Tony Conrad, CEO of Sphere, came by Yahoo! on Friday to give a few of us a look at the blog search technology they've been building.
He started off by explaining the company background (ties to Oddpost) and the small team that's been tackling the problem from a different angle. Unlike other engines that focus on trying to index every single blog, deal in link ranking, etc, Sphere puts an emphasis on trying to understand the content of individual posts. With that understanding, Sphere can then be smarter about finding related content from other sources, including some mainstream media.
I played with the engine a bit as Tony talked and tried to poke holes in it. Searching for "viagra" and "blog" didn't returned results that could have been far, far worse. Vanity searches worked quite well. Topical searches turned up relevant content rather than being completely biased toward the newest content. (You can sort by date, but it's not the default.)
Okay, enough background and theory. Is it good?
The interface, designed by Adaptive Path, is quite clean and understandable (see screenshot). They've added some nice features, such as a profile link that produces a small DHTML overlay with some vital stats for each blog: average posting frequency, posts per week, blogs recently linked to, etc.
In my testing so far, I see a lot of promise. Their technology seems far more splog (spam blog) resistant than many of the other engines. They don't actively filter it out, but the spam blogs end up being ranked so low that you rarely encounter them. That sounds like the right approach to me.
They currently don't offer several things that bloggers and journalists will likely want:
- Link-based searching
- RSS feeds of search results
- The ability to "claim" your blog
All of those are coming, if I recall Tony's answers correctly.
I'll be using Sphere more in detph in the coming days and weeks to get a much better idea how it stacks up against the other players. But so far I'm impressed.
Posted by jzawodn at October 17, 2005 01:24 PM