Accessmonkey: Scripting Accessibility

Post on 25-May-2015

3.087 views 0 download

Tags:

description

Presentation from the WebInSIght group at the 2007 W4A conference.

Transcript of Accessmonkey: Scripting Accessibility

Jeffrey P. Bigham and Richard E. LadnerUniversity of WashingtonComputer Science & EngineeringW4A 2007

Scripting Accessibility

2

Goals

Remove dependence on developers Enable users to improve content Enable users to share improvements

More efficient development process Enable developers to leverage

improvements Automated improvements as suggestions

Introduction

3

Accessible Content Production

User Influence Choosing and configuring technology Consultation and Lawsuits

Introduction

4

WebInSight1

Introduction

Making Images Accessible

WebInSight

5

WebInSight (continued)

Proxy-based system inadequate Wanted client-based solution Implemented as a Greasemonkey script

Alternative text provided for web users Automatic improvement for users Suggestions for developers

Other transformations share this property!

Introduction

6

Outline

Introduction

Accessmonkey

Example Usage

Future Work

7

Greasemonkey1

Firefox Extension

Allows users to inject custom Javascript Customize look and feel Combine data from multiple sites Alter pages to make them more accessible

Desired features Include developers Used in multiple browsers Enable collaboration

Accessmonkey

[1] Greasemonkey Firefox Extension. http://greasemonkey.mozdev.org

8

Accessmonkey

Preserves Greasemonkey Features Multiple browsers and platforms Framework for users and developers

Web Developers Separate Interface Edit and save

Accessmonkey

Scripting Accessibility

9

AccessmonkeyFramework

Accessmonkey

Web developers Suggestions

Web users Automatic

10

Writing Scripts

Who can write scripts? Anyone who can write Javascript Those who would directly benefit (and

friends) Web Developers

Developer interface Implementations provide area Allows users to toggle user/dev mode

Accessmonkey

11

More than Manual Annotation Removes dependence on Web

Developers Users independently improve content

Collaborative Accessibility

Harnessing Human Power Amazon Mechanical Turk1

ESP Game2

[1] http://www.mturk.com/[2] von Ahn et al. Labeling Image with a Computer Game. (CHI ‘04).

Accessmonkey

12

Outline

Introduction

Accessmonkey

Example Usage

Future Work

13

Example Usage

Web users Find/create scripts they want Scripts transform pages automatically

Web Developers Find/create scripts they want Use to edit pages Save changes

Example Usage

14

15

16

17

What else can scripts do?

Personalized Edge Services1

Many simple transformations Context-driven Web Browsing2

Direct to content that matters Site-specific Scripts

Anything!

[1] Iaccarino et al. Personalizable edge services for web accessibility.[2] Mahmud et al. CSurf: A Context-Driven Non-Visual Web-Browser.

18

19

Menu: No Style

Inefficient to Access

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

Outline

Introduction

Accessmonkey

Example Usage

Future Work

37

Next Steps

WebInSight Access Tools Script Writing Script Discovery

Writing

ScriptRepository

WebInSight Access Tools

Discovery

Web Developers Web Users

Future Work

38

Easier Script Writing

Existing tools Programming-by-Demonstration

Web Macros1, WebVCR2, PLOW3, etc. Platypus

Not targeted for improving accessibility Visual interfaces

[1] Safonov et al. Web macros by example: users managing the www of applications.[2] Anupam et al. Automating web navigation with WebVCR.[3] Jung et al. One-shot procedure learning from instruction and observation.

Future Work

39

Script Discovery

Online Repository User-contributed scripts

Discover As-You-Go Finds Applicable Scripts as a User

Browses Find even if you can’t write

Challenges Usability Maintaining Security

Future Work

40

Conclusion

Accessmonkey Extends influence to web users Allows web developers to benefit Leverages skill and motivation of users

Embrace Web 2.0 Challenges need to be solved Potential to increase web accessibility

41

WebInSight

http://webinsight.cs.washington.edu

Thanks to: National Science Foundation, T.V. Raman, Sangyun Hahn, Oscar Danielsson, Gordon Hempton, Ryan Kaminsky, and GW Micro.