homogeneous environment serving different purposes

A powerful KM tool should allow for

gaining access to the "organizational memory" of the company, e.g., the Bank Web,

sharing knowledge within a group of persons (one person may belong to different groups depending on the task at hand!), and

organizing personal knowledge (early drafts of current work, private stuff, etc.)6

Clearly, non-public information shall be hidden from others.

Seamless collaboration with other tools

KM comprises many different areas such as Information Retrieval, Document Management, Workflow Management, Calendar/Scheduling, Project Management, CSCW, and Text Mining. In order that the KM tool can serve as a homogeneous environment serving different purposes, it might be necessary to have access to other tools and services (e.g., databases, agendas or other KM tools).7 Thus, a seamless

3 Nobody can access a piece of information or knowledge who is not allowed to do so.

4 Guarantee that no unauthorized changes happened.

5 The identity of an entity (e.g. a person or a server) is certificated.

6 The HTML-based Bank Web will hardly ever be such a homogeneous environment. This is mainly due to the lack of user authentication.

7 I am currently involved in different projects and activities in different contexts (UBS and Knownet). Unfortunately, there is a sharp gap between the Knownet activities and my other activities. For instance, there is no link between my UBS agenda and the Knownet agenda. Of course, the Knownet members do not need to know minor details such as one hour meetings, but they should be able to know about major leaves of absences, which I enter in my UBS agenda. A second example regards project management. Within UBS certain tools are used for



collaboration with other tools and sources allowing for exchange of data seems important.

Support for user participation

The KM tool shall allow changes (comments, the adding new documents etc.) by the user wherever possible as this promotes the sharing of knowledge. The Bank Web, for example, is of static nature - the regular user does not have the possibility to add information/knowledge to the Bank Web. Taking an organizational/management point of view, UBS practices an in-house culture where the employees are passive consumers of information/knowledge. This definitively does not promote the sharing of knowledge.

Knowledge organization and reorganization

The KM tools should not impose any specific form of organizing knowledge. It should allow one to arrange knowledge successively into smaller pieces (e.g. by putting documents into folders and subfolders) 8 and make relationships explicit (e.g. by links).

It is extremely important that the KM tool allows for reorganizing the data/knowledge base. This includes the definition of new (sub-)categories, renaming and deleting of categories, moving documents to new places, etc. Such activities should be possible for running applications as well (which is not the case with Lotus Notes).

Such a reorganization does not necessarily require a physical reorganization. It could be achieved also on the basis of a simple static organizational schema by providing different dynamic views for the same knowledge pool (see 3.4.2 later in this paper for more detail) and adapting them as needed.

Document handling and management

Smooth integration of existing documents

As CRV already practices KM and has put together a knowledge pool consisting of a quite large set of documents, a smooth transition from the current environment to the new KM solution is extremely important. Further, one has to keep in mind that many of the documents are HTML documents with links to other documents. It is not clear yet what happens with these links when the documents are integrated into a Lotus Notes database. Leaving the documents where they are and using links to the original documents within the KM tool might not be a good solution neither, as this may weaken access control.

Ease of adding new documents and modifying existing ones

It must be easy to add new documents and to modify existing ones. Nevertheless, it must be guaranteed that the user is permitted to do so.

reporting efforts and expenses. However, for properly doing project management, additional tools are needed. A close collaboration between the different tools would be desirable, in order that one does not have to feed different tools with overlapping data.

8 To my knowledge, Lotus Notes has shortcomings here. It supports maximally four hierarchy levels.



External access to documents

The KM tool should allow the user to work with a document elsewhere (e.g. by providing export facilities). This should not only be possible for single documents but for whole document collections maintaining the relationships between the documents.

Conversion of printed documents into electronic forms

It should be possible to integrate printed documents into a document collection (by scanning the documents and applying OCR). The handling of such documents should be (almost) the same as for electronic documents, e.g. search on such documents should be possible.

