The ADAM Project is creating a subject-based information gateway service that will provide access to quality-assured Internet resources in the following areas:
- Fine Art, including painting, prints and drawings, sculpture and other contemporary media including those using technology
- Design, including industrial, product, fashion, graphic, packaging, interior design
- Architecture, including town planning and landscape design, but excluding building construction
- Applied Arts, including textiles, ceramics, glass, metals, jewellery, furniture
- Media, including film, television, broadcasting, photography, animation,
- Theory, historical, philosophical and contextual studies relating to any other category
- Museum studies and conservation
- Professional Practice, related to any of the above
The 3-year JISC funding for ADAM was awarded to a consortium of 10 institutions, each with a vested interest in the creation of the service, as part of the Access to Network Resources initiative of the Electronic Libraries Programme.
Each of the 10 Consortium Partners brings something useful to the project; for example subject expertise, technical capabilities or project management skills. Many of the representatives from the Consortium Partners who comprise the Steering Group also represent professional associations, for example the Association of Art Historians (AAH), the Council for Schools of Art and Design (CoSAAD), and Computers and the History of Art (CHArt).
The Surrey Institute of Art & Design, represented on the Steering Group by the Project Director Marion Wilks, is the lead institution in the Consortium, and is also the host institution for two of the salaried team members; one of the Project's two Resource Officers Rebecca Bradshaw, and myself.
The University of Northumbria at Newcastle is represented on the Steering Group by Chris Bailey, who is also managing the Consultation Process, described in more detail in the following section. UNN also hosts the Digital AlphaServer on which the service will reside, and the Project's Technical Officer Mark Burrell.
Glasgow School of Art, represented on the Steering Group by Ian Monie, hosts the other Resource Officer, David Buri.
The other Steering Group members are Coventry University Art and Design Library, Middlesex University, The National Art Library at the Victoria & Albert Museum, Southampton Institute of Higher Education, The Tate Gallery, University of West of England and Winchester School of Art.
The Story So Far...
Significant progress has been made since the Project's official start date of 1st December 1995 - a basic service that allowed users to search a database of 50 resource records using the ROADS software was available long before the first warm day of Spring 1996!
A considerable amount of work has already been completed in order to reach this stage; for example, setting up the server, conducting research into emerging standards for resource description, cataloguing Internet resources and producing a project plan. As a result, we now have a very clear idea of the challenges that need to be met in the coming months, and are working effectively as a team despite being geographically remote from one another - the effectiveness of the communications between team members can be visibly demonstrated by the web pages describing the project, which were designed and produced collaboratively.
This remote team working (tele-teamworking?) is enabled through the use of a diverse range of communication media - Internet applications such as electronic mail, FTP, and TELNET are used in conjunction with more traditional tools such as the trusty telephone and the occasional face-to-face meeting to promote an effective collaborative working environment. Currently missing from the list of useful communications media, however, is the Royal Mail: 'Snail mail' not only lacks the immediacy of the other media, it is also a hindrance in our attempts to move wherever practical to a 'paperless' environment. This is both for environmental reasons, and in an attempt to reduce the administrative workloads and costs typically associated with paper-based filing systems. The Post Office don't need to worry about lost revenue just yet though; the Consultation Process, described below, will make extensive use of the postal service!
Where we're going
The following sections outline some of the issues and activities, in no particular order, that we are currently addressing.
As this article goes to press, we are embarking on the first major stage in a Consultation Process that will enable us to focus the service we create on the needs of the likely user community, to monitor the usefulness of the Service, and to create a framework for feedback from its users. The Consultation Process forms the backbone of our Evaluation Strategy, and its successful completion is a key Project Objective.
In this first stage, we are currently conducting a survey of the information needs of a wide range of practitioners in all of the subject areas covered by ADAM, which will yield information about their awareness of and access to electronic information resources, their current information searching strategies and their views on barriers to access to networked resources. Background information about the respondents is also gathered, to enable the data to be analysed according to discipline, job function and working environment.
A consultant was employed to design a detailed questionnaire and a list of recipients that will produce a meaningful sample. The list of recipients was made up from the mailing lists of various relevant associations and professional networks, staff and students at Consortium Member institutions, and lists of commercial practitioners compiled by third parties.
Up to 5,000 paper questionnaires will be distributed by May 10th. In order to encourage respondents to reply promptly, we have provided reply-paid envelopes and will be conducting a prize draw on 31st May, with a prize of £100 for the winner.
The data from the paper questionnaires will be supplemented by electronically collected responses - an electronic version of the form is available from the ADAM web site, and announcements about its existence will have been made in various appropriate newsgroups by the time this article goes to 'press'. Although the electronic version of the form was time-consuming to set up initially, it will provide an efficient and cost-effective means of supplementing our sample - postage and printing costs are eliminated, and data entry consists simply of importing the data, which is already in an electronic format, into the main survey database.
The use of a large number of paper questionnaires was considered vital in order to collect representative information - respondents who fill in the web version of the questionnaire will, by definition, already have some experience and ability in using networked resources. These responses can be analysed separately, and will therefore still be extremely useful in determining the likely needs of users who are already network-literate.
Another benefit that will be derived from the Consultation Process is to raise awareness of the service, paving the way for the Dissemination Programme that will take place after the official launch of the Service for public evaluation, currently scheduled for 26 September 1996.
Interoperability & Standards
Two key interrelated issues for any Access to Networked Resources projects are interoperability across different services, and the adoption of various standards that enables such interoperability to take place.
The use of standards affects almost every aspect of the Project's development and implementation, although in almost all key areas standards are still emerging and developing, and compromises are sometimes required.
