Automatic Creation of User Profiles for Achieving Personal Web Accessibility

Automatic Creation of User Profiles Automatic Creation of User Profiles for Achieving Personal Web for Achieving Personal Web AccessibilityAccessibility

Markel Vigo, A. Aizpurua, M. Arrue and J. Abascal

Laboratory of HCI for Special Needs

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

University of the Basque Country

OutlineOutline1. Motivation

2. Introduction

3. Personal Web Accessibility

4. Context of Use Detector:

4.1 Components

4.2 Architecture

5. Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

MotivationMotivation

Introduction ConclusionsMotivation

• Accessibility is about accommodating people with disabilities

• User-modelling/user-adapted interaction practices can be applied to adapt web systems for people with special needs

• Context of use has to be captured to have acceptable accessibility levels

• A software component that collects System features Assistive technologies User characteristics

• Valuable information can be inferred

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Personal Web Accessibility

Context of Use Detector

IntroductionIntroduction

Introduction Conclusions

• User profiles are used (not only) in adaptive (web) user interfaces, intelligent help systems and explanation systems

• User models/profiles store information about users' characteristics preferences behaviour knowledge on a subject, etc.

• Data about the user can be collected Explicitly: filling in forms or questionnaires

precise, reliable, hard to collect, obtrusive Implicitly: monitoring user's interaction

natural, less reliable, unobtrusive

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation Personal Web Accessibility

Context of Use Detector

IntroductionIntroduction

Introduction Personal Web Accessibility Conclusions

• Related work on collecting data for user-adapted interaction applications for people with disabilities: Questionnaires (interviews) Filling in electronic forms Smart cards

• User profiles for accessibility store information about users' abilities software and hardware characteristics of the access device data regarding the installed Assistive Technologies

• Data are encapsulated in a CC/PP file

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation Context of Use Detector

Personal Web AccessibilityPersonal Web Accessibility

IntroductionPersonal Web Accessibility

Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• Personal Web Accessibility aims at providing user-tailored resources User profiles are the foundations to achieve Personal Web

Accessibility Guidelines are adapted to user and delivery context requirements

• Regarding guidelines such as WCAG, Satisfying them does not ensure accessibility Meeting them it is not necessary for certain users

• Why such contradiction? Guidelines are ambiguous Contain unresolved references to the delivery context Incompleteness

Context of Use Detector

Personal Web AccessibilityPersonal Web AccessibilityApplication ScenariosApplication Scenarios

Context of Use Detector Conclusions

• User-tailored Web Accessibility Evaluations Design-for-all vs. Design-for-one

• User-tailored accessibility scores Enriching Information Retrieval processes Accessibility observatories

• Adapting Web content according to a profile Server-side services are capable of adapting content according to a

profile

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation IntroductionPersonal Web Accessibility

Context of Use DetectorContext of Use Detector

Introduction Personal Web Accessibility

Context of Use Detector

Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• Objective: collect data about user's interaction context

• Requirements

Data have to be unobtrusively collected

Physical features and software support of the accessing device

Assistive technologies, be software of hardware

Multiplatform: Java on Windows or Unix systems

Output following standards such as CC/PP

Introduction Personal Web Accessibility Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• Assistive technologies product name and release version is of major importance New features of ATs make some accessibility barriers obsolete

positive dependencies Example for JAWS: there are not scripting issues for OnClick anymore

from version 5 onwards (WCAG 6.3, 6.4, 8.1, 9.3) False positives are produced

Older versions may suffer accessibility problems even if guidelines are met negative dependencies

Example for JAWS: older versions do not support acronyms (WCAG 4.2)

False negatives are produced

Context of Use DetectorContext of Use DetectorSome remarksSome remarks

Context of Use Detector

Introduction Personal Web Accessibility Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• A remote database storing information regarding ATs Maintained by an administrator Each record contains: name, vendor, version, description, modality,

etc

• A local database that mirrors the remote DB

• Assistive Technologies Detector Compares key values in the local DB with the installed ATs

• Access Device Features Detector Information concerning access device is retrieved

• Created profiles are stored in a repository

Context of Use DetectorContext of Use DetectorComponentsComponents

Context of Use Detector

user

Introduction Personal Web Accessibility Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

1. When the system initializes the coherence between DBs is checked

Context of Use DetectorContext of Use Detector Architecture Architecture

Context of Use Detector

remote

Local DB

ATs Information

Access Device Features Detector

Assistive Technologies

Detector

Profiles Repository

User Agent Evaluation Tools

developeruser

admin

<software features/><hardware features/><assistive technologies/>

CC/PP Profile

123

5 6

System API

2. DB entries are compared with key values in the Registry3. Information regarding access device features is collected4. Matches with key values and system information are put in a profile

4

5. User Agents can benefit and provide adaptive hypermedia techniques

6. Profiles are remotely stored and developers can make use of them

Introduction Personal Web Accessibility Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• CC/PP: gain interoperability with User agents Evaluation tools Server-side content adaptation services

