[thelist] Site search options

[Paul Bennett]

Hi Randal,

Good on you for looking at a cheaper option. However, may I share some bad experiences we've had with some of the 'free' solutions out there:

Caveats - the site I work on is large (around 1.6Gb of content, not including database data) and relatively high traffic, which may affect the performance of the solutions we've tried.

HTDig - spidered the search ok, but results were far from 'relevant', often giving thousands of results (dues to word stemming) rather then the half a dozen that it should have. Example given by my colleague was a search for 'oriental bay' which turned up results for: orient, oriental, orientation, orienteering.....

Mnogosearch - spiders ok, performance awful, results seem more relevant but the performance is so poor that the query either times out or you need to go and make your lunch while waiting for results. We had a db table error lately and found out when restoring from backup that one of the indexing tables has grown to over 40Gb...! Other people have had some success with this product, but I can't recommend it. In the meantime we're pointing our site search to Google.

Search software is a strange market with products ranging from free to $70K and it's often difficult to tell at a glance where the benefits lie.
After almost 12 months of off and on research, we're shooting for the Google Mini too. For the money, the benefits seem unbeatable. 

Paul