BUBL Information Service (BUBL)

Dennis Nicholson and Alan Dawson

Centre for Digital Library Research, University of Strathclyde, Glasgow G1 1XH



This book chapter appears in The Amazing Internet Challenge: How leading projects use library skills to organize the Web, published by the American Library Association, 1999


The BUBL Information Service (http://bubl.ac.uk/ or connect to link.bubl.ac.uk port 210 and database name ZPub) offers two major services for the U.K. higher education community: a directory of selected Internet resources covering all subjects of academic relevance (the BUBL LINK service), and the tables of contents, abstracts, or full text of hundreds of academic journals and magazines (the BUBL Journals service). Other services offered include an extensive directory of U.K. institutions (BUBL UK) and a specialist US service covering events, jobs, surveys, and mailing lists (BUBL News). This chapter gives details of the history, mission, funding, usage, collections, organization, policies, procedures, and future goals of the BUBL service, plus a summary of related projects that are investigating the use of LIS tools, standards, and expertise as aids to organizing the Internet. These projects include BUBL 5:15 (browseable access to key resources in over twelve hundred predefined subjects), CATRIONA I (Z39.50-based distributed catalogs of Internet resources), CATRIONA II (university management of locally created electronic resources), and CAIRNS (Z39.50-based integrated searching of dynamically generated clumps of catalogs enhanced through Conspectus-based collection descriptions).

1.0 An Introduction to BUBL

BUBL (http://bubl.ac.uk/ or connect to link.bubl.ac.uk port 210 and database name ZPub) is an Internet-based information service for the U.K. higher education community. When it was first established in 1990, BUBL was aimed only at library and information science professionals, and the name stood for Bulletin Board for Libraries. Since 1993, however, it has provided a service for the wider academic and research community, largely through the BUBL Subject Tree, and so the name changed to the BUBL Information Service, or BUBL for short. BUBL has long since outgrown its initial acronym, although a specialist service to the LLS community continues to be a significant function. BUBL is now run from the Andersonian Library of the University of Strathclyde, Glasgow, Scotland (http://www.lib.strath.ac.uk/).

1.1 Background History

BUBL began life as part of Project Jupiter, which was based at Glasgow University and aimed to train librarians in the use of JANET (http://www.ja.net/), the U.K. Joint Academic NETwork. When Project Jupiter funding ended in May 1991, a group of librarians from the Universities of Strathclyde and Glasgow, coordinated by Dennis Nicholson, saved BUBL from extinction by their voluntary efforts.

The service continued on this voluntary basis until early 1994, growing in popularity and attracting small amounts of sponsorship from commercial and professional organizations. In January 1994, one year's funding for the maintenance and development of BUBL was received from the Joint Information Systems Committee (JISC) (http://www.jisc.ac.uk/), which itself receives government funding via the U.K. higher education funding councils. This was followed by a further year's funding before JISC agreed that BUBL should be funded as a full U.K. national information service from August 1996, with funding pledged on a three-year cycle but allocated annually. At the time of this writing, BUBL funding was guaranteed until July 1999, with its future path beyond that date uncertain as JISC examines how best to develop services like BUBL LINK and related projects like EEVL (http://www.eevl.ac.uk/) (see chapter 5) and OMN1 (http://omni.ac.uk/) (see chapter 10).

Between July 1991 and August 1993, BUBL was run over a JANET link between the Universities of Strathclyde and Glasgow, before being moved to Bath University in September 1993. This move brought an upgrade from USERBUL software on a VMS system to Gopher and World Wide Web server software, which allowed the service to expand significantly. From September 1993 until March 1997, the BUBL service was run over a link between Strathclyde and Bath Universities.

BUBL was the first national U.K. service to offer its users subject-based access to the Internet through the BUBL subject-tree initiative, which began in 1993 and arranged resources together by subject area. The original Gopher-based subject tree was soon supplemented by a Web-based one. However, discussions over the limitations of the subject-tree approach led to the creation of the BUBL LINK service, accessible via Z39.50 as well as the Web and searchable in the style of a library OPAC.

The decision to move BUBL from Bath to Strathclyde and the diminishing use of Gopher precipitated the decision to relaunch the service. The new BUBL service, based entirely at Strathclyde University, began on March 23, 1997. The Gopher and Web subject trees were incorporated into the BUBL LINK service, with the remaining parts of BUBL revamped and the BUBL UK service introduced.

