RDN is just one of the many highly successful projects funded by the Joint Information Systems Committee (JISC) , which has also been negotiating UK-wide license agreements to databases for all students, faculty and staff at further and/or higher education institutions in the United Kingdom for a decade. The most recent nationwide license for further education institutes is for a subset of the Gale Virtual Reference Library from Thomson Gale. It well complements the quintessential British ready-reference sources offered by Oxford University Press and Xrefer Plus, which are also licensed by JISC for nationwide use. It seems that funding is usually mid-range for most projects, which provides predictability, although not perpetuity — even though JISC uses the adjective perpetual for its license with Thomson Gale (which extends to mid-2007). Then again, perpetual may mean that even if the license is not extended in a new contract, UK further education institutes will still have access to the collection of 21 titles without the updates.
In the United States, the Digital Library Initiative (DLI) was launched in the late 1990s, but with rather modest funding. DLI2 is much more generous and spawned many excellent research projects, but DLI2 does not have the same direct and wide-ranging impact on the college libraries as JISC.
Indeed, the best directories in the U.S. have been created as labors of love by a single, competent individual — like Mike Madin's Academic Info — or a group of librarians — such as the Librarians' Index to the Internet (LII). LII was founded by Carole Leita with the help of a dozen librarians in California; later, some funding was provided by various organizations. These are useful and very well-annotated, but relatively small guides and LII is not meant to focus on academic resources.
INFOMINE was launched by librarians from the University of California at Riverside in the mid-1990s and later received external funding. There are about 25,000 - 26,000 entries in INFOMINE about scholarly Web resources created by information professionals. Smartly, the directory search is limited to this segment of INFOMINE by default, excluding the much larger collection of entries hoarded by robots.
The superb Internet Scout Project with its Scout Reports also started out as a labor of love by its late founder Susan Calcari, and with partial support by the University of Wisconsin. Lack of funding made it necessary to suspend some parts of the project earlier. The good news is that it now seems to be funded as it deserves by a group of generous foundations. Its survival may have to do with the fact that the Scout Report was picked as a partner in the recently started, but promising, National Science Digital Library (NSDL) project that has been developing Web-based educational services and collections similar to JISC.
RDN is set up in eight hubs, each of which represents a group of specific disciplines that develop and maintain annotated directories of Web resources in their fields, as well as other educational materials, such as tutorials in digital format for Web distribution.
The hubs were developed at different times, starting with the Social Science Information Gateway (SOSIG) and Engineering, Mathematics and Computing (EEVL). They are obviously not created using a cookie cutter. Still, some of them have more resemblance to each other than others. For example, the Geography & Environment Gateway (GESource) and the Physical Sciences Information Gateway (PSIgate) feature similar layout and structure, as do SOSIG and Humbul, the gateway to humanities Web sites. The guide to Arts & Creative Industries (Artifact) has a distinct layout, as does Altis (which is not an acronym, but a reference to the sanctuary of Zeus), the guide for Hospitality, Leisure, Sport and Tourism Web sites.
Each hub has a leading organization that is in charge of the maintenance of the hub, often along with a number of partner organizations. The Consortium of Academic Libraries in Manchester (CALIM) is responsible for PSIgate, GESource and Artifact; the University of Nottingham for BIOME; the Heriot Watt University for EEVL (which originally stood for Edinburgh Engineering Virtual Library); the University of Bristol for SOSIG; the University of Birmingham for Altis; and the University of Oxford for Humbul. The number of sites within the hubs varies greatly, which reflects the popularity of the disciplines and their overall prevalence on the Web.
Altis has the fewest number of entries, about 3,500, and not surprisingly, BIOME, the gateway to Web sites related to Health, Medicine and Life Sciences, has the most: more than 25,000. The number of entries may be misleading, as in some cases separate entries are created for a single Web site, similar to creating a catalog record for every chapter of a book. For example, there are 60 entries for the British United Provident Association (BUPA), each of them a PDF discussing a common disease, condition or treatment. Each has a summary, not unlike the large sites or complete publications, although these are short entries.
The majority of resources have substantial and informative descriptions, as opposed to the too short and often erroneous entries in the widely adapted Open Directory Project (ODP), or in the ever-deteriorating Yahoo directory. Here is a characteristic example for the British National Corpus site from RDN, ODP and Yahoo. Even when Yahoo has an entry for a site, it is minimal compared to what RDN offers. For fairness, for some of the test searches both ODP and Yahoo came up with more sites than RDN. Then again, quantity and quality should be distinguished.
On the negative side, in spite of the richness of RDN, there are inexplicable omissions. You will not find information about excellent open-access British dictionaries and encyclopedias, such as the Collins English Dictionary, Chambers 21st Century dictionary, the family of Cambridge Dictionaries, and Chambers Biographical Dictionary. There is no entry about the free Concise Hutchinson Encyclopedia. Nor will you find the free Canadian Encyclopedia Online or the Encyclopedia of Music in Canada. After these omissions, it is no surprise that there are no entries for many of best free American ready-reference sources, such as the Merriam-Webster Collegiate Dictionary 10th edition, the Random House Unabridged Dictionary, Encarta Concise Encyclopedia or Britannica Concise Encyclopedia, to name a few.
