UNIVERSAL USABILITY IN PRACTICE
UNIVERSAL USABILITY IN PRACTICEIntroduction | Recommendations | Guidelines | Websites | Conclusions | Resources
Moustafa A. Youssef (moustafa@cs.umd.edu)
Department of Computer
Science
University of Maryland
College Park, MD 20742 USA
April 2001
We sometimes refer to our globally interconnected information infrastructure as the World-Wide-Web. At present, however, it is far less than that. For someone who reads only English, it is presently the English-Wide-Web. A reader of only Chinese sees only the Chinese-Wide-Web [1].
Researchers in Cross-Language Information Retrieval (CLIR) seek to support the process of finding documents written in one natural language with automated systems that can accept queries expressed in other languages [1]. CLIR has many useful applications. For example, multilingual searchers might want to issue a single query to a multilingual collection, or searchers with a limited active vocabulary, but good reading comprehension, in a second language might prefer to issue queries in their most fluent language. From the previous definition, it is clear that CLIR is a key factor to universal usability.
The first reported work on CLIR was the development in 1964 of the International Road Research Documentation system that used a controlled-vocabulary thesaurus [2]. However, multilingual thesauri do not completely solve the CLIR problem. The most serious limitation of thesaurus-based techniques is that non-librarians seem to have difficulty exploiting their capabilities [2] [3], which makes these techniques not universally usable. Searching free text is the obvious alternative to use of controlled vocabulary. Two dominant themes based on free text CLIR still guide research and practice: corpus-based and knowledge-based approaches [2].
To make a universally usable CLIR, many issues have to be studied. This paper, studies some of the issues concerning universal usability of CLIR systems, namely: level of users’ experience, users with disabilities, multilingual interface, and platform independence. In the rest of this paper, general recommendations for designers of CLIR systems are presented followed by more specific guidelines. Some examples of successful web sites are also presented. A section is devoted to resources about the paper topic. The paper ends with recommendations for future progress for researchers and practitioners.
There are multiple dimensions of accessibility: platform, language, disability, and levels of user experience. The highest priority goal of the user interface design is to be able to adapt the interface to inevitable requests for changes. The many unknowns during development necessitate the design for unexpected user needs, platform considerations, languages, and changing requirements [4][5].
- Multi-platform:
A good CLIR system should work on all current browsers including text based browsers such as Lynx. There are still many places (e.g. libraries with hundreds of old machines) that still use old browsers and would not be willing or able to upgrade their browsers.
Another issue related to the multi-platform issue is the screen size. CLIR systems should be designed to work with screens ranging from large (more than 900 pixels wide) to small screens (less than 700 pixels wide.)
Finally, a universally usable CLIR system should be designed to work across different hardware and operating systems (e.g. shortcut keys that are based on the operating system)
- Multi-lingual Interface:
Translating the interface and online help into different languages helps both novice and expert users to accomplish their task more easily than a single language interface and online help. Many considerations have to be taken care of. For example:
· The direction of the language: left-to-right (e.g. English), right-to-left (e.g. Hebrew), vertical (e.g. Japanese), or bi-directional (e.g. Arabic))
· Fonts: The fonts used in designing the interface and online help may not be available everywhere. Assumptions that fonts of the same point size will be rendered approximately the same size are wrong.
A CLIR system should be designed with internationalization in mind from start.
- Users with disability:
The system should be designed to support users with disabilities. For blind users, it should be compatible with screen-readers. Specially adapted graphical browsers (such as Microsoft Internet Explorer 4+) can serve people with other disabilities.
- Levels of Users:
The system should support users with different levels of experience ranging from novice to expert users (such as librarians).
The next section provides more specific guidelines to satisfy the previous recommendations.
These are not intended to be a complete set of guidelines. The issues involving universal usability of CLIR systems are diverse. However, the following guidelines can serve as a good base for system design and implementation.
1. Design for universal
usability from the beginning.
An architecture designed with the main goal of being able to adapt to changing requirements will facilitate exploration and implementation of individual issue.
2. Allow users to select the
number of results to be retrieved.
This allows users to choose the number that fits their needs. For example, if the response time is long, users can reduce the number of results to be retrieved to decrease the response time.
3. Provide different search
modes.
Novice users prefer minimal functionality to perform their limited tasks. On the contrary, expert users prefer to have full control of the system. Different search modes (e.g., basic (with support of most used features), advanced (with full features), and expert (to better support query language users)) will cover different levels of experience and increase users satisfaction.
4. Allow a user customizable
version of the interface.
For example, users should be able to select the default number of search results to be retrieved, the search mode, text only or graphics version, etc.
5. Provide a text-only
version of the interface.
This helps in many universal usability issues. It runs on any hardware, more suitable for users with slow connections, accepted by all web browsers (including Lynx), and compatible with screen-readers (to support blind users). In fact, a text-only version serves as the version with the lowest common denominator features for all browsers.
