We've spent most of the afternoon testing out a new search engine called Teoma. You may have heard of it but it seems unlikely. Only this month did it hit the search engine industry's consciousness (it first appeared in May apparently), so we suppose its techie sites like us and then the mainstream.
So what's the fuss about? Well, it looks as though it may give Google a run for its money. It's certainly improves on Google's methodology in one sense but it may end up being the ideal search tool if you know exactly what you are after.
We should explain here that we are converts to Google. It is easily the most efficient search tool on the Net. But then, as journalists, we invariably have a precise idea what we are looking for. Google works by taking your search string and then scouring the Internet for sites with those words. Then it counts up the links on other sites to those sites. The more links, the higher the ranking. It's a clever system and it works extremely well.
What Teoma does is extend this a little by finding the sites that concern themselves with the search string. Then it takes its rankings from the number of links within these sites to a particular site. It could probably be more easily explained by saying that Google asks about a certain expert in a field and then goes and takes a poll from people in the street over which one they think is best. Teoma would ask all the experts in the field which one of themselves they think is best.
The academic analogy is chosen on purpose because we reckon that rooting out precise information is where it comes into its own. It also offers three types of search result: normal rankings, Web pages arranged by topic and "experts' links".
We tested the site extensively and have to say, we are pretty impressed. Some abstract but extremely relevant links are found whereas on Google, they are stuck further down the line. It is also, due to its setup, far better at finding news stories. The topic folders are good and do take a step sideways, unlike most efforts at the same thing. The expert links run hot and cold. And of course it's fast because no one is using it at the moment (you remember when Google was faster than shit off a shovel?).
However, all is not sunshine and light and it has some way to go to match up to Google. First of all, advanced searches are non-existent (your Boolean etc). There is no cache, which is something we've been using more and more on Google. It has a limited database at the moment, which is occasionally apparent and a little annoying. It also, because of its approach, has a tendency to send you to Internet lists of information rather than direct to the original source.
However, Teoma definitely shows promise. Besides, it'll be nice to see Google have some competition.
Before we go, a very quick history of Teoma. It came from a project started in 1998 at Rutgers University (New Jersey). That project was headed by Professor Apostolos Gerasoulis, who is now the CTO. Professor Tao Yang from the university of California is VP of research and development.
It has got some private funding from Hawk Holdings - a joint company set up by Qwest Communications International and Baxter Investments - although we don't know how much, and is looking to sign up portals etc at the moment. It's still in beta.
Check it out at www.teoma.com. ®