Full support for all major document types

Within the Bank, the most frequent document types are Word, Excel, PowerPoint and Adobe PDF. A complete list of document types, however, cannot be specified. The KM tool should support the major document types and should be extensible such that additional document types can be handled.

Metainformation

For each document, it should be always clear

who is the author/owner,

when was the document created,

when was it modified, as well as

who is the intended audience (a particular group of persons, any staff member, outside world, etc.).

Further, it may be necessary or helpful

to show the status of document (draft/official),

to point to previous/later versions,

to specify the lifetime of a document (for instance, a proposal for a meeting agenda could be archived after a short time in order to keep the document collection slim and up-to-date).

Flexible access control mechanisms

Access control mechanisms are required, especially for guaranteeing confidentiality and for being able to offer a homogeneous environment serving different purposes (see 3.2.1). The access control mechanisms of operating systems are not sufficient, because they are typically too static and do not support well that a user may belong to different groups depending on the task at hand. Further, it may be necessary to define access rights on a document-specific basis.



Archiving

Due to new forms of distributing information and sharing knowledge, it may not only be necessary to archive documents of general importance (e.g. directives) but also documents which gained the interest of a single reader. Let us consider a small example: a user bases a decision on a piece of information found on the companies intranet. Some time later the user is asked to explain the decision. This may be difficult, when the piece of information has been removed or overwritten in the meantime! Individual storage would burden the user and would rise authenticity and integrity problems (HTML files and Word documents can be easily changed). Thus, the reader of a document should be able to request archiving. Of course, it should still be possible to request archiving on creation time.

Information access

Powerful search facilities

Search facilities should allow the user

to formulate different kinds of search criteria (e.g. criteria regarding document contents, document type, metainformation, etc.),

to combine search conditions,

to search for information stored in any kind of document or information source (Lotus Notes has severe shortcomings as certain document types are not indexed, for instance PDF documents, which means that information/knowledge stored in a PDF document can not be searched for!),

to search in a specified area of a knowledge pool or the entire pool, as well as to search in external sources, and

to order search results according different criteria.

Further, as frequent changes are to be expected within CRV, the internal indexes must be updated within short time intervals.

Improved access to information and knowledge

When storing documents, one usually applies an organizational schema. Such a schema could be oriented, for example, along topics, projects, document types, or the organization´s hierarchy. Unfortunately, each organization form has its strength and weakness, and no single schema will be able to serve all the different needs. Below, this problem will be elaborate in more detail.

Mandatory readings: More and more it happens that information/knowledge people should be aware of is not distributed individually but is posted on intranets, in Lotus Notes databases, etc. An example is the IPR proposal prepared by Ron Young for the Fontainebleau meeting. In addition to reading the proposal, the consortium members were expected to comment it, but apparently nobody did so. Personalized KM tools could make the user aware that there are readings the user is expected to have a look at, and remind the user that a response is due. No doubt, such facilities would be extremely helpful and could relieve humans from writing



notification mails, asking for replies, sending reminders, etc. However, there is a certain risk that such tools could be perceived as intruding. Thus, a careful design, implementation and introduction seems extremely important in order to gain the user's acceptance.

New or modified documents: Taking the Bank Web as an example, the user can ask to be informed when new documents are added or changes occurred. This is done by sending the user a notification e-mail. Unfortunately, the user easily can get overwhelmed by such e-mails (which typically arrive when one is busy with a different matter). A more user-friendly way of notification could be to provide a specific view which shows all the documents of a site (or part thereof) which have changed or have been added since the last visit of the user (one could call this a personalized pull notification).9

Interest in particular topics: The Bank Web -taking it as an example again- is hierarchically organized. This is, for instance, helpful when one wants to find out more about a department. However if one is interested in a particular topic which is important within different departments (for example, one would like to know what is going in the area of Java programming within the different IT departments of the bank), a topic-centred organization would be more helpful. Of course, one could perform a search. However, a list of documents dealing about the topic in some way is clearly not very helpful neither. A more suited form would be to organize the documents along an ontology or a thesaurus, this way constructing some sort of knowledge map.

