sxsw 2006 update

Yes, I made it safely to Austin. Blogging silence was due to massive overcommitment on Saturday (I ended up speaking on three panels), and then massive decompression on Sunday (I didn't even attend one panel, though I did make an appearance at a couple of receptions before turning in early again).

The panel I was most worried about, the book digitization panel, went (I thought) extremely well. Daniel Clancy from Google, Bob Stein from the Center for the Future of the Book, and Danielle Tiedt from Microsoft's book search program were my fellow panelists, and we tried to involve the audience heavily in the conversation. We capped it off with a delicious bbq lunch at Ironworks. The panel I didn't expect to speak on was danah boyd's "Designing for Local and Global Social Play," which went really well. The highlight was a game that we played called "the secret game" which I'll describe in a separate post.

Today I'm trying to shift back into conference mode, starting with Peter Morville's talk on "ambient findability" (the subject of his new book, which I really need to get and read). Peter and I went to the same library school, a few years apart, so while we don't know each other except peripherally, I'm always happy to see how well-respected his work and ideas are in the tech community.

As usual, Peter's doing a good job of using PPT as a way to show useful graphical examples, rather than bombarding us with bullet points. Yay.

He's talking about search right now, and about search as a system. He notes that companies spend more time on tweaking the search interface, and not nearly enough on the results interface. This includes not just ranking and clustering algorithms, but also the interface design that lets you pick out key information from the results screen.

He cites Marcia Bates' work on information seeking behavior, and points out that it's much more complex than search engine designers typically recognize. Search is an iterative process--searches don't exist independently of that iterative, linear process.

Talks about the problem with the term "usability"--what does that mean? What are the components of a good user experience? (Lists useful, usable, findable, valuable, desirable, accessible, credible as a few criteria.)

(I'm not providing a lot of the details, because I'm assuming that much of this is in his book...)

He shows some examples that he provided to NIH regarding searches for cancer-related information, and notes how many searches on cancer are done in public search engines as opposed to on their site, which doesn't come up high in the rankings. (Where did he get these search term frequency figures from August 2005? Isn't this the kind of data that the government requested and that everyone was so concerned about being provided?)

Finally gets into the ambient findability piece (halfway through the talk). "The ability to find anyone or anything from anywhere at any time." Hmmm. That doesn't sound so great to me, actually. The panopticon has a very dark side. (He does acknowledge that, to some extent, though without as much reservation as I feel at the prospect.)

"A wealth of information creates a poverty of attention." (Herb Simon) [searching on that phrase pulls up some interesting sites on attention...]

Shows a device intended to be locked onto a kids' arm to track their location and give you access via a web interface. Discusses the strong emotional response people have to this. Interestingly, he says, reviews on Amazon complain primarily about it not working well enough, not about potential privacy issues.

Cites David Brin's book The Transparent Society, with a long quote on "reciprocal transparency" that I don't have time to write down. (I didn't know about this book; will need to track it down)

What's going to solve information retrieval challenges, help us make the needle in the haystack bigger? He doesn't think it's going to be artificial intelligence (ms bob isn't going to save us), nor is it going to be information visualization (which is beautiful but not always useful; shows treemaps as an example). Even with maps--the applications are beautiful, but often not nearly as usable as we need. Most people jsut want to get from point a to point b--and they use the text directions more than the map itself. Collective intelligence and user participation need to be in the solution. (Hmmm. I can think of refutation to this...)

So, who's going to help us? the librarians, of course. (Puts up a slide with "Revenge of the Librarians"--where have I heard that before?) Has the web turned us all into librarians? ("Metadata is sexy now!" he proclaims.)

We should be careful not throw the old ways of organizing information out as we adopt the new ways. Quotes David Weinberger: "The old way creates a tree, the new way rakes the leaves together." What happens to piles of leaves, he asks? They rot, and become food for new trees.

(There's more to this presentation, but I tuned out for a bit while i dealt with incoming IM and email and text messages... My guess is this is one of the things that will end up with the podcast up on the sxsw podcast page, so you can watch it yourself if you're interested...)