Saturday, May 20, 2006

Search world less muddy (slightly)

We've learned quite a bit about the capabilities of modern search engines. I'll be attending the Enterprise Search Summit next week in New York, so I'm hoping that will clear up a lot of the nuts and bolts of it all. Sessions on constructing a taxonomy, etc.

While something like Autonomy would be great, we really don't have the resources to put it together just yet. Instead, we're probably going to go with more of the mid-range offerings. Leading contender at this point is Northern Lights. They are very low on the price scale (usually a warning sign), but I get good vibes from them every time we talk -- I have a very scientific vendor selection process, no?

They started as one of the dozen of Internet search engines in the 90s. Since then, they've gone through a few iterations to their current positioning as a Linux specialty search engine. It may be because they've always had an engineer on the phone call, but they strike me as a group that is passionate about search technology (weird as that may seem). Their engine seems to do everything that the ones charging $50k per year can do and they're genuinely interested in coming up with novel new uses for it. With the low price point, I see lots of potential to experiment with different options and gadgets on the search.

Back in the saddle

My apologies to my thousands of readers, I've been a bit slackery.

The audience driven content ideas are much more fully formed now than they were at my last post almost two months ago.

One lesson we learned is that blogs are VERY labor intensive. The AMA pub blogs are dying on the vine from lack of host activity (editors are just not bothering to post like they should, so no wonder there's no activity). As a result, we're approaching this from more of a team blog perspective -- either having multiple authors posting or using an RSS feed from an existing blogger to pull in commentary on the subject. I think that will work out a lot better.

Ran the plan by Toby Bloomberg and Dana Vanden Heuvel -- both were enthusiastic and offered helpful suggestions. In particular, they liked our idea of tying reader feedback and the site search algorithm into the quality assessment process. Also, the search alerts (either e-mail or RSS) to let users define search strings on the site and get an alert when new content meets those criteria.