Content descriptions: There are many occasions where short description of a document´s content would be extremely helpful. For instance, a user could better judge the relevance of a document before clicking on a URL and waiting for the application to be launched and the document to be loaded. Although much research efforts have been spent, textual content descriptions are out of the reach for broader applications, at least for the near future. However, content deceptions in form of a set of concepts/terms seem feasible. In addition to give the user a better idea of a document´s contents, such descriptions could also be used to provide alternative, content-oriented views of a document collection. Further, the same underlying technique also could be useful to indicate changes between two versions of a document.

9 In an e-mail Thierry wrote: "Is there any 'push' mechanism (such as an email) that notifies to the user that new messages have been posted, and provide an abstract view of all the new postings". Within the consortium we seem to have the problem, that too much discussion is going on per e-mail although we decided to use the Knownet site @ KA for discussions. To me the problems seems to be that one too often visited the site just to find out that nothing happened in the meantime. After a while, one cares less about possible changes. Thus, I fully agree with Thierry that we need notification services. However in most contexts, I would prefer a pull mechanism rather than a push mechanism.



Project management and support for team workFor the daily work of CRV, project management facilities would be helpful. Most urgent seem to be facilities for defining work packages and for maintaining to-do lists.

If we decide to work on this topic, we need to evaluate closer what commercial products (especially Livelink 8) already offer in order to prevent that we re-invent the wheel.

Better support for discussionsDiscussions are an important part of CRV's daily work. To a large extent, e-mail is used. In certain cases, discussion takes places in form of more formal memos. Unfortunately, contributions to discussions are stored at different places depending on the tools used. Further, discussions which involve several people and continue over a longer period of time are not well supported. This is mainly because the different contributions appear as a flat list of "Re: ..." units not showing the particular focus of a contribution neither making the relationships between contributions explicit.

Ontologies/thesauri and classificationClassification is important for CRV and they are willing to put efforts in classifying their documents. Classification could be done at several layers, for example specifying the document type (memo, presentation, etc.), status (draft, approved, etc.), and contents of a document. Such document descriptions would allow additional search facilities and would improve the access to information and knowledge. Content descriptions could be provided by hand or be derived automatically. As the automatic derivation of content descriptions is still experimental, a mixed mode with short, rather general, manually assigned content descriptions in addition to automatically derived ones is possibly the best solution.

The classification of document contents best works with a formal conceptual model of the domain in form of a thesaurus or an ontology. Such conceptual models have to be built up first and then to be maintained continuously. Both tasks require manual work but could be supported by text mining techniques.

Text miningText mining aims at deriving patterns from documents of a collection. Such patterns typically are collection-specific phrases or multi-word terms. Text mining tools could learn, for example, from analyzing the Bank Web that "service level agreement" is an organisation-specific phrase and SLA is its abbreviation. Such collection-specific phrases could be used, for instance, to augment a domain-specific thesaurus. Further, they could improve the automatic derivation of content descriptions.



Tools for knowledge managers

Access mining10

This idea came up during a discussion with the knowledge manager of one of UBS´ IT departments. The knowledge manager relies on statistics about accesses to his KM pages on the intranet. This helps him to improve the organization of the information presented, e.g. what should be included in a particular view. More sophisticated analysis tools could be employed in order to find out typical access patterns of different types of users or even single users.11

Although this idea is inspired by a particular case, it could be applied in other situations as well.

Document and link maintenance tools

Tools should help the knowledge manager to find dangling links, outdated documents (i.e. that were not modified after a certain date), cyclic links, etc.

Retrievability of documents

I have noticed that many documents on the Bank Web are not found by the official search engine, even if an appropriate query is given. The search engine relies on a list of locations to be indexed. Apparently, this list is not complete. To avoid write-only information/knowledge, there should be tools that make sure that a document is retrievable by the corresponding search engine. As I learnt, similar problems have occurred at other places, which suggests that such tools are really needed.

