BSCW for Disabled Teleworkers

Usability Evaluation and Interface Adaptation of an internet-based Cooperation Environment

Michael Pieper
GMD - German National Research Center for Information Technology
Institute for Applied Information Technology
Human-Computer Interaction Research Division
Schloss Birlinghoven
53754 Sankt Augustin, Germany
michael.pieper@gmd.de

Dirk Hermsdorf
Louisiana State University
298 Coates Hall - Baton Rouge
LA 70803, USA
dirkherm@bit.csc.lsu.edu

Abstract

TEDIS (Teleworking for Disabled People) is a research and development project of the research group on Human Computer Interaction at the German National Research Center for Information Technology (GMD). TEDIS is an assistive technology contribution to the promotional program 'Telecooperation - Value Added Services' of the German Federal Ministry for Education, Science, Research and Technology (BMBF) and accordingly funded. The general goal of the project is to implement a generic human computer interface for accessing Internet, which can be adapted to a variety of different needs of handicapped as well as elderly people. As part of a field trial, the internet based telecooperation environment BSCW (Basic Support for Cooperative Work) was installed to manage the teleworking process for two severely physically disabled teleworkers. At present, BSCW is adjusted to their special needs based upon data gained by structured usability-interviews. As a result, complete accessibility of BSCW by keyboard will soon be available, since operating a mouse causes many problems for motor-disabled, blind or visually impaired end users.

1. Introduction

Actually, 10% of the whole worlds population falls into the group elderly or handicapped. This number is based on the [BUEH96] definition of handicapped and the convention to count everyone about 60 years old as an elderly person. In Europe at present, there are about 100 million elderly and 50 million handicapped people. Recent official surveys reveal 6.4 million severely handicapped people in Germany alone. This tendency increases. Modern telecommunication technologies offer many possibilties for special needs adjustment. Thus severely handicapped as well as elderly people can be enabled to live a self-determined life and take part in social and economic affairs. For example, telecommunication technology can compensate for an often prevailing lack of physical mobility, which is detrimental to the vocational integration of handicapped people.

Unfortunately technological progress does not necessarily imply enhancement of applicability of information technology. In general, special needs of handicapped end users are not equally taken into consideration by system designers. Handicapped people are not usually included in design processes. Therefore, products are often inaccessible due to small but, with regard to certain disabilities, crucial shortcomings. In order to ensure the suitability of the teleworkstation developed within the TEDIS project, principles of participative systems design are applied. This refers to technical aspects as well as issues of usability. For instance, the Internet-based telecooperation environment BSCW is adjusted based upon data gained by structured usability-interviews [GAPP95].

2. The TEDIS field-trial

The adaptation of the teleworkstation is realized in cooperation with the FTB (the Forschungsinstitut Technologie-Behindertenhilfe in Volmarstein) an internationally renowned rehabilitation center in Germany. Two physically handicapped end users living in Dortmund, which is located 65 km away from Volmarstein, work for the FTB administration with support of BSCW (Basic Support for Cooperative Work), a telecooperation environment which has been developed by GMDís research group on Computer Supported Cooperative Work (CSCW). BSCW is available free of charge via common Internet browsers in the World Wide Web [URL01].

3. BSCW

The BSCW (Basic Support for Cooperative Work) project at GMD FIT is developing tools to support cooperative work over the Web. The basis for this work is the BSCW 'shared workspace' system - an extension to a standard Web server which supports document upload, event notification, group management, communication and much more. BSCW provides facilities for collaboration over the Internet. It runs across the most commonly used platforms on PC, Macintosh and Sun. This serves as an integration platform onto which a variety of CSCW applications can be added. The emphasis is on integrating existing tools, rather than constructing new ones [BENT95, BENT96, APPE96].

