There Is No Cat

Boogle-oo

Lou Josephs comments about the new web search engine provide by the BBC:

The BBC decided they wanted a search engine that did what they needed and brought up more information about uk topics. So they have their own customized google.

Google appears to be exactly what you get when you search via the BBC. Compare the search results for BBC World Service via Google and for BBC World Service via BBCi. Damned near identical. Same sites, same order, down to the fact that savebbc.org comes up as the number seven link on both.

One of the designers of the BBC search engine, well-respected blogger Matt Jones, gave a fascinating presentation about it at the ASIS&T Information Architecture Summit in Baltimore in March. You can download the PowerPoint file he used for the presentation from the front page of his site. You don’t get the benefit of his dry, mordant wit, but hey, there’s got to be a reason why it’s better to actually go to the conference. Matt mentions licensing as one possible strategy in the presentation.

Meanwhile, The Independent questions what the heck the BBC is doing in the search engine business anyway in an article entitled "Dyke said the BBC should focus on programmes, so why is he taking on Google? Search me...." They make explicit what we have divined on our own: the WWW part of the BBC’s search is licensed from Google. The BBC claims to exclude "derogatory" sites. Interesting that they kept savebbc.org in their results, then. (Well, really, we at savebbc.org love the BBC, that’s why we want to save them from themselves. No derogatory intentions.)

The Guardian, meanwhile, plays it a bit straighter and reports about other people being mad at the BBC for usurping a function from the commercial world. Mama BBC is going to take care of the British public, insisting that "[i]t is quite clear that the current search marketplace doesn’t have the needs of internet users at its heart." Curious then that they would serve the needs of Internet users by, um, replicating something that already exists.

Okay, that’s not fair. Matt’s presentation at the Summit made it clear that the BBC is in fact engaging in a value add proposition. The search I linked to above for BBC World Service doesn’t demonstrate it, but searching for Zimbabwe on the BBC’s search engine brings up an interesting feature: results from BBC News and Sport provided next to the regular engine’s results. And then there’s a very subtle refinement to Google’s results here as well: the first result has a bit of red text next to it that says "BBCi recommended". If you look at Matt’s presentation, you’ll notice that there’s actual human input into the search indexing process. BBC has a dozen people in an office in London beavering away, picking the best sites for given keywords so that when you search on one of those keywords, those hand-picked results show up at the top with the "BBCi recommended" tag added on. Theoretically, it’s an important addition. But most of the literature I’ve seen in the field deals with adding metadata and directing searches for comparatively limited sets of data. I’m not sure this approach scales well to the level of The Entire Internet. Maybe it does; conceptually, this is similar to what Yahoo does, where they search their own lovingly hand-crafted database first, then pass the search off to Google for the wilds of full-text search of the net as a whole (or at least that portion indexed by Google). But hey, if the British license fee payer is willing to foot the bill for the experiment, what the heck. Just don’t tell The Independent.

Posted at 11:17 PM