• CC/PP provides a framework to describe hardware, software and user agent characteristics

• Attributes should be defined by third parties

• Reuse is encouraged

• A new vocabulary has been defined

Context of Use DetectorContext of Use Detector

Context of Use Detector

Example of a profileExample of a profile

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

<rdf:RDFxmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"xmlns:ccpp="http://www.w3.org/2000/07/04-ccpp#"xmlns:access="http://sipt07.si.ehu.es/profiles/2008/access-schema#">

<rdf:Description rdf:about="http://sipt07.si.ehu.es/profiles/2008/user_0017"> <ccpp:component rdf:resource="http://www.w3.org/2000/07/04-ccpp#HardwarePlatform"/> <ccpp:component rdf:resource="http://www.w3.org/2000/07/04-ccpp#SoftwarePlatform"/> <ccpp:component rdf:resource="http://sipt07.si.ehu.es/profiles/2008/access-schema#AT"/> <ccpp:component rdf:resource="http://sipt07.si.ehu.es/profiles/2008/access-schema#AT"/> <ccpp:component rdf:resource="http://sipt07.si.ehu.es/profiles/2008/access-schema#AT"/></rdf:Description>

<rdf:Description rdf:about="http://www.w3.org/2000/07/04-ccpp#HardwarePlatform"> <access:CpuName>AMD Athlon(tm) XP 2400+</access:CpuName> <access:CpuConstructor>AuthenticAMD</access:CpuConstructor> <access:ramSize>1035172 kB</access:ramSize> <access:display>1024 x 768 pixels</access:display> <access:keyboard>AT Translated Set 2 keyboard</access:keyboard> <access:ColourCapable>Yes</access:ColourCapable> <access:ImageCapable>Yes</access:ImageCapable> <access:SoundOutputCapable>Yes</access:SoundOutputCapable ></rdf:Description>

<rdf:Description rdf:about="http://www.w3.org/2000/07/04-ccpp#SoftwarePlatform"> <access:OSName>Linux</access:OSName> <access:OSVendor>Unknown</access:OSVendor> <access:OSVersion>2.6.9-1.667</access:OSVersion> <access:user>root</access:user> <access:JavaVersion>1.4.2_10</access:JavaVersion> <access:JavaVendor>Sun Microsystems Inc.</access:JavaVendor> <access:JavaVendorURL>http://java.sun.com/</access:JavaVendorURL></rdf:Description>

<rdf:Description rdf:about="http://sipt07.si.ehu.es/profiles/2008/access-schema#AT"> <access:ATName>Brltty</access:ATName> <access:ATVendor>The Brltty Team</access:ATVendor> <access:ATVersion>3.6.1</access:ATVersion> <access:ATType>Output</access:ATType> <access:ATDescription>BRLTTY is a background process (daemon) providing access to the Linux/Unix console (when in text mode)for a blind person using a refreshable braille display</access:ATDescription> <access:ATIOtype>Braille</access:ATIOtype></rdf:Description>

<rdf:Description rdf:about="http://sipt07.si.ehu.es/profiles/2008/access-schema#AT"> <access:ATName>K magnifier</access:ATName> <access:ATVendor>Kde Access Team</access:ATVendor> <access:ATVersion>1.0.0</access:ATVersion> <access:ATType>Output</access:ATType> <access:ATDescription>KDE Accessibility Magnifier</access:ATDescription> <access:ATIOtype>Magnifier</access:ATIOtype></rdf:Description>

hardwarefeatures

softwarefeatures

assistivetechnologies

accessnamespace

Introduction Personal Web Accessibility Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• Guidelines manager: Guidelines that have an impact on the user are retrieved Guidelines that cause positive dependencies are not retrieved Guidelines that cause negative dependencies will be considered

• Evaluation engine Produces user-tailored reports and accessibility scores

Context of Use DetectorContext of Use DetectorHow can profiles workHow can profiles work

Context of Use Detector

Evaluation Engine

Guidelines Manager

<software features/><hardware features/><assistive technologies/>

CC/PP Profile

Guidelines Repository

<html><head><title>Test file<title><body>foo

(X)HTML

EvaluationReport

1

2

3

4

ConclusionsConclusions

Introduction Personal Web Accessibility

Context of Use Detector Conclusions

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

Motivation

• This component is part of a more general framework for user-tailored evaluations

• Profiles are useful to infer user's abilities Select an adequate guideline set for evaluation purposes

• More accurate using individual abilities rather than group membership

• These profiles are useful to obtain a first coarse grained profile

• Privacy issues

Automatic Creation of User Automatic Creation of User Profiles for Achieving Personal Profiles for Achieving Personal Web AccessibilityWeb Accessibility

Markel Vigo, A. Aizpurua, M. Arrue and J. Abascal

Laboratory of HCI for Special Needs

11th International Conference on Computers Helping People with Special Needs, ICCHP 2008

University of the Basque Country

Any question?