Mailing List Archive

Doug Cutting wrote:

>> - Added a few more words to the stop word list (MS'
>>contribution via Alan).
>>
>
>I don't think we should do that here. This could break any application
>which is already using this stop list when it upgrades Lucene, since it will
>no longer be possible to search for these words. What we need is a facility
>to load stop lists from file-based resources, and to include a new such
>resource that contains this MS stop list. But I don't think we should
>change the default stop lists. What do others think? Is that too
>conservative?
>
Sounds good. The real issue is making it easy to use custom stop lists,
not making Lucene compatible with today's flavor of some other search
engine. Besides, isn't there already a constructor that takes an array
of stop words? File reading stuff would be good, but I don't think we
have to have it before the release. Is that too conservative? :)

Dmitry



--
To unsubscribe, e-mail: <mailto:lucene-dev-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:lucene-dev-help@jakarta.apache.org>

Hello,

> > - Added a few more words to the stop word list (MS'
> > contribution via Alan).
>
> I don't think we should do that here. This could break any
> application
> which is already using this stop list when it upgrades Lucene, since
> it will
> no longer be possible to search for these words.

Ah, didn't think of that. I'll revert to the previous revision.

> What we need is a facility
> to load stop lists from file-based resources, and to include a new
> such
> resource that contains this MS stop list. But I don't think we
> should
> change the default stop lists. What do others think? Is that too
> conservative?

Couldn't applications already use this:

/** Builds an analyzer with the given stop words. */
public StandardAnalyzer(String[] stopWords) {
stopTable = StopFilter.makeStopTable(stopWords);
}

And we leave it up to them to figure out how they get the list of stop
words to this constructor.

> > - Re-indented the whole class.
>
> I don't think we should re-indent whole files. It makes it hard to
> figure
> out what's changed over time. We should try to use a similar
> indenting
> style, but if someone has written the code, they have the right to
> indent
> it. I generally only re-indent code if I'm committing it for the
> first
> time, or for those parts of the code that I change.

I knew I was walking on thin ice when I did this. :)
I was thinking more along the lines of making all code in the project
uniform, which may be different from the original.
I prefer the former, but either one is fine with me, as long as it's
clear what we're sticking to. I think the former would not be hard to
achieve since we have a fairly small number of developers with commit
priviledges.

Otis


__________________________________________________
Do You Yahoo!?
Yahoo! Sports - Coverage of the 2002 Olympic Games
http://sports.yahoo.com

--
To unsubscribe, e-mail: <mailto:lucene-dev-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:lucene-dev-help@jakarta.apache.org>