Qualification of the requirements

Title Importance Remarks

3.1 General requirements essential Should be guaranteed by the

underlying KM tool (e.g. Lotus Notes

or Knowledger). To my knowledge,

the scalability of Lotus Notes is

restricted. Thus, we should be clear

about the needs in the near future!

3.1.1 Support for major ardware and

software platforms

3.1.2 Security, integrity and authenticity

10 The term "access mining" has been created in analogy to data mining and text mining.

11 Hannu Toivonen et al. from the Univ. of Helsinki did first work in this area. They developed an approach to suggest with a certain confidence what URL will be looked up next given a time-stamped URL path.



3.1.3 Scalability

3.1.4 Reliability

3.2 Basic KM requirements

3.2.1 Homogeneous environment

serving different purposes

important As a prerequisite, we need access

control mechanisms. Then, this is

more a matter of the design of a

concrete application.

3.2.2 Seamless collaboration with other

tools

important Must be supported by the underlying

KM tool.

3.2.3 Support for user participation essential Again, this must be supported by the

underlying KM tool. Further, it has to

be taken into account when designing

an application.

3.2.4 Knowledge organization and

reorganization

important Commercial KM tools differ

significantly with regard to this. Lotus

Notes is weak here, as it supports

only a limited number of hierarchy

levels and it does not well support the

knowledge reorganization. When

developing an application for CRV, we

have to be careful about this topic!

3.3 Document handling and management

3.3.1 Smooth integration of existing

documents

essential Here again, commercial products

differ significantly. With Lotus Notes,

the integration of existing documents

seems to be tricky.

3.3.2 Ease of adding new documents

and modifying existing ones

essential All major KM tools seem to support

this.

3.3.3 External access to documents essential If we do not have this feature and later

decide to use a different KM tool, all

the knowledge would be lost! Lotus

Notes does not have a feature for

exporting documents, but it can be

extended by agents doing this. When

developing an application for CRV, we

should have this issue in mind and

make sure, that we can employ a

different tool later, if needed.

3.3.4 Conversion of printed documents

into electronic forms

nice to have This could be done quite easily with

commercial products without much

efforts.



3.3.5 Full support for all major document

types

important Unfortunately, Lotus Notes is weak

here, especially with regard to PDF.

However, there a helpful commercial

tools available. Possibly we could find

also a public domain solution. Hence,

this topic is more an engineering than

a research issue. Further, as

workarounds seem possible, this does

not disqualify a solution based on

Lotus Notes in no way.

3.3.6 Metainformation essential Metainformation can be provided

partially in an automatic way.

However, certain kinds of

metainformation require manual work.

We have to keep in mind that

metainformation has to be provided

consistently. Otherwise, tools relying

on metainformation (e.g. search) will

not work properly.

3.3.7 Flexible access control

mechanisms

essential This feature is provided by all major

KM tools.

3.3.8 Archiving important, but

not first priority

For the beginning, this is not an

important issue, but we definitely

should keep it in mind.

3.4 Information Access

3.4.1 Powerful search facilities essential Commercial products already support

search quite well. We have to see

whether extensions are necessary, for

instance for searching in sources

outside of the KM application.

3.4.2 Improved access to information

and knowledge

essential This is a topic which requires both

research and engineering work.

3.5 Project management and support for team work important / very

nice to have

With regard to this topic, we should

know more what other KM products

already provide. Livelink 8, for

instance, is pretty strong in this area.

What about Knowledge Associate´s

project management tools?



3.6 Better support for discussions important This is an important issue at CRV and

worth to do research!

3.7 Ontologies/thesauri and classification important We could work towards tools that

support the creation and maintenance

of ontologies as well as tools that

automatically derive content

descriptions. Both clearly are a

research issues.