For example, in designing the Project's web pages, a compromise was arrived at between the need for visual appeal (without which the Project risks the disdain of much of its potential user group!) and a rigid adherence to standards.
We took the view that the use of non-standard HTML 2.0 'extensions' in our web pages could be sanctioned, provided that no reduction in functionality was caused to users with more standards-conformant web browsers.
For example, we set the background colour of the pages to white because a non-graphical browser such as Lynx would simply ignore those tags. Conversely, we avoided the use of tables, because text formatted in tables is not represented clearly in certain browsers.
The pages were then tested extensively with a number of different browsers, on various platforms; Netscape, Mosaic, Internet Explorer and Lynx were used variously on UNIX/X-Windows, Macintosh, Windows 3.x and Windows 95.
The problems illustrated above pale into insignificance, however, when faced with a situation where there is no clearly defined standard, or a number of competing standards vying for universal adoption.
A particularly pertinent example of standards in flux for ANR projects are those for describing networked resources - how to catalogue the vast ocean of electronic stuff that is available over the Internet. These records are referred to as metadata, and there is currently much discussion globally about the best way to structure and communicate them.
The ADAM Project is currently using Internet Anonymous FTP Archive (IAFA) Templates, since this is the format designed to work with the ROADS software. A number of alternative metadata formats are now appearing, however; for example the Text Encoding Initiative (TEI) Headers, the Harvest Summary Object Interchange Format (SOIF) and the Dublin Core/Warwick Framework (discussion of these standards is thankfully beyond the scope of this article, but watch ADAM's web server for an Idiot's Guide to Metadata, due later this year).
This means that there's no guarantee that IAFA Templates will become the standard adopted by the wider 'net community. ADAM's approach to this problem will be to store our metadata records in a structured relational database, and to then create filters from this database that will be able to output records in a variety of metadata formats in response to queries from different systems. We are not planning to abandon ROADS, however, since we are keen to be compatible with the other ANR projects using the software, particularly once the WHOIS++ server functionality has been completed. Without the free availability of the ROADS software, and the experiences gained and shared by the ANR projects that preceded us, it would not have been possible to get the service up and running so quickly. It also has a number of other appealing features, for example the stemming algorithm used in the search engine is widely held to be the best for use with the English Language, and of course the software is available without cost!
We are also hoping to take part in the WHOIS++/Z39.50 Interoperability Trial, proposed by UKOLN. Z39.50 is an information retrieval protocol that exists as both an ISO and an ANSI standard, and there appears to be steadily growing support for it to be integrated into the WWW.
Vocabulary control is another application of standards that can become important as part of a cataloguing function. We are currently negotiating with the Getty Art History Information Program to secure the use of the Art & Architecture Thesaurus and the Union List of Artist's Names, both as tools to assist consistent data entry, and as tools to enhance information retrieval by providing the option to add synonymous terms to the search.
Some standards must be decided at the Project level, in response to the needs of the user community. For example, what quality criteria should be used to select the networked resources to catalogue? How comprehensively should web sites be catalogued? How should mirrors of catalogued resources be handled? Discussions on these topics and others are currently underway.
For the vast majority of the Service's users, the initial perception of the service will be largely based upon the web site's user interface. Of particular concern are those elements of the user interface that are used when searching for information held in the database.
One of the most interesting aspects of the project for me is the opportunity to carry out usability tests on a wide range of web-based search interfaces, to try and determine what information the users of the Service actually want, and what information retrieval capabilities and interfaces are the best for getting it, across a wide range of different users.
We plan to implement a whole range of search interfaces, and then invite users to attempt more detailed information retrieval exercises. This will take place in a structured way, for example through formal testing of volunteers under observation, and in a less detailed why, by analysing the log files on the server.
Initially we will be concentrating on user-friendly interfaces to handle boolean searches and the use of thesauri to expand searches to include synonyms, but if time and resources allow we will be looking at the feasibility of:
- 'Clustering' records into groups of 'like types' to allow browsing, as opposed to searching.
- The use of persistent 'recordsets' so that the user can build up a complex search in a series of simple steps.
- Ranking hits from a search in an attempt to present them in order of relevance to the user.
- Natural language processing, which allows the server to 'parse' or understand a search expression phrased in plain English. The additional use of semantic networks could also allow the data to be queried or ranked more intelligently.
- User profiling, to build up a knowledge-base about the preferences and habits of different users
- Communication with intelligent agents, should these become more prevalent on the Web.
A Drop in the Ocean
A fact that presumably every ANR project has to come to terms with at their outset is that it's impossible to thoroughly catalogue the contents of the Internet, no matter how hard you try. Attempting to catalogue ADAM's subject areas will be particularly frustrating, however, since they cover such a wide area and will tend to appeal to a wide range of users. New resources come on-line everyday.
In the long term, it is not practical to continue cataloguing networked resources by hand, except in very specialised cases. Some method needs to be found, therefore, that can automate, or at least speed up, the process.
The solution may lie in software technology; intelligent 'summarizer agents' may crawl the web, 'reading,' summarizing and cataloguing the resources automatically, sending back nicely formatted records to HQ.
On the other hand, the responsibility for production of metadata may shift culturally to the information provider, who could easily add information that would be helpful in cataloguing the resources.
One thing is certain - for the time being, it's important to keep the options open!
Further information about the ADAM Project can be found at http://adam.ac.uk
There is also an open mailbase list for the project at: firstname.lastname@example.org
The author can be contacted via e-mail on email@example.com, or by telephone on 01252 722441 x2427.