Long live indexing! You still need to to find stuff.

Music music, originally uploaded by angela marlaud.

This is a picture of the library at KALX. It’s my work away from work. Seriously. See that collection? It’s all alphabetized, and that’s it. Finding stuff is easy if you know we have it and what it’s called, thanks to an army of volunteers who spend hours every month shelf-reading. (Seriously, thank you!) The problem is discovery… we’re really bad on that, but alas… we have other things to worry about now… like running out of space. Don’t worry, we’re on it.

But think about what if my real work organized its collection like this… that would be horrible. How would anybody find anything? Luck? Memorization? Google? This is sort of how sometimes I think Google Scholar or article databases might work. It’s all just a big drive of loosely organized stuff waiting for you to ask for it specifically or to be stumbled upon. There’s a lot of “luck” and you might stub your toe.

For a while it seems libraries have been kicking around the idea that indexing is dying. The future will be the semantic web and indexers will be out of a job. Artificial intelligence will handle that or you won’t need it because you can search the full text. The future is going to be rad! (For more about the future of libraries, read Rochelle Mazar’s awesome post about it.) As an indexer and searcher, it makes me worried. Is indexing pointless? Well… bad indexing is, but that’s another story. (Some of the worst indexing I’ve seen has to come from machines, if it’s from people… yikes!)

Well, I had a moment a few weeks back that restored my faith in indexing. A student came in with a fairly straight forward question. They needed an engineering equation to analyze. No specific equation, just an equation from an engineering analysis that they could then discuss one of the variables. How do you find that? Scour engineering analyses? No need. We have mathematical models. That one term saved the day for all of the students working on this project, and it’s not really something I think our current or proposed methods for automatic indexing could easily pick up. Score one for the carbon filter. Keep on indexing.