3.8 Text mining important As classification is an important topic,

text mining could be a research issue

as well. However, there are

commercial products on the market,

e.g. IBM offers solutions. Thus, we

should see first, whether we can find a

commercal tool that fulfills CRV's

needs before starting any research

work.

3.9 Tools for knowledge managers

3.9.1 Access mining very important This is a research topic. Requires

expertise in data mining or similar

areas.

3.9.2 Document and link maintenance

tools

very important This is more an engineering problem,

but still a demanding task.

3.9.3 Retrievability of documents nice to have Probabely more important in other

contexts.

Conclusions

Lotus Notes, which is underlying Knowledger, seems to be a good starting point for KM in applications with smaller or medium-sized collections of documents. However, before employing Lotus Notes/Knowledger, we have to make sure, that the existing document collection of CRV can be easily imported. Further, we have to be confident that documents of a Lotus Notes database can be exported and used by other KM tools as well, in case we detect any major problems or find a product which is much better suited. If we find positive answers, an CRV-specific application based on Lotus Notes/Knowledger can be designed, which should be able to improve KM at CRV significantly within a short period of time. Several extension are desired. Some of them are more an engineering issue others are more a research issue. The most important



research topics are listed below. The more ‘*‘ symbols are assigned to a topic, the more important is it to CRV.

Better support for discussions (****)

Ontologies/thesauri and classification (*)

Support for the creation and maintenance of ontologies/thesauri

Automatically derived content descriptions

Access mining (**)

Document and link maintenance tools (***)

Improved information/knowledge access by dynamically generated views (supported by agents exploiting metainformation and classification results?) (****)

what´s new?

mandatory readings

topic-centered views, etc.

We summarize UBS’s requirements within our above introduced classification table:

Distribution, Publishing, and Filtering not appl. nice import. essent. notification / push mechanisms



****************

Collaboration e-mail communication real-time communication document-sharing, white-boarding net-based meetings, conferencing threaded discussions availability (who is accessible, how ) company yellow pages

******

******

**********

Search and Retrieval browser interface to KM services basic: keyword and boolean search basic: metadata based search advanced: combine criteria advanced: semantic analysis search over major document types restrict search area flexible ordering of search results smart pull mechanisms

****************

*******

***********

Indexing



basic: corporate taxonomy / ontology basic: metadata support, incl. versioning basic: short content description flexible knowledge organization & reorganization corporate thesaurus advanced: pattern detection / text mining advanced: document clustering user profiling map sources / contents to business views

********

*******

***************

Storage not appl. nice import. essent.

support for all major document types easy addition / modification of documents smooth integration of existing documents convert printed into electronic documents export facilities flexible and powerful access control mechanisms archiving

************

******

*****


******

*****


******

*****


************

Abbildung 18: UBS’s Compilation Table



Appendix E: FHBB’s Contributions

Awareness Creation: Individual Knowledge Portfolio (KnowPort)IntroductionAs soon as individuals begin working with a knowledge management infrastructure for sharing knowledge within their teams and enterprises, a profound cultural change in their competency requirements, roles and responsibilities takes place. The objective of the KnowPort project is to support these individuals in meeting the challenge of such a cultural change by helping them in the task of self-managing their own knowledge resources. For this we intend to cut across the classical domains of knowledge acquisition, human resource management, performance management, career planning, learning planning, computer based learning and self-management with the aim of explicitly transferring to the individuals the role of manager of their own knowledge portfolio.

KnowPort will allow individuals to manage their personal knowledge assets and in this way provide a basis for making them more effective in the tasks of developing learning capabilities and participating in knowledge sharing.

Knowledge Management in OrganisationsMany experts agree that the West shows now a clear trend from the information society towards an open knowledge society [ArthAnd96]. To meet the challenge of this information overflow and knowledge ocean, many international organisations are developing Knowledge Management strategies and systems which enable them to better manage their core competences and knowledge resources.

