Google Book Search slammed

A very interesting article from Thomson Gale and  Peter's Digital Reference Shelf which absolutely slams Google Book Search. It really is worth reading all the way through, and it's going to make very unhappy reading for the people over at Google. Bottom line is... it's badly broken. (Which does rather imply it might actually have worked once, but there ya go...)

What's the problem with it? Well... it's the poor relation in comparision to the greater functionality of other resources such as the Gutenberg Project, the Million Books project etc. The software is shot; Boolean doesn't do what you'd expect (which may actually be a problem with the original counts, and Boolean operators might actually be getting it right), limit fields don't work, confusing hit counts, poor content, the database size is questionable, the sources are sometimes strange etc.

None of this actually comes as any sort of surprise at all. The fact of the matter is that, and let me state it as plainly as possible Google is broken and doesn't work properly. It's not just Book Search, it doesn't work properly in normal search either, and this is something that I've banged on about before.

I really would recommend that you spend five minutes or so reading the original article - it does come from Gale, so there is the potential for bias there, but it's difficult to see exactly where when facts and figures are offered and results are replicable.

Comments

[walt crawford]

Without disagreeing on Google: Are there any of the major open web search engines that you regard as less broken? An honest question for which I'm not sure I have an answer...

[Phil Bradley]

Good question. I also like the phrasing 'less broken', with its implication that they're all broken to a greater or a lesser degree, something that I'd agree with.

I'd go with Exalead if pushed. That is to say, I put (a little) more trust in that search engine than I do in Google. This isn't to say that Exalead doesn't have its faults as well, because lets face it they all do. They are also more responsive than Google as well. I could also make the same case with an engine such as Clusty - they had a bug in their Shakespeare system that I pointed out to them and it was fixed inside 24 hours.

Consequently I don't think we can really have a discussion on what is/isn't broken until we've defined exactly what we mean (so I really should have done that in the first instance) by 'broken'. Part of that is certainly that search functionality doesn't do what it should, results are demonstrably incorrect and so on. Another part however is a much more difficult element to describe, define and quantify, and that's to do with a level of trust, the extent that problems are identified and acted upon and so on.

I see nothing to make believe that Google is taking any real notice of basic flaws in its systems, so that that extent I feel justified in saying that I think Google (both as a search engine, but more importantly perhaps as a company) is broken. And broken rather more than their opposition.