Since 1999, when it was first proposed to use metadata for collection-level description within the Research Support Libraries Programme (RSLP) , there has been steadily growing interest in this new method of supporting resource discovery. A number of collection-level description databases have now been created in the UK, funded through national initiatives. However, little documentation is available on how these were designed and created. This project explores setting up a database for a small geographic area, using best practice and with full documentation to support other local projects in this field.
The development of the Government's regional agenda with the creation of single regional agencies as strategic bodies for museums, libraries and archives has resulted in a growth of cross-sectoral collaborative initiatives. In May 2003, senior staff from public libraries and Further and Higher Education libraries in Bath met to discuss possible areas for regional cross-sectoral collaboration. This resulted in the formation of an informal network of libraries working in Bath and the North-East Somerset area called the Bath Area Libraries Initiative.
The aims are:
All libraries would not necessarily participate in every initiative, as this would depend on the relevance of the topic to the organisation's mission. The main areas of shared interest were identified as learning within the communities; cultural and community information; supporting private research and access to study space. The objectives of the Bath Area Libraries Initiative correspond closely with the recommendations of the LIC report "Empowering the learning community"  which investigated how to improve and stimulate co-operation between public and education libraries in support of life-long learning.
One possible initiative that emerged at the meeting was to establish a shared Web site with a community access focus to provide information on resources within libraries including a "Frequently Asked Questions" section and broad collection-level descriptions and strengths. It was thought that the Web site could, perhaps, become the kernel of an "access map" for libraries in the City of Bath and surrounding area of North East Somerset.
The "Empowering the learning community" report had also recommended that public and education libraries draw up access maps to enable learners to make use of the full range of resources available. The concept of an access map was further defined in a feasibility study by the Cultural Heritage Consortium in March 2003  as a tool which would enable a learner to answer the question "Where and how can I gain access to resources that will support my learning?" The Bath initiative was, at this initial stage, still confined to cross-sectoral library collaboration.
Given its expertise in library networking and collection description as well as geographic proximity, the University of Bath Library agreed to approach UKOLN for technical advice on setting up such a project. No funding had been identified so what the Librarians initially envisaged was a low-cost simple Web site with details about the libraries. It was hoped, however, that with more work and some funding, something along the lines of what the London Learning Group had achieved with their Find it in London Project might be possible .
UKOLN strongly advised against a simple text Web page and recommended the use of more structured data, such as a collection-level description schema, which would be easier to maintain, exchange and develop in the future. Furthermore, it was agreed that obtaining funding would depend on this interoperability and a standard format that could be used by others. What was proposed was a demonstrator project using best practice that could be used as a case study to support other local collection description projects. It was also agreed to widen the scope of the project to include collections in museums and archives as well as libraries and with that, Tap into Bath has progressed from the initial cross-sectoral initiative to a uniquely cross-domain project.
The City of Bath is a world heritage site with many rich sources of information relating to cultural, historical and local studies. It is also a respected centre for education with major HE and FE library collections. It was relatively easy, therefore, to compile a list of almost 30 potential participants. These include The American Museum in Britain, Bath Abbey Archives, Bath Central Public Library, Bath Record Office, City of Bath College, King Edward's School, The Roman Baths, The Royal United Hospital and Wessex Water. While information about some of these collections has already been collated in various publicly available databases, for example, Archon , Cornucopia , the SWMLAC Directory  and of course institutional Web sites, there is, at present, no single gateway for cross-domain searching. In addition, some collections are not listed anywhere and the searcher must know at least part of the collection name to have any chance of finding what information, if any, the collection itself has put on the Internet.
Although limited geographically to collections held in Bath with links to any associated collections, the intention is that Tap into Bath will be more inclusive in scope than many other collection description projects. Collections may include archives (which may be public or private or embedded within museum and library collections), all types of libraries (public, HE, FE, commercial NHS, government organisations) and museums (both public and privately funded).
Project management, technical expertise and data collectors are being provided by UKOLN and the University of Bath Library and a small amount of funding from existing budgets was set aside to pay for data collection work and development of the Web interfaces.
It was essential, however, to obtain the full co-operation of the collection managers (librarians, curators and archivists) to assist the project staff to describe the collections as accurately and consistently as possible. Each potential partner was contacted by telephone to introduce the project proposal and to check contact details.
The most efficient method of encouraging co-operation was to hold an introductory briefing session on Tap into Bath and the concept of collection description and its benefits. It was also an excellent opportunity to promote co-operation between colleagues working in the Museums, Libraries and Archives Sector in Bath. The South West Museums Libraries and Archives Council (SWMLAC) was approached for a small grant to cover this event and also a second event towards the end of the project to describe the project outcomes, demonstrate the database and discuss future possibilities.