A systematic account of knowledge management activities at department or enterprise level shows the following 6 core processes and 2 control processes [Probst97, 47-60]: (1) Identification, (2) acquisition, (3) development, (4) sharing, (5) application and (6) storage of knowledge, closed into a control loop by (7) the definition of knowledge goals and (8) the evaluation of goals achievement. Similar Knowledge Management frameworks distinguish between 4 and 7 processes [O'Dell97, 10-11], [Speck97].

Individual Knowledge ManagementThe basic idea from which the KnowPort concept has been started was that the same 6 + 2 classes of activities which help managers to deal better with knowledge in their enterprises and departments could also be used as a paradigm for the individual who aims at improving the management of his personal knowledge. Like in organisations [Hiebeler96], where the central issue is not creating organizational knowledge but figuring out what they already know ("if only we knew what we know", Jerry Junkins quoted in: [O'Dell97]), we suggest that the same approach should be applied to



individual knowledge ("if only I knew what I know"). Therefore the primary task of KnowPort that we designate by "knowledge survey" corresponds to the first of the above mentioned core processes of knowledge management at enterprise level, called there "knowledge identification".

At the beginning KnowPort will limit its scope to the "knowledge survey" task but our long term strategy is to extend its features step by step for supporting employees also in sharing personal knowledge, defining knowledge goals and evaluating goals achievement (see above core processes 4, 7 and 8 respectively).

Action oriented knowledge tracing The main innovative concepts on which KnowPort is based lie in a new approach to the method for obtaining a knowledge survey. Individual knowledge becomes manifest as any explicit knowledge in documents that can be seen as external memories. The ideal-typical knowledge is stored in encyclopedias, but the encyclopedia is necessarily context neutral and does not contain actual knowledge such as news from the stock-broker given by phone the day before or conditions to a tender received four weeks ago.

As it is common to refer to such context sensitive and actual knowledge we will take in the KnowPort project an action oriented constructivist approach, based on the following principle: "Trace your Tack". For sailing on the right tack in the knowledge ocean and mastering the information overflow by entering safely new ports of knowledge trace your knowledge in action (on tack). This means that KnowPort will have to support the user in producing a trace of knowledge relevant events during her "Knowledge Tack", i.e. her construction and use of personal knowledge in the context of a running task. Such traces will avoid the difficulty of retrospection, encourage the use of KnowPort, help the user in focusing on the process of knowledge construction and provide the basis for personal knowledge analysis.

KnowPort components and features

Our approach for supporting efficient and effective tracing of knowledge relevant tacks is basically to combine in a composite environment several independent tools designed to meet the salient requirements of knowledge work:

a communication tool for tracing different-time E-Mail conversations

an agenda tool for tracing contexts, objects and activities of tasks done

a file tool for tracing files (applications and data) used

a word tool for tracing and connecting critical concepts used

MailTack

This communication tool helps users in keeping a constant overview of different-time conversations and discussions done by E-Mail with several partners. Different topics that develop independently over time during a number of E-Mails and with contributions of different persons can be connected into independent argumentation chains; each



chain can then be viewed as linear text output (without the text segments of other topics) or as map of linked nodes over the global map of the complete discussion.

TaskTack

This agenda tool provides a daily overview of all tasks done. A task is specified as a record of date, context, object together with the activity performed on that object in that context; optionally duration and a comment can be added. The user can organize contexts, objects and activities into classification and aggregation hierarchies that can be browsed in a graphic display. When a new entry for a task performed has to be added, the system supports reuse of existing context, object, and activity specifications so that the same task will always be described univocally. With interfaces to the most common planning tools, TaskTack will help user to become aware of the knowledge hidden in their flat lists of daily tasks data.

FileTack

This file tool records the chain of file handling actions, like starting an application, working with a started application, opening or closing a file and working with it, accessing an internet site, etc. The record includes date, time, operation, duration, file name and size. The user can view charts of these records over a selected time interval (calendar days) and identify patterns of her use of software resources, for instance track down the historical progression of a particular file or find the amount of time spent using a specific application.