The BSCW system is based on the 'shared workspace' metaphor: an object store for group work, with some simple awareness functionality that allows users to keep an overview of what is happening in the workspace. A workspace user can browse through a shared workspace (if he or she has permission to do so) with an unmodified WWW client on any platform. The BSCW workspace allows one to (over)write objects by means of a small 'helper' application that is provided with the system.

4. Usability

Before starting the actual telework by using BSCW, a usability test was conducted in order to evaluate the user front-end of that telecooperation environment. The goal was to investigate whether the two teleworkers had learned to handle the system successfully as well as to assess the usability of BSCW for end users with special needs, since BSCW was not originally designed with regard to the special needs of disabled end users.

4.1 Methodology

Methodologically usability investigations followed two different approaches:

• The first approach relied on active participation of the two end users, who had to perform a predefined standard-task. Task accomplishment was then followed up by a structured interview to discover certain problems the disabled pilot-users encountered while using the system.
  


• The second approach relied on already existing guidelines [ACT96, ATRC95, ATRC96, DOIT96, FONT95, INFO95, THOR93, TRAC96, TRSE95] which focus on interface design for people with special needs. A special category of these guidelines related to motor-impaired people were taken to evaluate corresponding special needs usability of the user front-end of the BSCW system.
  
  

4.1.1 Usability Evaluation by Standard Task and Structured Interviews

The predefined standard task we asked the teleworkers to perform, was divided into three subsequent parts. During their course the main functional body of BSCW was tested. The content of the task related to results of our initial field exploration comprising a description of the social environment, the pilot-users educational background, their computer literacy, their attitudes towards telework and the like. Pilot-users were asked to validate corresponding findings by modifying research documentation and commenting on these modifications within a common BSCW telecooperation environment especially devoted to these purposes. Therefore accomplishment of the predefined task aiming at testing system usability can as well be seen to be crucial for the participative research and development approach, which aims at involving end users in an intensive dialogue on research validation during the process of designing technical and organizational systems adaptation.

The guideline for the interview, which was conducted after the teleworkers had successfully completed the given tasks, was developed in accordance to the software-ergonomic criteria listed within EVADIS II [EVAD92]. EVADIS II is a software-supported approach developed by GMD´s Department for Human Computer Interaction specifically to serve as a tool for evaluating dialogue systems. Most basically this evaluation procedure relates to the IFIP-Model of Human-Computer Interfaces, which already for years is being promoted by the International Federation for Information Processing, an international parent organization of national computer science companies.

According to the IFIP-model, the end user interacts with a computer system through at least four interface layers, the terminal (input/output)-, the dialogue-, the functional- and the organizational interface. The terminal interface regulates the way work data, parameters and commands are input and vice versa the way how system feedback is presented to the user. The dialogue interface regulates the nature and form of dialogue processes (i.e. command-, menue-, icon driven etc.), system aids, error handling and intervention options in the course of the dialogue. The functional interface regulates the handling of system tools (i.e. applications). The organizational interface reveals the relations between different end users within the same organizational or socio-technical context, which is for instance defined by the system borders of a telecooperation system. However, the structured interview was more closely focussed on the input/output - ,the dialogue- and the functional layer of the overall user interface.

Another important orientation aid of the EVADIS II evaluation procedure is a standard developed by the German National Standard Organization (DIN 66.234, part 8), which was much discussed in specialized circles. This DIN categories, which in a two factorial design can be combined orthogonally with the four interface layers of the IFIP-model, are in the form of explicit recommendations for the software-ergonomic design of the overall user-interface [OPPE89].

4.1.2 Findings

The following issues, which are only more or less intrusive examples for the layered evaluation approach outlined above, were among others identified as demanding for adapation:

• terminal (input/output) interface
  The cursor position serves within BSCW as a prompt for input. Yet, the cursor is not always positioned at the correct place for input, which - by the way - is more a general problem of the Internet browser than of BSCW itself. Moreover, it is not possible to keep windows open after having completed the input for instance by offering a 'More?' option, thus the windows need to be opened again and again. Considering that people with functional impairments in the upper limbs depend upon optimizing the required mouse-movements as well as keeping the number of characters to be typed as little as possible in order to work efficiently, software adaptations are necessary at this point.
  These findings interfere with the DIN-recommendation of Problem appropriatenes, which demands that the user's actual task should be supported and not made more difficult by specific system properties.