2.0 BUBL People

A great many people have contributed to BUBL over the years, mainly as volunteers in the early days, but more recently as paid staff. At the time of writing (July 1998), the service had two full-time staff, three part-time staff, and two vacant posts.

Dennis Nicholson has coordinated the BUBL service since May 1991 and since 1996 has had the title of BUBL director. He is head of the systems division at Strathclyde University Library, and he directs or codirects the CATRIONA and CAIRNS projects (described later). Alan Dawson is BUBL manager responsible for day-to-day running of the service. Alan joined BUBL in October 1996 and set up the new BUBL service at Strathclyde University. Andrew Williamson is information assistant with particular responsibility for the BUBL Journals and BUBL News services.

Many others too numerous to mention have contributed to BUBL over the years, particularly Fiona Wilson, Joanne Gold, and Jan Simpson, who each worked for over a year as BUBL information officer.

3.0 BUBL Mission

The BUBL mission statement, approved by the BUBL steering group in May 1996, is "to provide value-added access to Internet resources and services of academic, research, and professional significance to the U.K. Higher Education community by

To help meet these objectives, BUBL aims to offer fast, easy-to-use, and reliable access to selected high-quality resources of academic relevance, both on its own servers and worldwide.

4.0 BUBL Funding

BUBL is funded almost entirely by JISC, and currently costs £109,000 a year to run (or about $170,000). BUBL has not found it cost-effective to spend limited staff time and effort seeking commercial sponsorship, but it has received significant support from other sources, including the donation of NetPublisher (http://www.als.ameritech.com/) software from the Ameritech Corporation, educational discounts on hardware and software, free journal subscriptions (notably from Haworth Press), and institutional support from Strathclyde University such as office space, networking connections, and access to LIS journals.

5.0 BUBL Users and Usage

When BUBL was first established, its specific aim was to function as a bulletin board on JANET for the academic library community, but the approach was soon broadened to cover the U.K. higher education community more generally, though a specialist library and information science service is still provided.

Although aimed at U.K. users, BUBL provides an international service. During 1997-98, accesses to BUBL were recorded from 141 different countries (based on Internet host names). After the United Kingdom and the United States, the heaviest usage was from Australia, Canada, Germany, Spain, Denmark, Sweden, Ireland, Italy, France, and the Netherlands, in that order. This reflects the predominance of the English language in all areas of the BUBL service.

Total accesses to BUBL during the year 1997-98 were just over six million, more than double the figure for 1996-97. This compares with an estimated three thousand per year in 1991-92.

6.0 BUBL Collections


BUBL LINK (LIbraries of Networked Knowledge) is a catalog of Internet resources of academic relevance. It holds details of around eight thousand Internet resources and services and covers all main subject areas. Resource descriptions are searchable via Z39.50 as well as the Web, and the collection can be browsed by broad subject area, by keyword, or by Dewey Decimal Classification (DDC) (http://www.oclc.org/oclc/man/9353pg/9353toc.htm). Details of how BUBL LINK operates are given below.

6.2 BUBL Journals

The BUBL journals service offers tables of contents and abstracts (where available) of around two hundred and fifty current journals and other periodicals, with full text available for sixteen titles. The service acts as a current awareness tool for the library and information science (LIS) community in particular, since over 60 percent of titles relate to LIS. Other subject areas covered are agriculture, business, social work, and health. Holdings of many titles date back to around 1992 or 1993, and by mid-1998, there were over sixty-five hundred individual issues available. The content of each journal can be searched individually, and the complete collection can be searched as a whole or by broad subject area.


This is an extensive directory of U.K. organizations and institutions. it has a relatively simple structure but a large number of entries, currently organized under the following headings: Regional and Local Government, Political Parties, Banks and Building Societies, Charities, Companies, Newspapers, Television and Radio, U.K. Web Directories, National Information Services, Universities and Colleges, Schools, Churches, Hospitals, Libraries, Museums, Police, and Sports Authorities. It is popular as a simple and well-maintained reference service.