WordTack

This word tool supports the user in specifying her current understanding of critical concepts. The meaning of a word can be described by writing hypertext expressions and by drawing a semantic network. In the hypertext, the current description of a word is linked to the individual descriptions of the words it uses and in the semantic network a word is linked with the words of its description by different types of semantic relations.

These four tools for knowledge tracing, which will be implemented as Windows applications with interfaces between them as well as to the major office packages, constitute the kernel of the KnowPort architecture.



Figure 19: General KnowPort Architecture

Perspectives Around the knowledge tracing kernel of the KnowPort architecture, various analysis and validation tools will be added later. They will assist knowledge workers in modeling and managing their individual 'knowledge portfolio' on the basis of their traces of knowledge relevant tacks. This will include supporting individuals to:

review their own performance and identify their competencies,

compare these competencies and performance against the requirements derived from their company's strategic goals,

specify and prioritize learning and development goals according to corporate, team and personal needs,

find appropriate paths for achieving these goals,

control goal accomplishment and measure skill attainment and, last but not least, manage their activities of knowledge sharing (Fig. 1).

FAQ’sAt this point of our research many question are still open for discussion within our group, for instance:

Which are the advantages of tracking a knowledge tack?

What are the main principles of the methods for analysing the knowledge traces?

Which are the strength of this tool compared with other similar tools?

How to solve the problem of information overload due to a lot of irrelevant stuff contained in the traces?

Relevance and irrelevance



What we are interested in is primarily supporting the user to follow the course of development of knowledge-intensive tasks. For this reason, in the moment of doing such a task, a step is relevant to the user if it contributes to trace the course of development. Later the user criteria of relevance will change depending on what she will be focusing on: so, when she will look back at the tack followed, one and the same step will be one time relevant and another time irrelevant. In this perspective the traces do not contain a priori "a lot of irrelevant stuff".

Advantages and strengths

We expect that the main advantages of our tool will come from the main strength of our approach, namely its explicit foundation in knowledge theory. In fact, our method for obtaining a knowledge survey is strongly based on the constructivist view of knowledge developed by Ernst von Glasersfeld and known as Radical Constructivism [vonGlasersfeld95] which shows that knowledge and experience can never reflect a ready-made rationally structured real world. Instead, the rational structures that make up our knowledge are determined by what the subject makes in organizing and managing her flow of experience according to criteria of consistency and coherence. This is why focusing on knowledge construction is so important!



References

Abecker, A.; Aitken, S.; Schmalhofer, F. and Tschaitschian, B. (1998) ‘KARATEKIT: Tools for the Knowledge-Creating Company’, In: KAW`98, Eleventh Workshop on Knowledge Acquisition, Modeling and Management, Banff, Canada, 1998.

Abecker, A.; Bernardi, A.; Hinkelmann, K.; Kühn, O. and Sintek, M. (1998) ‘Towards a Technology for Organizational Memories. IEEE Intelligent Systems & Their Applications, Vol. 13, No. 3, May/June 1998.

Abecker, A.; Bernardi, A. and Sintek, M. (1999) ‘Enterprise Information Infrastructures for Active, Context-Sensitive Knowledge Delivery’, In: ECIS'99 - The 7th European Conference on Information Systems, Copenhagen, Denmark. June 1998.

Tschaitschian, B.; Abecker, A. and Schmalhofer, F. (1997) ‘Information Tuning with KARAT: Capitalizing on Existing Documents’ In: Enric Plaza and Richard Benjamins (eds.) 10th European Workshop on Knowledge Acquisition, Modeling, and Management (EKAW-97) . Sant Feliu de Guíxols, Catalonia, Spain. Springer Verlag, LNAI 1319.


REPORT COVER PAGE - DFKIaabecker/Postscript/D2_1.doc · Web viewthe project directors and project...

Documents

Transcript of REPORT COVER PAGE - DFKIaabecker/Postscript/D2_1.doc · Web viewthe project directors and project...