Quite a nice snippet from the php_london ML by Michael Stillwell
Site searches are really really bad. Even from organisations that should know how to do search. I really don’t understand why.
Try, e.g.
http://search.oracle.com/search/search?start=1&nodeid=&fid=&keyword=oracle+client+download&group=All
(Business Intelligence Beans Installation Guide and Release Notes as first result)
versus
http://www.google.com/search?q=site:oracle.com+oracle+client+downloadOr
http://www.mysql.com/search/?q=query+browser+download&charset=&x=0&y=0
(Greek as first result)versus
http://www.google.com/search?q=site:mysql.com+query+browser+download
July 2nd, 2008 at 1:43 am
Ha ha, I’ve been confounded by that exact same thing. There are a lot of other big name corporate sites with the same problem.
At the same time, as a developer myself I know how hard it is to create a good search routine/engine for one site (especially getting a good search result priority seems to be a problem), let alone for a cluster of different sites running different software/solutions.
Another part of the problem is that the users has become so accustomed to poor search engine results, that they stop using it, and rely on browsing the site instead to find what they are looking for. And since there isn’t any pressure to get better search results, developers can get away with half assed solutions.
I have also come across some people that seems to think that search results are always based on a single “magic” search algorithm, which doesn’t have to be adapted to the current solution. Many systems come with prebuilt search engines that are way to broad, and try to cover everything – but ends up doing nothing quite good enough.
Well, well. Enough ranting 🙂