• dialogue interfaces
  Some of the icons presented by BSCW as part of the graphical user interface did not imply the action, which is set off by them. Furthermore, some abbreviations, for example 'des' for 'description' were not comprehensible by the teleworkers. These issues are pointing at an underlying more fundamental problem, that is the lacking consideration of the native language of the end user. Most of the Websites available in the Internet respectively World Wide Web are presented in English. This causes a crucial obstacle in accessibility of the WWW for interested end users with little knowledge of English, which is particularly true for elderly and disabled end users. Most basically this is a policy issue of information politics which needs to be solved on a fundamental level for instance by offering versions adjustable to at least the most prevailing languages or by provision of translational assistance.
  Again these findings interfere with another DIN-recommendation denominated Transparency or Self descriptiveness. These criteria refer to the fact that the user should be able to 'see through' the system. Thus, the dialogue must be either directly comprehensible or, if this is not the case, the system should be able on demand to explain to the user the dialogue's purpose and method. Where the dialogue is not directly comprehensible, the user should, on request, be given a list of the available facilities (e.g., switch to native language or translation help), so that he has a good idea of the system context for his own task.
  


• functional interface
  Furthermore, he disabled teleworkers suggested the enhancement of the BSCW-system by offering the possibility of reading in input by cache memory, for instance, in case of long e-mail addresses, so that, as a side effect, on-line times can also be kept as short as possible in order to work economically.
  The DIN recommendation of Flexibility matches with this end users suggestion. It requires that software tools be adapted as far as possible to suit the users' individual wishes, experience, etc. Account should be taken of the fact that user control of a system grows gradually with use. The user should be able to influence the dialogue process, its speed, including interruptions, and the order of various dialogue operations.
  
  

4.1.2 Usability Evaluation by Guidelines

In order to achive Web accessibility for end users with special needs, it is necessary to distinguish between accessibility on the browser level and the HTML level. Nearly all of the guidelines related to Web design for people with special needs are HTML design guidelines. These guidelines can be divided into three groups:

• General Guidelines
• Guidelines regarding physically handicapped end users
• Guidelines regarding visually impaired and blind people
  

Hearing impairment is not mentioned because up to now acoustical elements play a less important role in Web documents and in general, hearing-impaired people have no special problems when they navigate through the Web.

General guidelines relate to demands that, for instance "Every (critical) object in an HTML document should be at least available in an alternative text version" (a critical object may then be defined as "if an object (picture, icon, button, table, image map, audio file, video, Java applet) is cut out of an HTML document and the left HTML document makes not much sense any more, the object is critical"). Other general demands are that "Web documents should be tested using different browsers on different platforms", that "HTML tags which are not supported by all common browsers-" and "detailed background images, requiring much download time should be avoided", that "a delete form contents function should be provided, when form elements are used" and the like.

Most important, however, for the purposes of the BSCW usability evaluation related to the adaptation of the teleworkstations of the two motor-impaired end users in Dortmund are the already existing guidelines regarding physically handicapped end users. These guidelines will be listed one after another with additional comments for each guideline followed up by some guideline-related illustrations of the BSCW user front-end:

• Buttons should be at least one inch long and a half-inch high.
  
  Comment:
  The main characteristic of physically handicapped people using computers is that they have problems with the mouse. For that reason the buttons have to be big enough so that they can hit them.
  
• Try to avoid the double click method of selecting objects.
  
  Comment:
  The double click is hard to perform for a lot of physically handicapped people although it is possible to adjust the time between the double clicks in the Windows system. Fortunately, all objects except Java applets do not use the double click method anyway.
  
