[WEB4LIB] Re: Good search engines functionnalities for relevant

[Stu Baker]

At Northwestern University we have a campus wide committee charged to 
evaluate and recommend a search engine for use for the entire campus. 
After much evaluation, testing and review we finally settled on 
Inktomi Search Software. We looked at commercial products, open 
source software (including HTDig) and remote search services. 
Besides, scalability, ease of administration and good results, we 
found that Inktomi could be tuned much better than any of the other 
search engines we tested when using metadata. Their Content 
Classification Engine (CCE) is also something that is very 
interesting in terms of providing an automatic means of creating 
browse structures based on rules and classifications that you can 
define as "topics". They have a java based API as well so you can 
write your own hooks into the system.

Our Library has also purchased the Inktomi Search software to use as 
our search engine and  potentially for digital library initiatives. 
The ability for Inktomi to index XML document stores was a very 
enticing feature.

We are in the midst of finalizing our report which I would be glad to 
share when it is complete. In the meantime you can look at our 
prototype search site at <http://search.library.northwestern.edu/>.