In chapter 10 of "Lucene in action" there are several case studies where
domain specific optimization is done in custom code (e.g. boosting
"new articles" or indexing certain field with a special analyzer).
I think we all have done something similar in our lucene application.
What i would love to see is a mini framework which frees us from all this
little standard things to be done (assigning an analyzer to a field for
indexing and query / boosting docs & fields at indexing time / defining
default search fields / ...).
The case study for "the server side" shows an approach where some of these
things can be done via an external config file. Unfortunately that config
is a little too specific ;)
Furthermore i think a user doesn't know about fields, boosts, phonetics or
fuzzy searches. And he shouldn't. A "search model" should do instead.
The search 'harry potter' should be translated for instance to
'title:"harry potter"^2 content:"harry potter" title:harry^2
title:potter^2' with the help of a search model.
Does anybody know of contrib classes (or a project) that do something