• When text is to be typed in by the user, the text to be typed should be as small as possible
  
  Comment:
  Writing long texts using the keyboard is hard for physically handicapped people. When a multiple-choice question is asked, the user should be able to check the corresponding answer or to type in a corresponding key. The key-character might be underlined in order to be identified
  
• In form elements, defaults should be available which allow the user eventually to confirm them.BR>
  Comment:
  Guidelines regarding visually impaired and blind people just claim the opposite, because Braille or speech output of already filled-in form-filling requests might confuse them. In order to solve this problem, the HTML document should provide a checkbox to turn on/off option values in form elements.
  
• Avoid answers with time dependency.
  
  Comment:
  Physically handicapped people might need more time to answer requests.
  
  

4.1.3 Findings related to the BSCW user front-end

5. Interface adaptation for motor-impaired end users

When a Web designer pays attention to the HTML guidelines mentioned above, the resulting Web document is better, but still not perfectly accessible for motor-impaired end users. The reason for this is that some adaptations - and that is especially true for interface adaptations for the motor-impaired - can not be realized on HTML but on browser level. The most comprehensive example for such adaptations is to enable total keyboard access of a Web browser for the benefit of physially handicapped people who wouldn't have to use the mouse anymore. For physically handicapped people, the keyboard is a much better input device. Furthermore, there are special keyboards with extra l;arge or small keys for people with cerebral palsy or muscular dystrophy, as it is the case for the BSCW end users of the TEDIS field trial.

5.1 Complete Internet access by keyboard

Therefore, in general navigation through BSCW can be eased dramatically for end users with physical handicaps by replacing cursor positioning by mouse, requiring good fine-motor coordination with navigation access by keyboard. TEDIS currently realizes complete access to an Internet browser by keyboard. However, such accessibility options are already partly realized on the OS level upon which certain browsers can be installed. The following adaptations for for physically handicapped end users, for instance, are already implemented into the 'Control Panel/Accessibility' options or 'Control Panel/Mouse' options of Windows 95:

• FilterKeys
• StickyKeys
• Keyboard mouse
• Adjust speed of mouse arrow
• Adjust time between the double click

FilterKeys means that after the user has hit any key some specified time has to pass before the next key can be hit. This is useful, because people with impairment in the upper limbs might perform uncontrolled movements which can be ignored using this feature. Additionally, the time between characters to be displayed can be specified, when one key remains pressed. Without this feature, a character is displayed serveral times, because the person cannot remove his finger quick enough from the key.

Regarding key-combinations like CtrlA, StickyKeys makes it possible for the user to press the required keys one after another. Therefore, the system just stores the corresponding information, when a key like Ctrl, Alt or AltGraph is pressed.

These and the other mentioned adaptions realized in Windows 95 resulted mainly from the the special legal conditions in the USA implied by the 1990 'American with Disabilities Act' (ADA). ADA claims free access to public places like theaters, restaurants and museums until the year 2000. Community and state buildings, public transportation services and telecommunications services are also affected by this law. Microsoft however realized some of these adaptations only after the state of Massachusetts threatened to boycott all Microsoft products if no adaptations would be made [HELI96].

Additional browser adaptations for the motor-impaired, which as far as we know are up to now (with one exception) not realized in common Web browsers, are currently under development by the TEDIS project:

• Approaching-tolerance of cursor positionning
• Tabbing through forms and checkboxes of an HTML document
• Providing a keyboard command in order to mark text
• Selecting a link by typing the first letter of the link's label
• Determining a special keyboard command in order to tab with a changeable step range through links
• Recognition of words after having typed in only some characters (definition of macros, provide dictionary, etc.)

Approaching tolerance realizes that an icon, button, link the like is activated, when the user executes a function and the cursor is close to but not necessarily exactly located upon a certain object. Every time, the object which is the closest to the pointer is marked so that the user knows which object is selected when a function is executed.

