Present: Tom Bolze, Daniel Lovins (chair), Youn Noh, Britta Santamauro (recording). Meeting began at 2.30 p.m. in Room 411.

Note: Action items in boldface green

Postponed XML/XSLT demos

Daniel updated task force members on his and Britta’s experiments with XML and XSLT which will be part of the Russian Avant Garde metadata pilot project. Daniel emphasized that some basic XML knowledge will be extremely valuable for catalog librarians. The person able to manipulate XML/XSLT serves as a kind of gatekeeper, since all metadata processing has to go through and is manipulated by the XML/XSLT configurations. Such knowledge will also possibly speed up any production process, since it will empower members of our department to repurpose their own metadata, rather than remaining dependent on other departments such as Systems and ITS. Daniel offered to write a summary/narrative of the tools and resources on XML/XSLT that he has started to put together. Britta mentioned and recommended the free XML tutorial on OCLC’s WebJunction (http://webjunction.org/do/Home): after having registered on the website, there are many (mostly technically-oriented) online tutorials, click on the tab “Learning Center” and then on online courses in the gray box in the upper left hand corner of the page. Daniel suggested that we should look at the XML tutorial in more detail at some other time.

GLBTQ book jackets project

Unfortunately, Tom’s meeting with Acquisition staff—i.e., the one to evaluate GLBTQ book jackets for a possible digital library experiment—had been canceled, so there was nothing new for him to report. Tom asked where he would find advice on the copyright issues involved. Daniel pointed out that the material should fall under the fair use doctrine. Britta mentioned that the copyright issues for digital material might be more complicated (e.g. restricted access, password protected etc.) and offered that she will bring a summary of the copyright laws that she still has from library school. Tom also asked about the process of writing the grant application. He voiced his concerns that the professor might not have the time to write the application, and that he is concerned about how to approach an offer to write the grant application on the professor’s behalf. All agreed that this is indeed a delicate situation but that it much depends on the professor himself and his attitude and enthusiasm towards the project. Daniel also mentioned the possibility of applying for a SCOPA grant in case the project fails to become part of the “Davis Grant”.

Tom pointed out that Agre's argument for greater subject specialization among librarians, particularly catalogers, seems to raise the danger of making cross-disciplinary (or inter-disciplinary) research more difficult, since more specialized librarians will probably adopt the unique vocabularies of the various disciplines. Daniel noted that Agre’s position seemed opposed to that of Marty Kurth.  The latter believes catalogers will flourish in their traditional role as conceptual “translators” between and among different individuals and communities. Through consistent use of controlled vocabularies, authority files, cross-references and assignment of subject headings, according to Kurth, catalog librarians ensure that different disciplines can “talk” to one another. Tom further commented that the subject headings for digitized bacteria slides, as demonstrated by the medical librarians earlier that morning, were hyperlinked, and that we still don’t know if anybody is actually using them, or if there have been any pertinent studies done on them.

While not anticipated in the agenda, Daniel mentioned that NCSU announced today that it is the first library to have implemented Endeca’s faceted navigation technology. Major retailers (including Barnes & Noble, Walmart, Amazon, etc.) as well as IBM and NASA are using this technology already. Endeca claims that its technology facilitates context-sensitive browsing and eliminates the need for repetitive and inefficient keyword searching.

After evaluating the results of a trial search on the word “cosmopolitanism”, Youn noted that she found the browsing difficult since it offers an overly broad point of entry, whereas the more specialized perspective of a discipline would have been helpful. Tom agreed that traditional subject heading linking somehow seemed easier. Daniel noted that the technique is celebrated from a database implementation and management perspective, but not necessarily (at least yet) by catalogers and end users. The classification categories are generated through matching the search term to the frequency of the word in the record. Tom concluded that it does not resemble browsing the shelves but rather pulling a selection from a given set of items. The question remained open, how they actually pull those records. It was also questioned whether this technique might work best with full text but not so well with bibliographic records.

Faceted Headings (FAST Project Demo)

Daniel then demonstrated the OCLC research project called “FAST” (Faceted Application of Subject Terminology) by searching a special FAST database that had been set up by OCLC, and was designed to be tested  by Arlene Taylor’s students at the University of Pittsburg. FAST algorithmically converts established and applied LSCH headings into faceted FAST equivalents. The idea is that LCSH is a good source of controlled vocabulary, but that the traditional subdivision syntax is too labor-intensive. The vast majority of FAST headings end up being just a single term or phrase, thus, in theory, allowing even non specialists to assign them to bibliographic records.

The test database shows how LCSH subject headings were broken apart and converted into FAST headings. Daniel presented results from a few simple tests he performed on this database for his SAC Subcommittee at ALA. Searching the combined (Boolean) keywords “Connecticut” and “Newspapers” both in FAST and LSCH indexes, he found that the FAST index produced more comprehensive results. The reason for this, it turned out, was that FAST regularly converts abbreviations like “Conn.” into the full form of the name, thus better supporting non-abbreviated search term queries.  He also searched the combined keywords “Clarinet” and “Jazz” in both indexes. In this case, it turned out that LCSH produced more comprehensive results that did FAST. The reason for this was that LCSH included some bound phrases (i.e., those precoördinated through internal grammar, not through subdivision) that could not be converted algorithmically into equivalent FAST terms. So, for the LCSH headings “Clarinet and piano music” and “Concertos (Clarinet with band) (Jazz)”), no corresponding FAST heading was generated. Daniel is attending the ALA meeting (and FAST, etc., subcommittee meetings) in San Antonio, and will report back to the task force on FAST project developments.

Meeting adjourned at 3.45 p.m.

Back to top