2001 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2001 Table of Contents


IMPLEMENTING A WEB ACCESSIBLE DATABASE

Mrs E.A.B. Draffan and Mr Robbie Corbett,
Assistive Technology Centre, University of Sussex, UK.

Abstract

This paper discusses the development and implementation of the National Internet Accessibility Database (NIAD), how the design of the database was based on ease of use by both its target audience and those working on the database, and also the approaches taken to ensure the successful implementation and launch of NIAD.

Introduction

The NIAD has been developed for those who have interests in assistive, adaptive and enabling technologies in the United Kingdom Further and Higher education sectors, as part of the Disability and Information Systems in Higher Education (DISinHE) Project. The NIAD aims to provide an on-line resource of information about products available to assist those with disabilities and/or special educational needs. In essence, the NIAD is a product guide with web links to external sites, and relevant support pages related to study needs rather than to daily living aids.

Background

The project team, funded by the Joint Information Services Systems Committee (JISC), began its work in 1999 and 'went live' in July of 2000. The idea, however, had been discussed many years earlier when colleagues in the assistive technology field admitted their dependency on American databases, such as Abledata and Closing the Gap. A sizeable collection of printed and electronic material had been collected over the years and the driving motivation was to collate this material, update the content and provide an on-line system for the education community in the UK.

Preliminary Planning

Koch (2000) describes the type of gateway the NIAD team planned to build as an internet-service that applies "a rich set of quality measures to support systematic resource discovery". He goes on to highlight several factors that make up this type of gateway:

A number of sources were referenced, primarily the Development of a European Service for Information on Research and Education (DESIRE) the efforts of which are geared towards the process of implementing information gateways. DESIRE provided some extremely useful guidelines but these rarely touched on accessibility issues, in terms of a user or a researcher with a disability e.g. making information accessible by the use of assistive technology such as screen readers. The project team soon realised how much 'manual effort' was required to find and accurately describe products within the criterion laid down. Less time-consuming methods had to be investigated to collect metadata and set in place checking systems to ensure the currency of the information. Categories, thesauri and search systems were considered, and accessibility options along with 'future proofing' mechanisms were designed. A web site was created to provide information about the project and Gantt charts with deadlines, targets and goals were used in our time management planning.

Team Building and Team Skills The DESIRE handbook has a 'Skills and people checklist', aimed at a large team. Given the limited availability of staff, the NIAD team shared the tasks they were capable of carrying out in-house:

A librarian's cataloguing skills should be considered as vital in this type of project especially if the research starts from a considerable amount of un-indexed data. Classification and categorisation of the data could not have occurred without this type of expertise and so this task was sourced externally.

There were insufficient funds to employ full time cataloguers so support was found in the student community, which was beneficial from a fiscal point of view but meant that there was a lack of continuity at times.

Kleim & Ludin (1992) cite the many reasons why project managers fail to fulfil the aims of the project despite the fact that they have "a wide array of project management tools and techniques at their disposal." "What project managers fail to realise is that their handling of people affects the outcome of their projects. Indeed, their neglect or mismanagement of people can affect schedule, cost and quality."

To help with project management meetings continued throughout the year and an Internet based 'chat forum' was designed to encourage open discussion, collaboration on policy, technical definitions resulting in clarification of protocols as well as reformulating the team's aims and objectives. The tools used for the forum were the same as those used for the NIAD's 'Tips and Tricks'. The latter is a modified one-way only message board provided by the NIAD editorial team for users.

Selection, classification and resource discovery for the database. "The problem with designing databases to answer specific or targeted questions is that invariably questions are left out, change over time, or even become superseded by other questions. Once this happens, a database designed solely to answer the original questions becomes useless. In contrast, if the database is designed by collecting all of the information that an individual or organisation uses to address a particular problem or objective, the information to answer any question involving that problem of objective can theoretically be addressed." (Ashenfelter, 1999) Addressing this requirement became an objective before we even began to decide how to set up the database.

The Project's Librarian reported that the Dewey Universal Decimal system offered possibilities for the NIAD (Jenkins 1999). However the notation is such that it would not transfer easily to a database. After visits to a number of charitable organisations, that used general library techniques with paper-based systems, it was decided that a simple tailor-made index system was the only option for the NIAD. Simplicity was the most important criteria but the team were warned that this would mean that the structure would not be as efficient at the retrieval time as it would be when filing the original data. The latter was a constraint that had to be overcome when developing the web-based system from the same material. Classification eventually occurred on the basis of functional difficulties e.g. visual impairment and then categories of technology types followed by sub-categories e.g software such as brainstorming/mindmapping, or audio equipment, such as loop systems. The file-based system eventually mirrored the relational database system using a unique identifier.

Choice of tools and design issues Having spent time looking at various information handling procedures, it was vital to make the right choice of relational database that could have its contents published on to the Internet. Having investigated a number of 'off the shelf' systems, it was realised that we would have to design the infrastructure in-house using Microsoft Access database, which was chosen for three reasons:

  1. The project team already had experience with Microsoft Access;

  2. Microsoft Access integrates flawlessly with the web development software (see below), thus reducing development time by utilising skills already learnt;

  3. Upsizing to Microsoft SQL Server is a simple task, requiring little or no modification of the underlying source.
Server choice was based on what was already available, the existing expertise, and access to the Joint Academic Network, through the university. The decision to use Allaire's ColdFusion (CF) as the active server web development environment stemmed from the available in house skills of the team. CF frees the web developer to produce dynamic web applications by using a flexible server-side mark-up language, which integrates with standard Hyper Text Mark-up Language (HTML). However, the challenge for the developer has been to make the database not only universally accessible but also capable of operating on multiple platforms and browser types.