Some directories, for example, government and political parties, are maintained and updated by BUBL, whereas others are simply pointers to resources held elsewhere. Most users do not care where a set of links is held, and there is significant value in maintaining a central repository of resource links. Regular link checking ensures that resource links are kept up-to-date and this is supplemented by periodical searching and checking for new institutional pages. The directory of names of all U.K. Internet sites (http://www.hensa.ac.uk/uksites/) is helpful for this purpose.

6.4 BUBL News

BUBL began life as the Bulletin Board for Libraries, and this element of the service is retained in the BUBL News service. This service holds details of job vacancies (mostly United Kingdom or United States), forthcoming conferences and workshops, details of current offers such as surplus journal disposals, and other current news items relevant to the library community.

6.5 BUBL Archive

This provides storage for thousands of old files that are rarely needed but may be of some historical value. The four main sections of the archive are journals, Internet, LIS, and other subjects. The journals archive holds abstracts or full text from over one hundred titles that are no longer held in the main BUBL journals service.

6.6 AcqLink

AcqLink (http://link.bubl.ac.uk/acqlink/), run by Catherine Nicholson, is a welcome addition to the online resources available to all those involved in library acquisitions and collection development. It supplements the well-established AcqWeb (http://www.library.vanderbilt.edu/law/acqs/acqs.html), which although excellent is naturally aimed primarily at U.S. users. Now British and European acquisitions personnel have a comparable resource directly related to their needs.

7.0 Collection Policy

The philosophy underlying BUBL LINK is that for many purposes it is better to have a relatively small and well-maintained collection than a large arbitrary selection. The broad guidelines to use when considering resources to add to BUBL LINK are

Within these guidelines, certain types of resources are usually included, such as

In contrast, various resources are usually excluded, such as

Although such items are not normally included for their own sake, they may be included if they contain any of the desirable items, for example, if a company or local society maintains an extensive link collection or publishes a full-text online journal or magazine of wider interest. In such cases, the resource itself, rather than the company or society, is perceived to be of value and would be added to the catalog.

One of the biggest problems in applying such a collection policy to Internet resources is that of granularity, or deciding the level of detail to catalog. For example, one must decide whether to include a single catalog record for a learned society, or an individual record for each journal published by that society, or a separate record for every major article in every issue of every journal. In practice, the latter is not feasible with limited resources, but there are nonetheless numerous online articles or papers that may warrant individual cataloging.

7.1 Resource Discovery

BUBL is able to draw on several sources of information for locating Internet resources, such as

Incoming e-mail provides a regular source of suggestions, but these usually quote only a URL and title. All resources need to be checked and cataloged before they can be added to BUBL LINK. The above sources provide a useful but somewhat eclectic set of new resources. When time permits, we take a more coherent approach to a particular subject area by carrying out subject-specific searches and following up manually compiled subject-specific indexes. This is time consuming, but it is the type of activity that users would need to undertake for themselves in the absence of a service such as BUBL LINK. Time spent by BUBL staff can mean time saved by others. The more difficult it is to locate information, the more worthwhile the task.

7.2 Resource Evaluation

Finding Internet resources usually takes less time and effort than the evaluation and classification of them. At some point, BUBL staff need to make a judgment about whether an item is suitable for inclusion in the service. Even though a resource may appear to fall within the BUBL LINK collection policy, there are several reasons why it may fail to reach the quality threshold required for inclusion. In fact, BUBL keeps a list of rejected links, so in theory they may be checked at a later date, but in practice this rarely happens as there are always other new resources awaiting evaluation. Examples of reasons for rejection include

Although this does not mean that all such items are of no value to anyone, they do not warrant cataloging and inclusion in a selective resource collection such as BUBL LINK.

However, evaluation criteria are not constant. Some subjects are so well covered on the Internet that a new set of general links is of no great value. If the catalog already holds details of ten sets of general business and economics links, then there is little value in adding an eleventh. However, a set of links of similar quality for a less common subject, such as radio astronomy or physiotherapy, would have a higher relative value. On the other hand, full-text online books, tutorials, or reference data of academic relevance are almost always included, whatever the subject matter.

