The Virtues and Limits of Cataloging

Christine points me to a post a by Eli Jacobowitz on The Virtues and Limits of Cataloging. From the post:

First of all, current implementations of tagging are "flat" – there is no meta-meta-data about what type of label a tag represents. Is "Mona Lisa" the title, author, location, genre, art movement?

I agree!

Second, and fairly obviously, tags are susceptible to spelling errors and multiple listings for the same category ("Italian", "italian", "italy", "Italians""¦ what do I search for?).

I agree!

Third, even if you know what attribute you're supposed to label, and have a controlled list of values to pick from, ambiguity may persist in how to summarize multiple or conflicting facts about the object.

I agree!

The bottom line is, you need a degree in Library Science to do this right.

I disagree!

What you need is training by an experienced cataloging librarian! I’m in a cataloging class right now and I’m enjoying it – but in the end I’ll never remember everything I’ve learned or learn everything I need to learn – only a veteran cataloger can teach me how to do it right and how to do it efficiently. It all comes back to what I’ve always said – you need hands on work experience to learn how to do a lot of the things librarians do – and most librarians I’ve spoken with agree.

That said (sorry for the rant). It’s a great article and I think an alternative that Eli overlooked is to have librarians work in conjunction with average internet users. This can be done many different ways.

One option is to have librarians create the controlled vocabulary that is used on a particular site.

Another option is to have librarian editors go through and clean up tags (yikes!).

Lastly, you could just have librarians and users cataloging and tagging items in tandem. Then you have the authority control and you have all those other words that non-librarians think of. This comes back to a bit of what Tim Spalding showed us in one of his presentations at CIL. The LCSH does not have a heading for CyberPunk – so how do you find a book like Neuromancer by William Gibson if you’re searching by subject? You look for:

Computer hackers–Fiction.
Business intelligence–Fiction.
Information superhighway–Fiction.
Nervous system–Wounds and injuries–Fiction.

Need I say anymore??

Technorati Tags:


  1. Tags are not subject headings. Neither should be seen as a replacement for the other. Tags help people refind things. They are personal, they should not be cleaned up, nor even suggested. What comes 1st to a person’s mind is most likely what they will use to refind the item.

    Computers can do interesting things with large groups of tags. Clustering, for example. That gives an interesting view of a information space. The ability to pivot from a tag to a tagger to their other tags is also a useful navigation tool. There is no training required to tag things in your own personal information sphere.

    One benefit of thesauri is that there is an hierarchical structure. Did something find too much, just get more specific. A search find too little use a brader term. They are all there to easily choose.

  2. I agree – but some librarians I have heard speak don’t like the fact that tags are so free form – and so the suggestions I listed above were some that I had heard at conferences and meetings over the last few years about how to find a nice balance between the strict rules of cataloging and the “no rule” form of tagging.

    That said, I’m in your ballpark – I like tagging things whatever I want – but in the end that doesn’t always help others find information in my collections.

  3. First, thanks for the comments!

    I hope that the rhetorical point about ILS degrees doesn’t overshadow the conclusion that cataloging the internet is a mismatch in scale — it will never be as organized as a library.

    So we make compromises. One compromised attempt at a controlled tag vocabulary is‘s suggested tags feature. I have found myself changing my tag vocabulary to better match their suggestions.

  4. I do the same thing (with!! You are right – we can’t catalog the internet the way we can our collections.

  5. While I think I would enjoy being one of those “librarian editors that go through and clean up tags”, I find the third option you presented most attractive–librarians and users working in tandem. I think library catalogs would have a much richer vocabulary if, along with subject headings, we allowed users to add tags to bibliographic records.

  6. I’m so pleased this example is spreading.

    LibraryThing is going to release a new “tagmash” feature soon, creating pages—not searches—combining tags.

    So, “Greek” and “history.” This can be better than the lesser pile of things tagged “greek history,” and really comes into its own with something like “france + wwii + memoirs” or “japan + love – fiction.” I haven’t “opened” it yet, however, because I wanted to add two things (1) weighting of factors, and (2) adding subjects and pieces of subjects too—”france + wwii + subject:congresses” or “subject:Nervous system–Wounds and injuries + tag:mountaineering.” The system naturally computes “relatedness” to “single” subjects, tags and to other tagmashes.

    The goal is to create something between the fluidity of the tag and the fixity of the subject—a fixed criterion for a fluid list—a fluid LCSH, a dynamic playlist, Listmania will legs!

    Still thinking it through…

Leave a Reply

Your email address will not be published. Required fields are marked *