CF's advantage, from an accessibility point of view, is that the user (on the client side) is sent pure HTML. No plug-in or additional software is required to access the HTML pages created with CF. Indeed, there is a seamless integration between a static, non-interactive web page, and a 'potentially' interactive page created using Cold Fusion, since they look and feel the same. This fitted the project team's aim that the design of the product and environment should be usable by all people, to the greatest extent possible, without the need for adaptation. (Connell et al, 1995).

It was felt that for the application interface (the web front-end) to be a success it would have to be:

Information collected for the database had to be validated, and verified by reviews, which were often included on the catalogue pages. A product formed the core of each record to which pictures, web links and other elements were added. Basic searches, advanced searches and browsing are classic parts of an information gateway and these sections had to be easy to use. Queries were designed to form the search engine for the NIAD with an underlying use of Verity '97 technology. Each search is fully customisable by the user in real time.

Having investigated various options, the decision was made to base the metadata layout on the Dublin Core, as the field types fitted well with the type of data the NIAD contained. In fact an on-line 'meta-grabber' was designed so that researchers could use this method to easily add general web links directly to the site.

Maintenance

The maintenance of any database requires systems and scope procedures to ensure that quality control is in place. These protocols are vital for handing over research to new staff members as well as for reminding those more familiar with the systems in place. Thus a series of checklists were developed with efficient on-line support strategies (collectively known as the NIAD Administration). This approach not only ensured they were accessible, as they used the same technologies that were being used with the back-end database, but the researcher could also work from any machine that had an Internet connection.

As the project grew, so did the reliance on the Internet-based Administration, which was adapted, expanded and modified in-house to meet changing needs. These advancements actually resulted in features that had been designed for the development team, eventually making their way into the final application because they were so useful e.g. searching by company and those that supply specific items for various functional difficulties.

Evaluation

Various evaluation strategies were set in place, including off-line paper-based material as well as interactive features on the NIAD website.

The aims and objectives of the evaluation were to:

The on-line evaluation is based on a System Usability Scale (DEC 1986) to provide standardised statistical evidence over a period of time. This scale covers four out of the five characteristics linked to 'Usability': ease of learning, use, memorability and subjective satisfaction. However, it does not check for 'error frequency' and this is something that will need to be investigated through interviews. Nielson (1998) alerts us to another interesting point indicating that there may be a case for saying that no more than five users are required to test for usability.

Retrospective

Much time has been spent on the subject of accessibility and web page design, while less has been said about the processes and support mechanisms for those building databases and on-line information systems. We have learnt that web based administrative tools are as important to the success of this type of project as the focus that is normally aimed at the application itself.

As the NIAD's database grows, so the complexities of search engines, metadata and methods of automatically monitoring data increase. Metadata vocabularies to introduce quality ratings for automatically searched data have been researched, and if more time were allocated to this aspect of the database, these methods may eventually be very productive and save on researchers' time.

Statistics show that only 16.2% of projects are completed on time and within budget ("Charting the Seas of Information Technology" The Standish Group 1994). The NIAD project was ready on time within its budget and can be found at http://niad.disinhe.ac.uk26

References

Abledata http://www.abledata.com 

Ashenfelter, J.P. (1999) Designing Good Databases, Webreview.com http://www.webreview.com/pub/1999/03/26/feature/inex2.html

Brooke, J. (1986) System Usability Scale (SUS): A Quick-and-Dirty Method of System Evaluation User Information, Digital Equipment Co Ltd, Reading, UK

Closing the Gap http://www.closingthegap.com

Connell, B.R. et al (1995) The Principles Of Universal Design

Corbett, R., (2000) Working Paper on Audit protocols for NIAD, University of Sussex

DESIRE handbook, Development of a European Service for Information on Research and Education (DESIRE) Information Gateways handbook http://www.desire.org/handbook/welcome.html 

Disability and Information Systems in Higher Education (DIS-IN-HE) http://www.disinhe.ac.uk 

Dublin Core, http://purl.oclc.org/dc 

Jenkins, S. (1999) An Enquiry into the Organisation of Special Collections: Designing a system for the ATC. Unpublished work for MA Information studies; University of Brighton.

Kleim, R. & Ludin, S. (1992) Introduction, The world of project management, from The people side of project management, pp1-36. Gower, Aldershot

Koch, T. (2000) Quality-controlled subject gateways: definitions, typologies, empirical overview. Manuscript of the article published in the Subject gateways special issue of "Online Information Review" Vol. 24:1, Feb 2000. MCB Univ. Press

Nielson, J. (2000) Why You only Need to Test with 5 Users, Useit.com Alertboxhttp://www.useit.com/alertbox/20000319.htmlAlertboxhttp://www.useit.com/alertbox/20000319.html

Slater, D.(1999) Comparison of Database Systems Produced for the NIAD team. University of Sussex.http://www.atc.sussex.ac.uk/cfdocs/niad/baseline<


Go to previous article 
Go to next article 
Return to 2001 Table of Contents 
Return to Table of Proceedings


Reprinted with author(s) permission. Author(s) retain copyright.