Don’t be evil


Not a lot of people are aware of Google’s corporate motto, don’t be evil, but it seems that a lot of librarians actually think that Google is doing the exact opposite. In fact, a while ago I suggested dropping all of our bibliographical databases in favour of Google Scholar*, the response was of course underwhelming, but a few people were really dead against the idea and not for a good reason — like it not being any good — but for ideological reasons.

In this case, I think that ideology is a dangerous thing because it allows us to think of Google and its ilk as “the enemy”, which they’re clearly not: they’re just a bunch of folk who manage to make information appear accessible to the vast majority of computer literate people. The thing that bites is that they’re doing this better than the so-called information professionals (read: librarians like us). I know that a lot of people have said this same thing before, but we need to learn from Google, figure out what they’re doing well, simulate this and improve it (I’ve written about a presentation at ILI2007 that made the point that Google doesn’t do everything well, so there is room for improvement).

One of the very interesting things that I’ve often come across is the perception that keeping track of what a user does in order to present them with a. relevant hits, and b. relevant advertising is particularly insidious. I probably think this myself on occasion. But, what’s the real problem: improving the results, tailoring them to a particular audience — isn’t this what we should be doing? I think we should.

Now, you don’t need to be a genius to recognize the value of Bayesian learning in information filtering. A simple implementation of Bayesian learning — attached by cookie and user profile (and perhaps IP address!) — in your OPAC might make the world of difference to the user experience for a lot of people. Imagine being presented with hits on the right topic even if you entered a hopelessly bland keyword. We’re not in it for the money, we’re just here to help, so this doesn’t seem quite so insidious, particularly if you offer users the opportunity to change their profile in a meaningful way.

It’s not evil, it’s sensible. Be sensible.

* Probably not a very good idea, but when you consider a. the state of play in bibliographic databases, b. the innovation that takes place at Google and c. the ability of most students to use plain old Google instead of a proper source, it maybe doesn’t seem so daft.



%d bloggers like this: