My recent week chock full of Code4Lib goodness was sponsored by Amanda Vizedom of Wind River Consulting. She’d been lucky enough to grab one of the precious few available registrations in the hour or so before they sold out, but she couldn’t attend. She generously offered up her spot to a student who otherwise would be shut out, and my friend and favorite infod1va spoke up on my behalf. I got to hang out with the cool kids!
They’re all the cool kids of their industry now, for sure, but given the nerd quotient in the room, I’m fairly sure no one there qualified as cool back when they really were kids. And that’s a good thing.
Anyhow, it was a fascinating experience! I can’t remotely keep up with these people in the code-monkeying department, although I can just barely hold my own in the beer drinking. I’m rather proud of the latter, as it’s a definite accomplishment — these people set a strong pace. And I’m inspired to build out some skills in the former department, just because so many really impressive things are possible. All the ideas in my head will come to be in the world only if I make it so; nothing will happen if I wait for the universe to manifest the code.
Individual highlights for me maybe start with a presentation on the search infrastructure at the Hathi Trust, which gives access to 450+TB of digitized scholarship through the same sort of html search interface that turns up last week’s top 10 music CD at lesser websites. Later informal discussion on search usability was productive, too, and I eagerly hope that Tom Burton-West will lead us in reviving the C4L mailing list on search usability, given his demonstrated level head and deep understanding of the implications that small design choices can have.
I also appreciated the discussion of relevance ranking in a presentation by Mendeley on applications of its platform. They make some useful points about inferring relevance by connections to other users and the resources they prefer, a point that’s clearly not lost on Google, considering the G+ annotations on its search results. Judiciously applied, this could be a useful layer in filtering people’s access, but it’s subject to selection bias in a big way. People we know already share many of our interests, so the precision gained could come at a huge cost in lost recall. It won’t succeed if it’s done in a way that narrows the radar screen, but it could help to brighten some of the best blips out there.
Other useful stuff included a discussion of microdata and schema.org, which seems like a fine, lightweight way to annotate content with useful metadata that cues searchers where to find the stuff they want and also supports outreach by content providers. In particular, I see lots of potential for raising the visibility of libraries out on the wild, wild web. I also enjoyed a review of Harvard Library’s StackView project, which tries to visualize the complexity and popularity of resources using only basic html and css. Simple concept, fancy result!