FUDforum
Fast Uncompromising Discussions. FUDforum will get your users talking.

Home » FUDforum Development » Bug Reports » Search function -- inconsistent behaviour
Show: Today's Messages :: Polls :: Message Navigator
Switch to threaded view of this topic Create a new topic Submit Reply
icon5.gif  Search function -- inconsistent behaviour [message #15931] Fri, 09 January 2004 20:47 Go to next message
wfjmueller is currently offline  wfjmueller   Germany
Messages: 95
Registered: December 2003
Location: Darmstadt, Germany
Karma: 0
Member
I posted a message with the string (taken from the error log)
was not locked with LOCK TABLES

When I do a search for "was not locked with LOCK TABLES" the search function finds several postings, but not the one I posted today.

When I search for just "locked with LOCK TABLES" it does find my posting.

A search for "was" should find many posts because the word is very common, but the newest one is dated Tue, 14 October 2003. I checked for other usual 3 letter words, like has or had, and the newest post was always 14-17 October.

So far the observations, now some speculation:

It looks like, that around this time a kill list was introduced to avoid indexing of trivial words, and that this was done inconsistently, such that these words never go into the index for new posts, but are still handled as required when entered in the search field.

Re: Search function -- inconsistent behaviour [message #15965 is a reply to message #15931] Sat, 10 January 2004 20:04 Go to previous messageGo to next message
Ilia is currently offline  Ilia   Canada
Messages: 13241
Registered: January 2002
Karma: 0
Senior Member
Administrator
Core Developer
Fixed in CVS. The problem was that the message indexer ignored words shorter then 3 chars, but the search query's word size minimum was 2 chars. I've made the limit 2 chars in both instances.

FUDforum Core Developer
icon4.gif  Re: Search function -- inconsistent behaviour [message #16000 is a reply to message #15965] Sun, 11 January 2004 20:38 Go to previous messageGo to next message
wfjmueller is currently offline  wfjmueller   Germany
Messages: 95
Registered: December 2003
Location: Darmstadt, Germany
Karma: 0
Member
From this I gather that now only words more than 3 characters are indexed. Before October the limit was obviously >=2.

I'm not really happy with this move. The reason is, that the world is full with 3 letter acronyms, and it is a pitty if the search function simply ignores them. In the project I'm involved in we have
LOI MOU CDR TDR STS TRD RPC CBM DSP HDL IOB LCB LHC SIS
and this is just what I remembered in less than a minute.

So ignoring 3 letter words and thus acronyms really weakens the index when it comes to technical forums. Imho a kill list with words to be ignored in the index would be a much better and cleaner solution. This is a widely adopted strategy, very efficient to implement, and usually such a kill list is not very long. However, it is certainly language specific, so should be chosen based on the board language.
Re: Search function -- inconsistent behaviour [message #16001 is a reply to message #16000] Sun, 11 January 2004 20:41 Go to previous messageGo to next message
Ilia is currently offline  Ilia   Canada
Messages: 13241
Registered: January 2002
Karma: 0
Senior Member
Administrator
Core Developer
Like I said in my previous message, this is something that is fixed.

FUDforum Core Developer
Re: Search function -- inconsistent behaviour [message #16004 is a reply to message #16001] Sun, 11 January 2004 22:19 Go to previous messageGo to next message
wfjmueller is currently offline  wfjmueller   Germany
Messages: 95
Registered: December 2003
Location: Darmstadt, Germany
Karma: 0
Member
Ooops, my mistake, I misread your message, stating that now again all words with more than two characters will be indexed consistently again, so no problem with 3-letter agencies and the like Smile .
Re: Search function -- inconsistent behaviour [message #16005 is a reply to message #16001] Sun, 11 January 2004 22:22 Go to previous messageGo to next message
wfjmueller is currently offline  wfjmueller   Germany
Messages: 95
Registered: December 2003
Location: Darmstadt, Germany
Karma: 0
Member
To avoid a lot of 'the', 'has', 'had', 'was', and 'and' bloating the index by using a kill list is on the long run probably still an option to be considered, but that should be in 'suggestions'.
Re: Search function -- inconsistent behaviour [message #16006 is a reply to message #16005] Sun, 11 January 2004 22:24 Go to previous message
Ilia is currently offline  Ilia   Canada
Messages: 13241
Registered: January 2002
Karma: 0
Senior Member
Administrator
Core Developer
I do not want to implement kill lists, as they over complicate the logic and the advantages are very minimal.

FUDforum Core Developer
  Switch to threaded view of this topic Create a new topic Submit Reply
Previous Topic: Spell checker and comma
Next Topic: Anchor text mangled in auto-created [url] tags
Goto Forum:
  

-=] Back to Top [=-
[ Syndicate this forum (XML) ] [ RSS ]

Current Time: Fri Nov 22 01:00:51 GMT 2024

Total time taken to generate the page: 0.02513 seconds