Category Archives: linked data

More on Serials and Linked Data

Last year I wrote an article on serials, FRBR, and linked data in the Journal of Library Metadata. My main goal was to re-think how libraries can make connections between articles and the journals in which they’re published using linked data. I used the FRBR model to link the article and the journal together at the Item level, envisioning both the article and the journal being positioned as Works.

I never felt entirely happy with my model, but I couldn’t figure out a better way at the time. I recognized several months ago that my thinking, when I wrote the article, was limited because I was focused on trying to create some kind of symmetry in the model.

Recently, I came up with another way to think about connecting journals and their articles, still using the FRBR model, and I think this makes a lot more sense. In my original article, I looked at the journal from a FRBR perspective and saw each individual issue of a journal as the Item in the FRBR hierarchy. But it was awkward, and I don’t think it worked particularly well.

In re-imagining this, however, I realized that an individual issue of a journal is really an expression of that journal.

A visual diagram of the FRBR hierarchy for a journal and an article in that journal

Serials FRBR model to link articles and journals together

The journal itself (“The New Yorker,” “The Paris Review,” “The New England Journal of Medicine”) is a work; it is a conceptual thing that doesn’t have expression outside of the issues that are published as part of its run. Each issue that is published is another expression of that journal. Similarly, if you think of an article as a work, they are published as an expression in a particular issue of a journal.

I think this model works much more organically, and makes a lot more sense that what I was originally trying to force to make sense because I was fixated on symmetry.

The other question I asked in the article was how we can deal with journal changes using linked data in the FRBR model. Merges, splits, and title changes can still create problems for someone in a library trying to find a particular resource. But I think linked data itself can solve this problem, without us needing to change the FRBR model by creating something like “super works” or “journal families.” We have a good way of linking former and succeeding titles together, but it doesn’t work as well when our metadata is contained in independent catalogs. However, if our “records” exist on the web and are openly linked, we can link to a former or succeeding title even if it’s not held in our own unique collection.

I don’t know if an idea like this will be picked up by the people who are currently arguing about the models we should use in a linked data environment. I suspect it’s too simplistic for them, which is what makes it appealing to me, but catalogers seem to like to make things as complicated as possible. But I felt that the niggling annoyance about my previously published model disappeared when I started thinking about linking resources together this way.

I’d love to hear your thoughts. Do you think this model makes sense?