The first SWMLAC Briefing session was held in April 2004 and was very well-attended. The response to the project was extremely positive and a number of benefits from participation are envisaged. Tap into Bath will enable collection providers to disclose information about their collections to users, some of which have not been catalogued at item level. Collection level description is a new concept for some of the smaller organisations and all the information professionals involved will benefit from the expertise that UKOLN has developed in this area. Many participants indicated that they had already been asked to provide collection descriptions to other projects and the Tap into Bath Project will support them to complete high-quality descriptions using recognised standards for their collections that can be used elsewhere.
As the intention was to use the database as a demonstrator project, it was important that recognised standards, such as the RSLP Collection Description Schema , and open source software were used when available. In the early stages of the project, the option of re-using databases created for other collection description projects was examined. However, this proved unfeasible; some databases had either extended, cut down and/or modified the RSLP Schema, while in others the database software could not be reused for a variety of reasons.
In the field of collection description the RSLP Collection Description Schema has become the 'de facto' standard. The schema uses metadata to identify all the elements of information, also known as attributes, that relate to a collection. It is not tied to a specific implementation but is usually used to create a relational database accessible online via the Internet.
Using the RSLP Schema means that descriptions created for Tap into Bath could, at a later stage, be contributed to other collection databases. This might be a single record contributed to a subject specific database, or the whole set of records contributed to a geographically based database that included Bath.
The schema does not prescribe whether elements are mandatory or may have multiple occurrences, nor is it tied to specific software; these decisions are left to the implementer. The first stage therefore was to review the schema and decide on mandatory elements, repeatable elements and which would contain free text and which controlled vocabulary, together with the sets of controlled vocabulary terms.
Since no suitable database had been identified for re-use, the project had to design the database from scratch. The open source software MySql was chosen as the relational database. This is hosted on a Unix server.
UKOLN and the library drew up the details of the database structure, identifying mandatory and repeatable elements and notes on content. A member of library staff then created the database in MySql using the schema and the project decisions on mandatory and repeatable fields.
Three different interfaces are required for data entry, for public search and display and for administrator access. The interfaces will use XHTML or HTML pages and will be designed with accessibility in mind. At this point the partners had no one with the required expertise and time to construct these interfaces, so a programmer is being employed on a short-term contract to do this.
Data collection was carried out in parallel with the design and creation of the database so that data would be available for entry at the point the database was complete.
Since administrators for the target collections had no prior knowledge of collection description, it was decided to produce a draft description using details that could be found on the Internet and a post-graduate student was recruited to collect this initial data. Information found in this way proved to be variable, and in the case of opening hours and admission charges, it was unclear how up to date these were. A member of the library staff searched for appropriate Library of Congress Subject Headings (LCSH). All data was recorded in table form in Word documents.
It was decided that in certain areas (for example the Strength element) specific wording was required for consistency and several lists of terms were created and the draft descriptions edited. The administrators of individual collections were then contacted and arrangements made to send each their draft description and to set up a telephone interview to review, supplement and correct those descriptions. The revised descriptions will undergo a final editing before data entry.
Phase 2 is still in progress, so no data has yet been entered. A prototype data entry interface has been created and early drafts of collection records have been used as test data. The final data entry interface is expected to be available in late June.
Using recognised terminology sets or schemas for subject indexing supports interoperability in the contribution of record(s) to other database or cross-searching with other databases using the same sets. Terminology sets may be general (LCSH), or subject-specific, for example Medical Subject Headings (MeSH) or the Art and Architecture Thesaurus (AAT).
Potential users of the Tap into Bath database are expected to include a range of people - the general public, schools, tourists, students, researchers - with a variety of searching skills. It was therefore decided to use a general set with wide usage in libraries (LCSH) for initial subject indexing. In consultation with partner collections, it may be that terms from other sets will be added for some collections at a later stage.
At mid-June 2004, the project is still in progress. The interfaces for data entry and searching are due for completion at the end of June. Data collection should be almost complete at that point and data entry can then take place. The project is hoping to launch the database in July.
Having encountered minimal documentation on the creation of other collection databases, the project is committed to providing full documentation. The programmer will provide a generalist and a technical version of his documentation. Tap into Bath will also feature as a Collection Description Focus Case Study  and further documentation on the project will go into the Collection Description Online Tutorial .
This project has already proved to be a worthwhile venture. The opportunity for information professionals from archives, libraries and museums from all sectors to work together and to learn about a new tool has been widely welcomed. It is also evident that more people are becoming aware of the potential of collection descriptions and the project has already had a number of enquiries regarding re-use of the database and its interfaces.