Xapian discuss - Apr 2005 - multilanguage support

Hi.

I am considering runing omega with xapian backend as a basis
of the search engine tailored for my specific needs.

After some research I found that omega is stemming text
before adding it to xapian database.

I was thinking to write a small code to detect web site language
and according to it use special stemming algo for that text.

My question is the search query itself.
Search queries are too small to detect language, so I can not
decide what stemming algo to use.

Adding the possibility to choose the language while entering
the search string is not appropriate in my case.

So, I whant to ask if the are any workaround for this problem ?

On Sat, Apr 09, 2005 at 03:56:22PM -0400, links@bannershift.com
wrote:
> After some research I found that omega is stemming text
> before adding it to xapian database.
This can be disabled fairly easily.  You need to tweak the source
currently (stemming language "none" instead of "english") -
I've
not got around to making this a config file option yet.
> Adding the possibility to choose the language while entering
> the search string is not appropriate in my case.
> 
> So, I whant to ask if the are any workaround for this problem ?
Really, I think you need to design your UI so that the language for a
query does get specified (either explicitly or implicitly), or not stem.

You can try to be clever - see the latter half of this message (and the
replies):

http://thread.gmane.org/gmane.comp.search.xapian.devel/447

But I've seen that tried and I'm not convinced it's the right
approach.

Cheers,
    Olly