mifluz-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Mifluz-dev] (your mail)


From: Geoff Hutchison
Subject: Re: [Mifluz-dev] (your mail)
Date: Wed, 29 May 2002 14:46:42 -0400 (EDT)

On Wed, 29 May 2002, Yann Poupet wrote:

> > - is mifluz actively maintained at the moment?
> NO!!!

Agreed, at the moment Loic seems to be AWOL. That's not to say that it
would be hard to restart development if interested parties could be
found. The ht://Dig group is probably the most interested party. :-)

> > - is mifluz aware of handling unicode (utf-8, uftp-16)?
> no, don't think so

I'd agree--I think at the moment it can handle any 8-bit locale. I can
certainly think of a few places where 8-bit assumptions are made and these
will need to be fixed. But no one has seriously come forward as far as
tackling this. (I'd guess it's more a matter of someone who understands
coding issues with UTF-8, etc.)

> > - how does mifluz support relevance ranking?
> that's my main problem...

IMHO, this is *not* something for mifluz. Either you insert a score with
the word records, or you do the scoring from the results, like ht://Dig.

> > - does it support pattern matching or globbing in some way?
> supports suffix globbing 

I think you mean prefix globbing.

Again, look at the ht://Dig fuzzy algorithms. The easiest way is to
construct an "alternative" list and execute multiple queries.

In a B-Tree, it's easy to do a prefix match and pull out the subtree
of all matches. Any other sort of globbing would require a different
database structure *OR* multiple queries.

<http://www.htdig.org/>

--
-Geoff Hutchison
Williams Students Online
http://wso.williams.edu/




reply via email to

[Prev in Thread] Current Thread [Next in Thread]