8.0 BUBL Operations

8.1 Hardware and Software

Most BUBL pages are stored on a Viglen XX server running Windows NT 4 (http://www.microsoft.com/ntserver/) and Netscape Enterprise Web server (http://merchant.netscape.com/netstore/servers/enterprise.html). Search facilities are provided by the Verity search engine, which is distributed with the Enterprise server. The service aims to be available to users twenty-four hours per day throughout the year, although some service breaks inevitably occur. In practice, over the course of the year, the BUBL service is available over 99 percent of the time.

BUBL LINK runs on a separate Viglen Genie 2 Plus server, running Windows NT 4 and Ameritech NetPublisher software, which acts as both an HTTP and Z39.50 server. A third Viglen Genie server acts as backup for the two main servers: all BUBL content is replicated there, and it may be brought into operation on short notice in case of failure of one of the other servers.

The rest of this section focuses on operation of the BUBL LINK service.

8.2 Database Software

For any large collection of links, there are several advantages in using database software to generate menus, rather than manually editing and maintaining numerous HTML pages. Two obvious benefits are the use of templates and the provision of search facilities.

Templates allow a consistent appearance to be given to a large number of Web pages, which can be dynamically generated by the database software. This obviates the need to design and create numerous individual pages, it allows the inclusion of a standard header or footer on every page, and it allows any changes to be made only once while appearing throughout the service.

The software used for BUBL LINK, Ameritech NetPublisher, has flexible template options that allow different subsets of information to have a distinct design. This preserves consistency, yet allows for variety where required (e.g., in the AcqLink service). Most other Web database publishing software offers similar features.

The structure offered by database software allows field-specific searching, as in a traditional library catalog. This is only effective if the content is similarly complete, for example, if the author, abstract, and subject keywords are recorded in the database along with the title. This structure enables users to find, for example, documents written by Einstein, as opposed to those written by others about him, or those that happen to mention him, or pages written by people living on Einstein Avenue. Large-scale Internet search engines can be very effective for many purposes, but they are rarely able to make the distinction between different occurrences and meanings of a specific word.

8.3 Database Design and Operation

The BUBL LINK database aims to offer users far more than a mere collection of links, yet remain relatively simple in structure so that rapid data entry and updating is feasible. The database holds ten fields for every record in the catalog.

Item name:
The title of the resource. This is not always obvious, as there may be one heading in the title bar and a different title on the Web page itself. Sometimes acronyms need expanding or further context needs adding to make the title more meaningful.
Most items are Web pages, but Gopher, FTP, and Telnet resources may be included if still relevant and up-to-date.
Item abstract:
This is written or edited by the BUBL cataloger and is intended to be concise, accurate, and meaningful. If extracted directly from the resource, it usually requires editing to remove value judgments and excess verbiage.
May be an individual or an institution or both. Unattributed resources are occasionally good enough to be included, but most high-quality resources have a clear author or owner.
Author type:
Classifies the author into a predefined category, for example, person, society, company, university, government, and so forth.
One or more of around twelve hundred subject terms, described below. The intelligent completion of this field enables effective operation of the BUBL 5:15 service (http://bubl.ac.uk/link/five/). Selected major resources may be allocated numerous subject terms.
Dewey class:
The Dewey Decimal Classification number (edition 21). Items are filed in menus based on this number, but cataloging need not be as detailed as that required for physical book shelving.
Resource type:
One or more of around twenty-five predefined categories, such as bibliography, biography, book, dictionary, document, forum, guide, image collection, index, journal, mailing list, news, reference data, or software. This enables retrieval of specific types of resources, as opposed to those on a specific subject.
The physical location of the server that holds the resource. Around 50 percent of resources in BUBL LINK are located in the United Kingdom or elsewhere in Europe, though some of these are mirror sites.
The date an item was added to BUBL LINK or was last checked or updated. The main value of this field is to enable BUBL staff to review all items that are twelve months old. This field is also used to generate a list of recent additions that is published on BUBL twice each month (http://bubl.ac.uk/news/updates/) and also distributed to the lis-link and lis-subjects mailing lists (http://bubl.ac.uk/mail/) so that BUBL is able to provide an Internet resource current awareness service.

