From Blog Herald:
Clusty is a cluster search engine. Carnegie Mellon computer science researchers began researching search clusters in the 1990s and eventually brought the first “high-quality text clustering search engine” online through Vivisimo in 2000. The idea behind clustering is to gather related information into groups or folders, thus directing the searcher to more specific information rather than just a big list. The result eventually became Clusty.
I tried Clusty out today, playing around with several searches I conducted recently for one of my blawgs, The SC Bankruptcy & Consumer Law Blog. Here’s what I experienced, in a nutshell.
Using Clusty For Law-Related News
One of the biggest topics in bankruptcy-land recently is HR 3609; here’s a recent post of mine on the subject.
I performed several searches on Clusty for each of the following assorted keywords:
- HR 3609
- Emergency Home Ownership and Mortgage Equity Protection Act of 2007
- Bradley Miller D-NC (NB: the bill’s original sponsor)
For purposes of this post, however, I’m only covering the first and the last searches, as they had the most interesting results.
For the “HR 3609″ search, we had some interesting results. The two paid links at the top of the page had nothing to do with HR 3609 (HP printer toner cartridge and basic human resources training from something called “EmployeeHandbooks911.com”). The top 4 organic links were related to the bill; number 5, oddly enough, was about the herpes simplex virus.
However, on the left side of the page, we see the “clusters” list. This list included several categories that could be incredibly useful in a blog search, as shown in this screenshot:

I can imagine this feature giving some welcome context to law bloggers. Being able to distinguish between links associated with the Act itself and those associated with foreclosure generally. Let’s test it out.
Testing the Clusters
I click on the first link in the clusters list: Emergency Home Ownership and Mortgage — an abbreviated form of the Act’s name. Here are the results:

The first link from OpenCongress.com is an information and status page about the bill. The second link from The Orator is the text of the bill. The third link is a call to action from the Center for Responsible Lending.
What about the “Legislation/Law” cluster?
The first link is the text of the bill, a link that was #2 under the first cluster. Number 2: again from Open Congress, showing a collection of news articles about HR 3609. Third up is what looks to be the suggested text of a letter opposing 3609 from the Arkansas Bankers Association. The fourth link is a PDF from the US International Trade Commission — a memo on “proposed tariff legislation.”
Let’s test the “Foreclosure” cluster next.
It looks as if these links do indeed focus more on the effects of the bill on the foreclosure crisis. The first two links are from a blog linked to something called the Philadelphia Unemployment Project, and both advocate for passage of the bill in light of its effect on homeowners facing foreclosure (among other social effects). Third is a press release, also focused on the bill’s effects on foreclosure; fourth is from Banks.com but also stresses the foreclosure link.
“Blog” Cluster: Helpful or Misleading?
Now we’re getting somewhere:
These links all tag off the word “blog,” or some variation thereof, in the linked-to page’s title. This could be incredibly helpful for bloggers who want to seek out what other bloggers have to say on a particular issue. That’s a necessity for bloggers who are writing on any debated issue; presenting the “other side” is a critical part of any attempt at persuasion, after all, as well as for any attempt to cover an issue impartially.
So let’s see if that Blog Cluster repeats itself with another search.
The “Person” Search
Well, drat:

No blog cluster. (And I checked the full list. Nada.)
However, looking at the main results, as well as the clusters that do show up, it’s easy to see how this particular search could be useful to someone blogging about the legislator’s background and voting record.
As for the blog cluster, it may well be that no one’s blogging specifically about Representative Miller, apart from his connection to the main topic of HR 3609. So, don’t rule out the utility of blog cluster just yet.
Conclusion: Handy Tool for Bloggers Of All Stripes
Will Clusty replace Google? I don’t think so. For one, the results in many instances seem to mirror, if not straight out duplicate, Google results; at least, they’re not so different for some searches to make a meaningful difference. For another, Clusty is a good bit slower on my machine (a 13″ MacBook early ‘07 model with 2 GB memory) than Google is; analyzing those clusters might be quite a task.
But as an additional tool in the blogger’s toolbox, it’s a valuable addition, and definitely worth playing around with from time to time.
Share and Enjoy:
These icons link to social bookmarking sites where readers can share and discover new web pages.