The other listed browser adaptations should be self explanatory. NCSA Mosaic 2.0 allows one to tab through links using the arrow keys and to select a link by RETURN. Lynx and www are browsers which provide the same functionality as NCSA Mosaic 2.0 [URL02]. However, they are pure text browsers and a physically handicapped end-users probably does not want to abandon all graphical objects and elements in an HTML document.

The adaptations currently under development by the TEDIS project will be implemented by using 'MS Internet Control Pack' [URL03], especially the HTML Control in combination with Visual Basic 4.0. Alltogether, the Internet Control Pack comprises eight OLE controls which provide a lot of useful methods when a developer wants to create an Internet application with Visual C++, Visual Basic, Fox Pro or Microsoft Access. However, while working with these tools it became obvious that some of the mentioned adaptations can not be realized because of the restricted HTML control functionality. The current version of the Internet Control Pack (beta 2), for instance, does not allow one to mark text in an HTML document. A solution to overcome this and other shortcomings would be to modify the source code of an existing Web browser. Unfortunately, the source code of common browsers is hardly available. Currently TEDIS considers modification of the HotJava browser source code which should be available as public domain software by the W3-consortium.

6. The generic approach

In the long run, development efforts within the scope of TEDIS seek to develop a generic human-computer interface for a reliable Internet browser. The concept of generic human computer interfaces most basically means of implementing a software layer between the user´s front-end and the application, which can easily and universally be adapted to a variety of different needs of handicapped people and other end users with special needs [STEP91].

To at least partially comply in the short run with the demand of implementing a generic interface for Internet access, besides the adaptations for motor-impaired end users the following adaptations of WWW front-ends (like BSCW) to visual impairments and blindness are taken into consideration:

• Labeling of listings (size of list, number of sublists, numbering of list items)
• Definition of a browser-transformation table like & means 'and', $ means 'dollar' and the like, because screen readers might have problems with symbols like §, $, % and &
• Use of natural speech in case of an error message
• Space insertion betwen each character in abbreviations like SLAC, GMD etc., so that a screen reader may pause between each character and will not try to read the whole abbreviation as one interconnected word.
• Adding dots to the end of list items, so that synthetic speech output lowers the sound of speech at that point, thereby indicating the end of the list item
• Listing of all Links incorporated into a HTML-document (by URL) to get a quick overview or impression of a document
• Conversion of tables into sequential representations
• Listing of all 'Header x' titles in HTML code
• Keywords like 'Link', 'Radiobutton' etc. at corresponding locations within HTML documents

The approach of modifying HTML layout for the visually impaired has already been realizied in the WAB (Web Access for the Blind, [WAB]) project. If one specifies the proxy server [URL04], the HTML layout will change according to some of the interface adaptations listed above. One disadvantage of WAB is that it is not possible to use only some of its features. Furthermore, all data requested by the browser have to pass the server in Switzerland, which leads to a further time delay in a Web, which is still slow enough. Another disadvantage is that in case of a crash of the server in Switzerland the whole browser will be blocked.

References

[ACT96] ACT Center, 'Accessible Web Page Design', http://www.igs.net/~starling/acc/index.htm

[APPE96] Appelt, Wolfgang; Busbach, Uwe: The BSCW System: A WWW based Application to Support Cooperation of Distributed Groups (postscript, 914 K), presented at WET ICE 96: Collaborating on the Internet: The World-Wide Web and Beyond, Stanford University, June 19-21, 1996

[ATRC95] Adaptive Technology Research Center, University of Toronto, 'Guide to Accessible HTML, Accessibility for Persons with Visual Disabilities', http://www.utoronto.ca/atrc/rd/html/html.html

[ATRC96] Adaptive Technology Research Center, University of Toronto, 'Making the Web Accessible for the Blind and Visually Impaired', Mike Paciello, http://www.yuri.org/Webable/mp-blnax.html