In Aurora [6], they took this a step further by introducing the concept of user agents which can adapt the data for display and interaction according to various needs and preferences (e.g., as speech for blind users, as text for sighted users, etc.) Consider persons who are blind and want to access an online auction site. If they use their voice-browser to access the site directly, they will encounter many obstacles (e.g., site-maps, advertising, banners, and other extraneous data.) With Aurora, users identify themselves as users of generic voice-browsers. Hence, Aurora selects an adapter that generates simple, graphic-free HTML page of the auction site.
6. Make use of accessibility
support tags in HTML.
Some HTML tags are useful to support accessibility:
·
title: The title attribute provides extra information about
what it is attached to. For example, FirstSearch [4] uses
title tags for input fields to provide more detailed prompts, and on links to
explain where they will lead the user.
·
label: The label tag allows a label to be more formally
associated with a form element with which it is logically associated. Screen
Readers knows that a label is associated with a checkbox, and some browsers
allow users to control form elements by clicking on their labels.
·
accesskey: The accesskey attribute allows Alt-x keys to be
associated with form elements. For example, FirstSearch associates Alt-s with
submit buttons, and Alt-c with clear button.
· alt: The alt attribute is used with the images tag in HTML to give a textual representation of the image.
Though these tags are not supported by all web browsers, they provide features useful to all users.
7. Design the system to
provide multi-lingual interface and online help.
This helps users with different languages to use your system.
8. Do not assume single-byte
characters.
One byte allows only 256 characters. The problem is that many languages, such as Japanese, require many more symbols. It is recommended to support Unicode [7].
9. Do not make assumptions
about fonts.
Many users have different standard fonts. If your system assumes the existence of certain fonts used in the interface, the results will be unexpected [7].
10. Provide a universal
usability statement for your web site
Universal usability statements are declarations by web site designers of the usability measures and concerns associated with a given web site. By knowing what is needed for a successful web site visit, users will be more likely to have a satisfying and successful experience [8]. A universal usability statement provides a summary of universal usability support provided by your site.
This section lists examples of some web sites that provide information related to universal usability of Cross-Language Information Retrieval.
· Cross-Language Text Retrieval Systems
http://raven.umd.edu/dlrg/clir/systems.html (accessed April 16, 2001)
This page lists cross-language text retrieval systems which can either be used over the web or obtained from commercial sources. In some cases, papers describing the theory and/or implementation details are also available on these pages
· FirstSearch service
http://NewFirstSearch.oclc.org/ (accessed April 16, 2001)
Online Computer Library Center (OCLC) FirstSearch allows users to search for bibliographic and full text records in over 80 online databases. FirstSearch was designed to adapt to unexpected user needs, platform considerations, languages, and changing requirements. It also includes a listing of recommended browsers, viewers, and plugins.
·
ARCTOS
http://messene.nmsu.edu/ursa/arctos (accessed April 16, 2001)
An interactive search engine that illustrates a selection interface that shows how thumbnail images can be used to support selection without knowing the document's language.
|
|
However, thumbnails will not work with text-only browsers and is not suitable for blind users. The system offers one search mode only and the number of results to be retrieved is fixed.
· ERIC
http://searcheric.org/ (accessed April 16, 2001)
The Educational Resources Information Center (ERIC) Clearinghouse on Assessment and Evaluation supports Spanish or German queries for English documents, and then translates the results into the query language. The user interface is available in English, German and Spanish.
 |
The number of results to be retrieved is under user control. The system provide the thesaurus based search as an option (with a wizard to facilitate the searching process.) The site does not offer a text-only version.
·
Strategis
http://strategis.ic.gc.ca/ (accessed April 16, 2001)
A bilingual system that searches a large collection of English and French documents using either English or French queries. When the query translation option is selected, the user is prompted to select the desired translation from a list of available translations for several technical domains.
 |
The interface is provided in both English and French. The site does not offer a text-only version. However, all images have the ‘alt’ tag.
·
ArabVista
http://www.arabvista.com (accessed April 16, 2001)
A search engine in which users enter a query in either Arabic or English and the actual search is done in many languages. The site is an example of supporting right-to-left and left-to-right languages. The site provide a virtual keyboard (an applet) to support the Arabic language.
|
|
The site also offers: multiple search modes, interface in both English and Arabic. No text-only version is provided. (A complete English version can be found by following the link to Emirates Internet (http://www.emirates.net.ae/) )
The unprecedented international exposure afforded by the Web increases the designer’s responsibility for insuring universal usability. To search a collection of documents in a foreign language, more needs to be done than merely retrieving foreign documents relevant to the query posed in the native language. There is ongoing research in the area of Cross-Language Speech Retrieval [11-13], in which speech documents are retrieved in response to user queries specified in text of another language. This opens a new area for universal usability issues of these systems.