8.4 Update Procedure

To maximize flexibility and resource quality, additions to the BUBL LINK database are made in batches on a daily or weekly basis. Most resource descriptions are compiled using Microsoft Word as a powerful plain text editor, though Windows Notepad or Wordpad could be used. This means that anyone can compile a cataloging record, using a simple template format, and send it to BUBL via e-mail. For example:

<name>Beginner's Guide to Organic Synthesis

<abstract>A guide to organic synthesis, including chapters on Grignard reactions, enolates, the Wittig reaction, linear synthesis, and convergent synthesis.

<author>Otto Meth-Cohn, Sunderland University


<subjects>organic chemistry

<deweyclass>54 7




Once a set of resource descriptions has been compiled, the spelling and subject terms are checked carefully, and they are translated in bulk to an import format understood by the NetPublisher software. This translation is carried out by a fairly simple Word Basic program specially written for the task. Addition to the database is then simply a question of filing items in the appropriate menus according to the DDC number. Once the database has been saved and reloaded, new items are automatically available to users. Web-page design and the display of search results is handled automatically by the database, so no HTML editing is required beyond initial template design.

8.5 Link Checking and Record Maintenance

Many regular Internet users find the most frustrating aspect of information gathering is the broken link problem. All large directories and search engines suffer from this. In the time lag between items being added to their database and users searching it, the Internet changes. Files are deleted, file names change, sites are reorganized, people change jobs, software is updated, and new policies are introduced. The net result is a proliferation of broken links.

Tackling the broken link problem is not trivial. Software can help, but needs supplementing with manual labor. The policy with the BUBL LINK catalog is to run link-checking software at least once per month and to manually check each link at least once per year. This all takes time and effort, but the big benefit for users is the guarantee of less than 1 percent of links broken at any time (usually less than 0.5%). Even checking links with software takes time. The excellent Linkbot program (http://www.tetranet.com/products/linkbot-main.htm) is able to check eight thousand links in less than an hour, but the follow-up takes much longer. There are numerous possible reasons for a link to be broken at any given time, so resources cannot be deleted automatically; you can't knock down a house just because no-one answers the doorbell. Further investigation is required, since sometimes a redirection is provided and the URL can easily be updated; sometimes more detective work is required; sometimes the title or author has changed; or sometimes deletion is warranted.

Internet information is volatile by nature, and any resource directory needs a strategy for dealing with this. In practice link checking is so time-consuming that at BUBL the effort is concentrated on following up "file not found" errors (error code 404) and permanent redirections. Link failures due to time-out or host-name errors are usually temporary and only require action if failure persists for two or three months.

9.0 Taxonomy

Database software provides searching facilities and ensures that a resource directory is far more than just a collection of links. But users want to be able to browse too, which means the directory requires some clear, browseable structure. The database software may or may not make such structure easy to provide, but it still requires some manual filing, just as in a physical library. BUBL LINK uses the well-established DDC number to provide a browseable structure, with a series of hierarchical menus based on DDC class numbers. As most users are not particularly familiar with DDC, this is supplemented by a simple subject keyword index so that it is easy to find the relevant menu without knowing its location in the hierarchy.

9.1 Subject Classification

Apart from broken links, perhaps the biggest problem faced by Internet researchers is that of terminology. Language is ambiguous, imprecise, context dependent, culturally dependent, and full of synonyms and subtlety. This richness has made it very difficult to develop software that understands natural language, and it also poses problems for Internet search software that relies mainly on simple word matching. Anyone searching for information on the Latin language will find thousands of resources about Latin America. This problem can be partially solved by specifying more complex searches, such as "Latin NOT America' but most users fail to do this, and in any case there are more fundamental problems. Many topics and concepts are not easily captured by a single word or phrase, especially in the social sciences or library and information science. Terminology is a big problem for any automated search system.

The approach that has evolved for dealing with this problem at BUBL has been to develop and apply a large but controlled set of subject keywords to supplement the resource abstracts. Originally we used an existing standard, Library of Congress Subject Headings (LCSH), but this had to be modified significantly to make it appropriate for Internet use. In some areas it proved far too detailed, in others not detailed enough. Currently, BUBL uses a set of around twelve hundred subject terms to describe Internet resources. As these terms are applied by BUBL staff after consideration of the content, users are able to retrieve only items that are about the specified topic, as opposed to items that happen to mention it. This important distinction is very difficult to achieve with a software-based system, though of course software is able to handle a larger number of resources. The benefits of this approach are visible to users via the BUBL 5:15 service.

