[Snowball-discuss] Re: wrapper references

From: J Smith (jsmith@tutorbuddy.com)
Date: Thu Mar 14 2002 - 16:54:21 GMT


In an attempt to be less secretive (which I suppose is the purpose of my
Hushmail), I'm replying to via my work account.

I'm glad to see that Snowball is getting more and more support. Personally, I
have no idea how to use Snowball itself and I wouldn't know where to begin
devising a language stemmer, but I use the stemmers themselves quite a bit at
work. I'm basically responsible for writing and maintaining a search engine
for indexed videos, and the original Porter algorithm found it's way into the
search engine at some point about six months or so ago.

I've never written any sort of search engine/IR system before, but the end
result is pretty decent. The entire thing is written in PHP (the search
engine itself) and Perl (the database/keyword indexer). Originally, I
implemented the stemming algorithm in PHP, which didn't last, as it was slow
and produced results that didn't match 100% with the stemmer available for
Perl. (If the PHP and the Perl stemmers don't match, the search engine can
produce some odd results.)

I re-wrote both the Perl stemmer and the PHP stemmer (this time in C) and
synched up their output. After a while, I took interest in Snowball and wrote
the latest stemming extension to PHP. I haven't actually used it in a
production system yet, but the next version of our search enginep probably
will, as we have a need for multilanguage searching. (Specifically, English
and French -- we are a Canadian company, after all.)

Anyways, as mentioned, the stemming extension to PHP is available at
http://209.202.82.229/software, and has recently been updated with the latest
Snowball sources, which fix the small memory leak in api.c and such. Details
are on the site.

-- 
J Smith
programmer guy
Tutorbuddy, Inc.
http://www.tutorbuddy.com

> -----Quoted Message----- > From: Martin Porter <martin_porter@softhome.net> > Sent: Thu, 14 Mar 2002 09:26:10 -0700 > To: Snowball discuss <snowball-discuss@lists.sourceforge.net> > Subject: wrapper references > > Okay, > > Teodor, Oleg/Dark Panda, I've put in a page about wrappers on the Snowball > site to cover your Perl/PHP work. > > I believe we are, at long last, fully up to date on the Snowball site. > > Martin > > >

_______________________________________________ Snowball-discuss mailing list Snowball-discuss@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/snowball-discuss



This archive was generated by hypermail 2.1.3 : Thu Sep 20 2007 - 12:02:41 BST