I took time yesterday to visit a few colleagues at Ball State University.
![]() Ball State, the movie! |
Over the past few months the names of some fellow librarians at Ball State University repeatedly crossed my path. The first was Jonathan Brinley who is/was a co-editor on Code4Lib Journal. The second was Kelley McGrath who was mentioned to me as top-notch cataloger. The third was Todd Vandenbark who was investigating the use of MyLibrary. Finally, a former Notre Damer-er, Marcy Simons, recently started working at Ball State. Because Ball State is relatively close, I decided to take the opportunity to visit these good folks during this rather slow part of the academic year.
Compare & contrast
After I arrived we made our way to lunch. We compared and contrasted our libraries. For example, they had many — about say 200 — public workstations. The library was hustling and bustling. About 18,000 students go to Ball State and seemingly many of them go home on the weekends. Ball State was built with money from the canning jar industry, but upon a visit to the archives no canning jars could be seen. I didn’t really expect any.
Shop talk
Over lunch we talked a lot about FRBR and the possibilities of creating work-level records from the myriad of existing item-level (MARC) records. Since the work-related content is often times encoded as free text in some sort of 500 field, I wonder how feasible the process would be. Ironically, an article, “Identifying FRBR Work-Level Data in MARC Bibliographic Records for Manifestations of Moving Images” by Kelley had been published the day before in Code4Lib. Boy, it certainly is a small world.
I always enjoy “busman’s holidays” and visiting other libraries. I find we oftentimes have more things in common than differences.
Tags: Ball State University
The conclusion I come to reading that article is “not so much.” They sort of kind of got it to work–by taking an awful lot of free uncontrolled text in the MARC and applying heuristics to it.
To be honest, “sort of kind of” is about all we were hoping for. It’s better than zero.
For some materials, there are probably alternative ways of doing this (say, getting a dataset from someone who already has work-based records and just matching them to manifestation records), but not probably for the more obscure stuff. We either try to get it out of the MARC records automatically or we do it manually or we stick with what we have now (which doesn’t work so well).