9.2 BUBL 5:15

BUBL 5:15 offers a new and unusual approach to Internet resource discovery, yet it is basically a new interface to the BUBL LINK service. The reason for the name is that users are guaranteed at least five relevant matches for any subject, and in most cases will get no more than fifteen (the upper limit is not rigidly applied, and a few subjects may produce up to thirty-five hits). BUBL 5:15 is not intended as a substitute for Internet searching, but as a fast, simple, and reliable alternative method for finding information on around twelve hundred different predefined subjects via a point-and-click interface.

The twelve hundred subject terms are divided into the nine top-level headings identified by JISC (BUBL's funding body): Creative Arts; Engineering and Technology; Health Studies; Humanities; Language, Literature and Culture; Life Sciences; Mathematics and Computing; Physical Sciences; and Social Sciences. BUBL has added two further main headings: library and information science (BUBL's main subject specialization) and area studies (resources about all 193 independent countries), as well as a full A-Z list of all subjects. Some subject terms are allocated to more than one category.

The use of a tightly controlled vocabulary for BUBL 5:15 is highly significant as it removes the difficulties and uncertainties of searching. Users do not need to know or guess the precise term to type as all subject terms are visible on screen via a set of twelve pull-down menus. In this respect, the philosophy of BUBL 5:15 is similar to the techniques of faceted classification and view-based searching, which are currently being studied for application to large resources such as a Library catalog and the EMBASE biomedical reference database (http://www.hud.ac.uk/schools/cedar/hibrowse.html).

BUBL 5:15 is particularly useful for library staff, information professionals, and Internet trainers who need to cover a wide range of subjects, as well as to students and infrequent Internet users. Librarians have found BUBL 5:15 valuable when dealing with face-to-face user inquiries as it guarantees that they will find relevant resources for a large number of subjects, removing the uncertainty of general Internet searching and the frustration of finding broken links.

The controlled approach offered by BUBL 5:15 also allows users to be directed toward important resources relevant to a specific topic, a traditional activity for librarians. It also makes resource discovery easier for users themselves by providing a more helpful interface than a search box.

BUBL 5:15 has been described as a browse engine (http://www.ilrt.bris.ac.uk/roads/news/issue6/bubl/), but the underlying implementation is by database searching. Whenever a user selects a subject term from one of the twelve menus, this term is sent to the BUBL LINK database as a request to search the subjects field only for that term. The effectiveness of the service is therefore dependent on the subject classification carried out by BUBL staff. Display of results is handled by a specially designed database template, which uses HTML tables to display both the list of titles and resource descriptions on the same page.

10.0 Strengths and Weaknesses

The main strengths of BUBL are the quality and breadth of its services, as described above, and its links with the library and information science community. For example, BUBL runs lis-link, the U.K's major library and information science mailing list, which carries discussion and information on all aspects of library operation. It is also used by BUBL to send out news relating to the service, notably the weekly BUBL Updates and the fortnightly BUBL LINK Updates. Details of lis-link and other mailing lists managed by BUBL are available via the BUBL Mail service (http://bubl.ac.uk/mail/).

Other strengths of BUBL include its

BUBL has been highly praised for its extensive and timely publication of meaningful usage statistics, which include details of the most popular services, files, and journals, as well as a monthly summary of accesses. These figures are published online each month in the BUBL Admin pages (http://bubl.ac.uk/admin/usage/).

There were over six million accesses to the BUBL Web service in 1997-98 (6,070,917), an increase of over 120 percent from 1996-97. BUBL regularly features in the "Top 50 UK Web Sites" (http://www.top50.co.uk/). Over twelve thousand separate pages on the BUBL service were accessed during August 1997-July 1998, with BUBL Journals and BUBL LINK being by far the most popular areas. Around 60 percent of accesses to BUBL can be attributed to a specific country, based on the ISO country code in the domain name of the system making the connection. Accesses were recorded from 149 different countries in 1997-98. The top six were the United Kingdom, the United States, Australia, Canada, Germany, and France.

Another method of determining our success is from our users.