The presence of a record for the top-notch Bartleby.com site gives hope that users will discover many of the gems therein that did not get full-entry treatment, such as the American Heritage Dictionary, but at least appear in the resource description. However, only through serendipity will users bump into resources for Bartleby.com that are not mentioned in the RDN records at all, such as the Columbia World of Quotations.
In light of these omissions, it is more disappointing that RDN has multiple descriptions for the same resource in the catalog of different hubs. The Oxford Reference Online (ORO) site is undoubtedly worthy of inclusion, but not in the catalog of three of the RDN hubs. No matter how impressive the Web site of the National Museum of Science and Technology in Milan is, it is not the best use of catalogers' time to create an entry in four of the RDN hubs. There are many other duplicates as well, such as the ones for the Oxford English Dictionary. The entry in Humbul is far better (it was created by information professionals at University of Oxford, the publisher of OED) than the harried one created in Artifact.
Such duplicates and triplicates may come with the turf when so many organizations are involved in a large-scale project. One may also argue that entries should be present in two or more of the hubs, as a user may be looking for the same resources through different hubs. While that may sound reasonable, it would be much better to create a link to the best version of the record and display that record in each hub when the users are browsing in one hub. Applying the concept of "one record, multiple displays" from online catalogs that display the same full record for the primary, secondary and tertiary author(s) would be better solution. In addition, RDN has a simple search engine, ResourceFinder that searches across the collections of all of the hubs. It is more capable than its Help file may lead you to believe, as I will discuss in the software section.
Duplicates within the same hub, such as the pair for the Oxford Companion to English Literature, are less excusable. It makes the case even stranger when you find out that although one record was created on July 11, 2001, and the other a day later, they were both updated on the same day. In the update, the duplicates must have been, or at least should have been, discovered and one of the entries should have been deleted.
In testing, I found some outdated, and hence inaccurate, records, such as the one about the Xrefer.com site. Xrefer was free when the record was created and included many titles from the Oxford reference stable, but later these were withdrawn. A year thereafter the free Xrefer.com — as contemplated at launch — was replaced by a fee-based version that kept getting enhanced by many more excellent reference works. Neither the Humbul hub nor the SOSIG hub updated the information or removed the outdated record. At least Humbul added a new entry for Xrefer Plus (which also has multiple records). The one created for Humbul is much more detailed than the two entries for the fee-based version described in GESource and Artifact, not only in terms of the description but also from the perspective of the other metadata.
Errors and typos were rare in the 120-130 sample entries that I looked at. The misspelled format of the name of Dante Alighieri as Aleghieri, both in the title and in the description of the catalog record, seems to be an exception.
I am not insensitive to the difficulties posed by cataloging Web sites. Cataloging print books is considered by many to be child's play compared to cataloging serials, as the metadata remain the same once the book is published, while with serials, the title, subtitle, editor, publisher, place of publication and frequency often keep changing. In turn, cataloging serials is a walk in the park compared to cataloging Web resources. As for the subject terms in the metadata: they often show redundancy and inconsistency, but traditional indexing/abstracting services have also struggled with these problems for a long time.
It would be very useful to have a specific metadata element to describe (and make searchable) the access type of the resource, such as free for anyone, free for FE/HE institutions in the UK, or individual/institutional subscription required.
In spite of some structural differences, the gateways are so intuitive that jumping from one to another is not a problem. Navigating through the subject trees is straightforward in all of the hubs. This is particularly true for those hubs that display how many entries there are for each section, subsection and subject heading.
SOSIG offers three thesauri, including the most comprehensive General Social Science Thesaurus, with posting information and optional term explosion (for extending the search to the records that have the narrower terms of selected descriptors, like suffrage for the broader descriptor of political representations). BIOME features a subset of the 2005 edition of MeSH terms, limited to the subject headings assigned to one or more sites. This is a reasonable compromise to keep the MeSH file quickly browsable in an unusual, but practical, three-column layout — although without posting information that would be a useful hint in fishing for the best MeSH terms in BIOME.
The search engines of the hubs are not identical, although many use a similar layout and offer similar features. Most have an advanced search mode. BIOME, SOSIG and GESource offer sorting/ranking and other output options for result display.
The separate search engine Resource Finder searches across all of the hubs. It is a no-frills piece of software, but it is more capable than the help file leads you to believe. The help file claims that phrase searching is not possible. I tried searching for "English dictionary" without quotes around the query words and received 134 results. Performing the same search with quotes produced 16 hits. This leads me to believe that the software can indeed distinguish between phrase searching and Boolean searching (the default operation for a space in the query) — an important feature given the substantial and long descriptions in RDN.
In spite of the deficiencies, RDN belongs to the few good Web directories and subject guides dedicated to scholarly resources that are not polluted by uninformative entries about mediocre or inferior sites.