On March 29th and 30th, I attended the 16th North Carolina Serials Conference for the first time in three years. A few years ago, the registration fees jumped considerably from one year to the next, and the content did not seem strong enough to justify continuing to go. Thanks to a bit of prodding, I attended this year and was delighted to find that the depth of the content was greatly improved and fully justified the registration fee.
Calhoun and Reynolds on Catalogs
The most interesting sessions at the conference were the opening keynote and the closing keynote, which both complimented and contradicted each other. Because they are so linked in my thinking, I’ll write about both sessions in one entry. The opening keynote was by Karen Calhoun of Cornell University, and incoming Vice President for Cataloging and Metadata at OCLC, entitled “The Changing Faces of Catalogs: Accelerating Access, Saving Time,” and the closing keynote was by Regina Romano Reynolds, head of the National Serials Data Program, the U.S. ISSN center at the Library of Congress, entitled “To Boldly Go: Transforming Catalogs and Cataloging to Meet User Needs.”
Calhoun based most of her presentation on a report she prepared on behalf of the Library of Congress called “The Changing Nature of the Catalog and Its Integration With Other Discovery Tools” which was released about a year ago. Calhoun argued that we need to rethink the catalog in light of a changed information world and that users are not getting what they need from the OPAC. As information systems, catalogs are hard to use, and, in an academic setting, are used primarily by faculty and graduate students, with fairly low undergraduate use. She discussed many of the known features of Millennial students, that they enjoy interactive tools, collaborative learning, etc.
Calhoun then discussed the results of a poll question posed to students, “Where do you begin a search for a information on a topic?” 84% said web search engines, while only 1% said library web pages. This struck me as exactly the sort of poorly worded poll question that is designed to induce panic in the library world. I find these sorts of apocalyptic pronouncements about the future of libraries to be quite tiresome and far off the mark (our profession seems to take a perverse joy in predicting our own demise). The question doesn’t allow for comparison to information seeking behavior in the pre-Web age, doesn’t consider that most information needs are fairly shallow and don’t require a library, etc. Lest I be accused of explaining away the problem, Reynolds addressed these very issues in her presentation, which drew heavily from a study conducted at Ohio State called “Sense-Making the Information Confluence: The Whys and Hows of College and University User Satisficing of Information Needs”. Reynolds pointed to survey data that showed that students go to a variety of information sources teachers, friends, “Dad,” etc. and that many information needs are simple to meet. We will never get back patrons to answer simple ready-reference questions in the same numbers we had in the pre-Web days. However, when it comes to in-depth, thorough searches, the library was the number one choice. And many people still trust library sources more than random things found on the Web.
Calhoun claimed that library catalogs must come to terms with the facts that content has changed, user needs have changed and search behavior has changed. Accordingly, she argued that catalogs should incorporate social networking software, to allow for simple things like “people who searched this title also searched” functions as well as more complex models where the configuration and design of the catalog adapts and changes based on the what users search for and what they do with the catalog. Calhoun also argued that libraries should try to pursue a user-centric model and push our services out, making our catalogs visible to users where they’re searching. That is, that integration should be outward rather than inward.
While pushing our services out to users does make sense, Reynolds pointed out that libraries may not get credit for this work. Students often use databases provided by an academic library and don’t even realize that the library is providing the service, even if it is clearly marked and branded. Reynolds essentially agreed with most of Calhoun’s comments about the need to change the catalog, but gave more concrete examples of changes that could improve the search experience, such as spellcheck, word stemming, relevance ranking, better labels, reduced jargon, utilizing serial links already present in bib records, using FRBR granularity, icons to represent formats, “word cloud” results a la AquaBrowser, adding tables of contents, reviews, and recommendations, improved subject access, metasearching functionality, user tagging, folksonomies, etc. Reynolds also argued that one thing library catalogs do very well is to collocate large amounts of information and that we should retain and expand on that function in future catalog models.
Calhoun agreed that library catalogs do certain things very well and will remain necessary for tracking local holdings for print books, serials, videos, etc. But, she argued that cataloging departments need to make much better linkages, need to simplify and exploit all sources of cataloging data, need to automate and streamline workflows, need to explore automatic classification and need to mine rich catalog data. Reynolds touched on many of these points and gave a concrete example of a project she was involved in that attempted to simplify cataloging data, the creation of the CONSER standard record. The goal was to create a less costly record that would be compatible with existing standards, applicable to all formats and focused on user needs. The resulting record stripped out some of the intricate detail of serial bib records and left only those fields that were genuinely of use to users. It has sparked some controversy in the serials cataloging world, but it represented a fairly bold attempt to respond to the changing library catalog, which was the shared theme of the two keynote addresses.
Empowering the Library Search Experience
In this session, Holly Johnson of the Howard County Library in Columbia, MD discussed her libraries implementation of an AquaBrowser based catalog, and Kristin Antelman of NC State discussed some of the issues with their implementation of Endeca.
Johnson began the session with an in-depth demo of AquaBrowser. I have to say it looked really nice. I particularly liked the “discover cloud” results that gives a cluster of related terms, which are each searchable. The system also has relevance ranking, the ability to refine options by format, and can support RSS feeds. You can look at their AquaBrowser catalog.
Antelman discussed the Endeca implementation and pointed out a few problems with the system. For one thing, relevance ranking can be tricky with serials. For example, a search of “New York Times” brings up the record for the newspaper in twentieth place. Why? Because the phrase “New York Times” can occur many times in the bib records of related titles like “New York Times Book Review,” because of title changes, related titles, key titles, etc., and the more times the phrase appears the more “relevant” the record is. Also, the natural language query functionality of Endeca doesn’t necessarily effectively use the LC Subject Headings. A search on the phrase “causes of the Revolutionary War” brings up only three records (and two of those are for the same title in print and on microfilm). But, using a subject authority search brings up 388 records. The shallow skimming of the LCSH by Endeca can be really shallow.
Other Thoughts on the NC Serials Conference
I also went to a panel discussion on institutional repositories that drew heavily on a preconference that I didn’t attend. The panel boiled down to the following:
- Pro: Institutional repositories can change the scholarly communication model by making faculty works publicly available, they can make material widely and permanently available, and they can preserve at-risk material.
- Con: A lot of universities talk about wanting to have IRs, but few actually do. The IRs that have been created are lightly populated with material, and very little faculty-produced, peer-reviewed material is available on IRs.
- Response: Perhaps IRs don’t have much material in them because they’re still in the embryonic stage. When the Web was new, there was very little useful stuff on it, and many “experts” said that there would be little need for most universities to have anything up on the Web. Give IRs some time.
And I went to a session that introduced me to SUSHI, or the Standardized Usage Statistics Harvesting Initiative Protocol. The talk was a bit over my head, but I at least got that SUSHI is a standard that vendors and other content providers can use for communicating usage statistic reports (COUNTER reports) to libraries. Beyond that, I get a little fuzzy.