My experience since you launched BUBL 5:15 has continued to be positive, and I regularly introduce it to students and colleagues as a quick and easy way to find a selection of quality resources.
(Alison McNab, Loughborough University)
I'm impressed with the depth of such fine areas of your information service, such as drama, palaeontology, and oceanography. These confirm not only the comprehensiveness that is BUBL but also its determination to service all areas of knowledge with equitable attention. The effort and energy invested in BUBL is readily apparent and appreciable. I'm sparse with accolades but will readily recognize excellence wherever it arises. Keep up the great work.
(Robert J. Tiess, The New Athenaeum)

The main weakness of BUBL is perhaps its vulnerability to staff turnover, an inevitable consequence of short-term funding. This also makes innovation difficult in certain areas, for example, in providing additional value-added services such as mirrors or subject-specific search services. BUBL sometimes suffers from being a small organization in a fast-moving, ever-changing environment.

11.0 Future Goals of the BUBL Service

BUBL's funding body, JISC, is currently considering how best to organize and fund subject-based information gateways over the next few years, so the future of BUBL and of projects such as EEVL and OMNI (see chapters 5 and 10) will develop in this context. Discussions to date have been based on the idea of a federation of information gateways integrated at some level to provide a single service to users. The question of whether this will offer a common user interface, subject-specific interfaces, or both, is one of the threads of the discussions. BUBL has been invited to contribute to these discussions and has suggested a number of possible roles in any such federation of information gateways, including

In the meantime, with the shape of future developments in this area still unclear, BUBL's strategy is to build on existing strengths and to focus effort on areas of the service most likely to be of future value. In summary, BUBL is aiming to

This strategy is designed to keep options open during a period of uncertainty while continuing to strengthen and develop the various services valued by users. At the same time, the service aims to develop knowledge and experience in the application of LIS skills, tools, and standards to organizing access for Internet resources, and to contribute to discussion and discovery in these areas by being involved in research and development initiatives of the kinds described below.

12.0 Associated Projects and Future Developments

Because of its history, BUBL's position on Internet resource location and description has, over the years, tended to be heavily influenced by the library and information science perspective, with its strong tradition of concern for user needs, and this continues to be true of its current approach to such issues.

Thus, while monitoring the potential importance of developments in the community at large, such as the use of metadata and the development of Resource Description Framework (RDF) (http://www.w3.org/RDF/), the efforts of BUBL, and its vision of the future, are focused on the concerns of librarians and information specialists and the users they serve, and on the use of LIS skills, tools, and standards to solve resource location and description problems. This general approach is illustrated in a number of recent or current initiatives in which BUBL staff have been involved, initiatives that have either exemplified or helped form a vision of the future in this area.


The original CATRIONA project (now known as CATRIONA 1) investigated and demonstrated the feasibility of a distributed catalog of Internet resources based on library standards Z39.50 and MARC, and integrated with library OPACs. CATaloguing and Retrieval of Information Over Networks Applications (CATRIONA) was based at Strathclyde and Napier University Libraries in Scotland and ran from mid-1994 to early 1995. It envisaged that, in the long run, resource discovery on the Internet would be based on local cataloging and control of locally created resources (considered essential if the problems of broken URLs were to be reliably solved), and would be based on distributed searching of groups of Z39.50-compliant catalogs that would include, but not be limited to, library OPACs - groups of catalogs that would be created by users searching catalogs of such catalogs. Although it had no immediate follow-up, CATRIONA influenced a range of other future developments and initiatives, both within BUBL and within the Scottish library community:


CATRIONA II (http://catriona2.lib.strath.ac.uk/catriona/) is a project investigating university approaches to the management of locally created quality teaching and research resources, examining a range of issues, including: associated policy, strategy, and organizational infrastructure; the role of the library; and service design, resource description, and inter-university integration.

Among other things, the project is examining whether universities will choose to manage services to deliver locally created electronic resources beyond the local campus, and whether the library will have a role in such services. The indications are that, in many instances, the answer to both questions will be yes, outcomes that, if put into practice, may well influence some aspects of resource location and description, at least in the United Kingdom, and may also have a bearing on the future activities of services such as BUBL.