[BENT96] Bentley, Richard, Busbach, Uwe and Sikkel, Klaas: The architecture of the BSCW Shared Workspace System in Proc. 5^th ERCIM/W4G Workshop "CSCW and the Web", Arbeitspapiere der GMD 984, GMD, Sankt Augustin, April 1996, pp. 31-42, http://orgwis.gmd.de/projects/W4G/proceedings/bscw.html

[BUEH96] Christian Buehler (FTB), 'STOA: Scientific And Technological Options Assessments. The Information Society. Participation of Disabled and Elderly in the Information Society', Luxemburg, Februar 1996, PE: 165.285

[DOIT96] DO - IT (Disabilities, Opportunities, Internetworking & Technology), 'DO - IT HTML Guidelines', http://Weber.u.washington.edu/~doit/Other/design.html

[EVAD92] R. Oppermann; B. Murchner; H. Reiterer; M. Koch: Softwareergonomische Evaluation. Der Leitfaden EVADIS II. (Mensch Computer Kommunikation 5/2), de Gruyter, Berlin: 1992

[FONT95] Paul Fontaine, Center for Information Technology Accommodation, General Services Administration, Washinghton, DC. USA, 1995, 'Writing Accessible HTML Documents', http://www.yuri.org/Webable/htmlcode.html

[GAPP95] H. Gappa; D. Hermsdorf; S. Mermet; M. Pieper:
Teleworking for Disabled People - GMD's TEDIS project, in: Österreichische Computergesellschaft (ed.): Interdisciplinary Aspects on Computers Helping People with Special Needs, Oldenbourg, München, Wien 1995, pp. 267 (ISBN 3-486-23797-7)

[HELI96] Journal 'HELIOSCOP', Herausgeber: HELIOS (Handicapped People in the European Community Living Independently in an Open Society), Edition Fruehjahr 1996 Nr.7, CE-CU-96-007-DE-C

[ICCH96] Oesterreichische Computer Gesellschaft, ICCHP '96, 5^th International Conference, Linz, Austria, July 1996, ISBN 3-486-23797-7, Henrike Gappa, Dirk Hermsdorf, Stefanie Mermet, Michael Pieper, 'Teleworking for Disabled People - GMD's TEDIS project', Page 267 - 276

[IGEL] Brochure from IGEL, Elektronische Kommunikationshilfen GmbH, Osterdeich 222, 28207 Berlin, Telefon 0421/413373, Fax 414824

[INFO95] InfoUse, Center for Accessible Technology (CforAT), 'The User Interaction Design Guidelines Document', 28 Dezember 1995, Version .955

[OPPE89] R. Oppermann; B. Murchner; M. Paetau; M. Pieper; H. Simm; I. Stellmacher:
Evaluation of Dialog Systems, GMD-Studien Nr. 169, Sankt Augustin,1989 (ISBN 3-88457-169-9)

[STEP91] C. Stephanidis; J. Ekberg; K. Nordby: Towards a Model of Terminal Design. In: S. v. Tetzchner (ed.): Issues in Telecommunication and Disability, Luxembourg, 1991, 141 - 156 (ISBN 92-826-3128-1)

[THOR93] Clas Thorn, 'Nordic guidelines for computer accessibility', ISBN 91-86954-15-6

[TRAC96] Trace R&D Center, University of Wisconsin - Madison, 'Design of HTML (Mosaic) Pages to Increase Their Accessibility to Users With Disabilities', http://www.trace.wisc.edu/text/guidelns/htmlgide/htmlgide.html of this paper

[TRSE95] Jutta Treviranus, Chris Serflek, University of Toronto, 'Alternative Access to the World Wide Web'

[WAB] Projekt 'Web Access for Blind Users', Institut fuer Informationssysteme, ETH (Eidgenoessische Technische Hochschule) Zuerich, http://www.inf.ethz.ch/department/IS/ea/blinds/

URL section: