At the Jefferson Library, we’ve been on a quest for several years now.
As a major information gateway and one of the chief organizers of knowledge on the life, times and legacy of Thomas Jefferson for the Thomas Jefferson Foundation, our key goal and indeed mandate is to provide scholars, teachers, students, and a worldwide audience with access to the universe of knowledge and research on Jefferson. With the enormous growth of relevant information on the Web, finding search tools that automate and help organize an otherwise seemingly disorganized web of resources in multiple formats continues to pose a major challenge. It doesn’t help that while our information users are mostly interested in Thomas Jefferson, the third president of the United States, and often his home Monticello, his “essay in architecture,” they’re not exactly looking for Thomas Jefferson Medical College in Philadelphia, PA or Monticello High School in Monticello, MN to turn up in their search results of over 16.5 million Google hits.
Well, it looks like things they are finally a-changing. Kosmix is a new web aggregator that may actually put us one step closer to the one-stop web search utopia we’ve been searching for. I test drove Kosmix with a simple search phrase, “Thomas Jefferson,” and the topic guide that appeared after about 4-5 seconds is here. Right at the top, I get “At A Glance” that tells me that I’ve got the right Jefferson. I get a short bio and links to Grolier (which was unfortunately a dead link when I last checked) and the website for Monticello — call me biased, but Kosmix would fall by quite a few notches if Monticello didn’t appear at this point!
Okay, so that’s just basic. Nothing special so far. What’s exciting to me is what follows. There’s a results section for blog posts, images, videos, podcasts, newsfeeds, what they call Conversations (essentially Q & A from sites like AllExperts.com and Answers.com), web links, and even tweets from social networking site, Twitter. Yes, the results are far from perfect and don’t necessarily appear in the order I would place them in, but for a generalist, the results for Jefferson are actually not bad, especially the relevant video lectures and talks which turn up from YouTube and through the web video search engine, Truveo. There are also links to the article postings from Helium, the community publishing site. I didn’t mind as much the ubiquitous e-commerce links to Amazon, Ebay, Shopping.com or Audible.com, as they were fairly unobstrusive and actually relevant if that’s what one is looking for.
Kosmix characterizes itself as a browse engine, as opposed to a search engine like Google. At its heart is a categorization engine that crawls the web, categorizes and aggregates them, and then presents results in a magazine format in response to a user search. It uses these categorization algorithms along with editorial input to create these topic guides. How we could use a tool like Kosmix is to feature a link to its topic guides from our site, and possibly an RSS feed on updates if that becomes available. Depending on how Kosmix develops, this could potentially be a powerful way to pull in relevant web connections in any single-search (or federated) solution we develop in future for all things Jeffersonian.
There’s also the Related topic guides that appear on the right hand side. At first glance, I got really excited to see links to separate guides for Jefferson family members and even slaves. This is where Kosmix falls short. Granted I wouldn’t expect results to be as numerous and hence less relevant the further one gets away from the luminary Thomas Jefferson, but you would reasonably expect to get what you think you’re getting. Results for Jefferson’s daughter Martha Jefferson Randolph get mixed in with results for his wife Martha Wayles Skelton (Jefferson), and vice versa. Reliability is an issue. Bios for Peter Jefferson and other family links are drawn from Wikipedia, so perhaps that explains why Peter Jefferson’s parents were one Hugh Gubutt and Carmen Electra (!), at least according to the Wikipedia entry last indexed by Kosmix.
In addition, the topic guide for Thomas Jefferson lacks (at this point) top-level links to the Thomas Jefferson Encyclopedia, and fails to pick up content and blog postings from the brand new Jefferson Today site, both sources originating from the Thomas Jefferson Foundation. The Encyclopedia is a resource that provides authoritative and trustworthy information about Jefferson and his world. On Kosmix, the Encyclopedia is featured in the topic guide for a search on “Thomas Jefferson and sheep,” but missing for Thomas Jefferson himself. The Jefferson Today site launched in mid January 2009 examines Jefferson’s ideas and their relevance in our contemporary world. Somewhat of a disservice to not list these resources if you’re touting your site as “the place to start when you want to browse and discover everything the Web has to offer.”
Serious flaws no doubt, but I have to say that the aggregation I see on Kosmix is at a level that is a definite step up from the typical unorganized Google search results up to this point. I think the folks from Kosmix will be getting some feedback, and we’ll be keeping an eye on this site …