The usage of the Extensible Markup Language (XML) to semantically transcode existing Web contents [6] and using that to adapt to different users need is still an open area for research.
[1] Douglas W. Oard, 'Evaluating Interactive Cross-Language Information Retrieval: Document Selection,' in Proceedings of the first Cross Language Evaluation Forum, Lisbon, September 2000.
http://www.clis.umd.edu/dlrg/filter/papers/clef2.ps (accessed April 16, 2001)
[2] Douglas W. Oard, Anne R. Diekema, ‘Cross-Language Information Retrieval,’ Annual Review of Information Science and Technology (ARIST), Volume 33, 1998.
[3] G. Perlman, ‘Web-Based Bibliographic and Full Text Retrieval: Which Features/Which Users,” Slides available online at:
http://www.acm.org/~perlman/patronlib.ppt (accessed April 16, 2001)
[4] G. Perlman, ‘The FirstSearch User Interface Architecture: Universal Access for any User, in many Languages, on any Platform,’ ACM Conference on Universal Usability, Arlington, Virginia, USA, November 16-17, 2000.
http://www.kric.ac.kr:8080/pubs/articles/proceedings/chi/355460/p1-perlman/p1-perlman.pdf (accessed April 16, 2001)
[5] G. Perlman, ‘The FirstSearch User Interface Architecture: Universal Access for any User, in many Languages, on any Platform,’ Slides available online at:
http://www.acm.org/~perlman/fsmulti-cuu.ppt (accessed April 16, 2001)
[6] Anita W. Huang, ‘Aurora: A Conceptual Model for Web-Content Adaptation to Support the Universal Usability of Web-based Services,’ Proceedings on the conference on universal usability, Washington, DC United States, November 16 - 17, 2000.
http://www.kric.ac.kr:8080/pubs/articles/proceedings/chi/355460/p124-huang/p124-huang.pdf (accessed April 16, 2001)
[7] Tony Fernandes, ‘Global Interface Design,’ Academic Press, 1995.
[8] Harry Hochheiser and Ben Shneiderman, ‘Universal Usability Statements: Marking the Trail for All Users,’ ACM Interactions, 8(2), March-April 2001.
http://www.kric.ac.kr:8080/pubs/citations/journals/interactions/2001-8-2/p16-hochheiser/p16-hochheiser.pdf (accessed April 16, 2001)
[9] Debbie Hysell and Gray
Perlman, ‘Lessons Learned from Internationalizing a Global Resource,’ 1st
International Workshop on Internationalization of Products & Systems,
Rochester, New York, USA, 21-22 May 1999.
http://www.acm.org/~perlman/fstrans.html (accessed April 16, 2001)
[10] G. Perlman, ‘CHI 99 SIG: Universal Web Access: Delivering Services to Everyone,’ SIGCHI Bulletin, 1999, 31:4, 53-54.
http://www.acm.org/~perlman/access/ (accessed April 16, 2001)
[11] Douglas W. Oard, 'Speech-Based Information Retrieval for Digital Libraries,’ Technical Report, University of Maryland at College Park.
http://www.clis.umd.edu/dlrg/filter/papers/speech/paper.html (accessed April 16, 2001)
[12] Helen M. Meng, 'Intelligent Speech for Information Systems: Towards Biliteracy and Trilingualism,’ Proceedings on the conference on universal usability, Washington, DC United States, November 16 - 17, 2000.
http://www.kric.ac.kr:8080/pubs/articles/proceedings/chi/355460/p91-meng/p91-meng.pdf (accessed April 16, 2001)
[13] Paraic Sheridan, Martin Wechsler, and Peter Schauble, ‘Cross-Language Speech Retrieval,’ Proceedings of the 20th International ACM SIGIR Conference on Research and Development in Information Retrieval, pages 99--109. New York: ACM, July 1997
http://citeseer.nj.nec.com/89636.html (accessed April 16, 2001)
[14] Web Accessibility Initiative (WAI)
http://www.w3.org/WAI/ (accessed April 16, 2001)
A good place to look for information when designing for accessibility. It also contains a guidelines section.
[15] ‘Microsoft Accessible Web design guidelines‘
http://www.microsoft.com/enable/dev/web/howto.htm (accessed April 16, 2001)
Includes information about HTML tags that supports accessibility.
[16] Universal Usability.org , ‘Guide to Universal Usability’
http://www.universalusability.org/ (accessed April 16, 2001)
A good start to issues regarding universal usability. It also offers a template for a Universal Usability Statement.
[17] ‘Universal Usability in Practice’
http://www.otal.umd.edu/uupractice/ (accessed April 16, 2001)
This site offers a practical guide for different universal usability issues.
Moustafa
A. Youssef
Department of Computer Science
University of Maryland
College Park, MD 20742 USA
moustafa@cs.umd.edu
April 2001