CATRIONA II surveys at six Scottish Universities have shown that:

Clearly, a lot of valuable and in-demand material exists on U.K. campuses, and it is important that this material be made more accessible, whether or not universities themselves choose to play an active role in this. At present, a number of outcomes are possible, and some or all of them may have a bearing on the future of BUBL:

1. Universities may choose to manage services themselves, in which case services like BUBL may:

  1. disappear entirely as cataloging of Internet resources in the United Kingdom becomes the province of universities, and resources in other countries are handled by services in those countries;
  2. continue in a new role, offering value-added services such as BUBL 5:15 or a Z39.50 distributed search service;
  3. become a mechanism through which the universities might sell access to their resources;
  4. become integrated within the services offered by universities and run by groups of universities, with funding direct from service management rather than through top-sliced funding.

2. Universities may choose not to manage services themselves, in which case services like BUBL may become the mechanism whereby resources created locally at U.K. universities are cataloged and made accessible.

In either event, it is envisioned that responsibility for resource location and description will become more local. Either U.K. services like BUBL will be increasingly responsible for providing and maintaining access to U.K. resources and offering integrated access to resources worldwide through gateways offering distributed searching of foreign services, or the responsibility may devolve even further to the universities themselves, an outcome that the CATRIONA I project saw as the only practical way of controlling the problem of broken links (if both the resource URL and the resource description are controlled by the local library, then in theory, broken links should not occur).

The mechanisms whereby such resources are made available to the wider world are obviously important, and Z39.50 is viewed as a key standard in CATRIONA II. The project is examining two methods of managing resource location and description and one method of inter-university integration:

  1. the centralized model, being investigated at Napier University, where all locally created resources are cataloged by the library in the library OPAC, a Z39.50-compliant server;
  2. the devolved model, being investigated at Strathclyde University, where resources are cataloged on departmental servers using embedded metadata (probably the Dublin Core [http://purl.oclc. org/metadata/dublin_core/] or IMS metadata [http://www.imsproject.org/metadata/], which is a superset of DC). The project is investigating the use of Harvest software (http://www.tardis.ed.ac.uk/harvest/) to collect embedded metadata from distributed sites on campus, with a view to incorporating it within a database mounted on a central Z39.50-compliant server (the CATRIONA II demonstrator server).

In both models, inter-university integration, the mechanism whereby locally cataloged resources are made accessible beyond the local campus, is centered on the Z39.50-based approach being investigated within the CAIRNS project.


CAIRNS (http://cairns.lib.gla.ac.uk/) stands for Co-operative Academic Information Retrieval Network for Scotland. The project aims to integrate the twenty-five Z39.50-compliant catalogs or information services of CAIRNS sites across Scotland into a functional and user-adaptive test-bed service. This will offer efficient and effective search and retrieval capabilities across a clump of services comprising all of the individual CAIRNS bibliographic databases and also across various smaller sub-clumps (groups of servers), and will provide

The project will take advantage of SCURL's Web/Z39.50 conspectus-based Research Collections Online (RCO) service as the basis of a subject-based, dynamic clumping service. (BUBL received funding from the National Library of Scotland to purchase, set up, and manage a server to hold the RCO data on behalf of the Scottish Confederation of University and Research Libraries.)

This will provide users with dynamically generated, subject-based sub-clumps of SCURL catalogs to search via Z39.50. An extension of this would enable users to be presented with subject-based sub-clumps dynamically generated via searches of catalogs of electronic resources: BUBL LINK, SLAINTE, CATRIONA II demonstrator and Strathclyde University Library's Z39.50-compliant Web server.

13.0 Conclusions

The mechanisms by which Internet resources will be published, organized, and accessed in the future are unclear. One of the ultimate goals is to combine the quality and precision of resource description offered by services such as BUBL with the huge scale of information handled by current Internet search tools. The extensive use of metadata by resource providers offers the promise that this goal may be reachable by enabling software to carry out the filtering, evaluation, and organization of resources that is currently a time-consuming human activity. Increased use of Z39.50 offers another means of distributed searching and resource discovery.

These approaches will only be successful on a large scale if the necessary software is in widespread use, whether incorporated into desktop browsers or used at an intermediate stage to offer enhanced user services. Information specialists will still be required to direct the operation of this software, to customize it to specific groups of users, and to design interfaces that assist users to drive the software effectively. The nature of Internet resource directories and gateways will continue to evolve, but they will still be need to be guided by the hands of librarians and information professionals.