UVa Blacklight – faceted browsing and prospect
If you were at code4libcon you’ve already seen Erik Hatcher’s initial foray into the world of indexing library catalog data with solr. I am pleased to announce that the unofficial UVa team we’ve assembled has come even further with that effort, and for the next couple of weeks we are demonstrating Blacklight, UVa’s solr based faceted OPAC. (Solr? UVA? Blacklight. Get it? Erik came up with it. Pretty clever, no?)
You can take a look here: http://aleph.lib.virginia.edu/blacklight/
We know we have some problems, e.g., our marc data isn’t utf-8 and that caused some trouble with the diacritics. There is currently no advanced search, but you can mimic some of that functionality by searching for author_text:glass if you want the name Glass only in the author field (similarly, title_text:glass to find it only in the title field). We need significant data cleanup and rationalization to make the facets user friendly. But overall, I’m impressed with what we’ve been able to accomplish in a few months, with no offical committment of staff time, using only our spare time, borrowed hardware, and good will.
Ultimately, we do not want this to be a rogue effort, and we’re hoping to find common cause with LibraryFind, the eXtensible catalogue folks, and/or many of the other excellent open source faceted opac projects happening in code4lib land.
We’ve indexed ~3.7M marc records + 320 Tang Dynasty Chinese poems (from a website you can find here). The marc records are indexing through a ruby script Erik wrote, directly from marc, without transforming into marc xml first. The box it’s all running on is a development server, with four medium fast CPUs and 3.5G of RAM. Solr is running through Jetty, Blacklight is running on a pack of four mongrel instances, with apache doing load balancing out front. Sessions get written to a mysql database, and we use capistrano for deployment and versioning. And it sends us all an email every time an error gets generated.
I have long been fascinated with the idea of giving users prospect on large collections of data. He hasn’t published a book yet, but Stan Ruecker‘s dissertation Affordances of Prospect is an excellent discussion of this idea, and is where I was introduced to the concept. What I like about faceted browsing of our library collections is that I start to get that feeling of being up on a hill, able to see everything laid out in front of me. Prospect.