[wp-trac] Re: [WordPress Trac] #8571: Plugin search results from within WP are horribly bad

[WordPress Trac]

#8571: Plugin search results from within WP are horribly bad
---------------------------+------------------------------------------------
 Reporter:  markjaquith    |        Owner:  mdawaffe
     Type:  task           |       Status:  assigned
 Priority:  high           |    Milestone:          
Component:  WordPress.org  |      Version:          
 Severity:  major          |   Resolution:          
 Keywords:                 |  
---------------------------+------------------------------------------------
Changes (by mdawaffe):

  * owner:  matt => mdawaffe
  * status:  new => assigned

Comment:

 For the curious, the main reason it's so bad is because of the way the
 information is stored.  The readme.txt file is parsed, broken into
 sections ( installation, description, faq, ...) and then mooshed into one
 big string that's stored as the post content of a bbPress post.  In this
 string, the sections are delineated by html comments... which have very
 generic words in them.

 Try searching for "title".  You will get almost every plugin back since
 the string "title" appears in each of those html comments.

 That, and bbPress search has always been pretty poor.

 I'm working on addressing this now by setting up a
 [http://www.sphinxsearch.com/ sphinx] index.  This first pass will give
 much better results.  And look for continued improvements over the coming
 months.

-- 
Ticket URL: <http://trac.wordpress.org/ticket/8571#comment:2>
WordPress Trac <http://trac.wordpress.org/>
WordPress blogging software