Re: Prefix algorithm and other tweaks


Esa Ahola (esa@cyclone.mindspring.com)
Wed, 17 Dec 1997 14:22:43 -0500 (EST)


On Thu, 11 Dec 1997, Esa Ahola wrote:

> Haven't heard back from you; that's quite okay, just wanted to make sure
> mail was not getting lost in one direction or another.

If you have sent me mail, I have received none of it. :-\ But I see that
you (rodan.contigo.com?) paid a short visit to my test page.

I have now upgraded my "production" site

   http://mercedes.mindspring.com/mercedes/archives/mq.html

to support prefix matching. It uses

   exact:1 endings:0.8 synonyms:0.7 prefix:0.5

I have also enhanced the prefix feature to support either explicit
or implicit prefix matching:

- If you specify a prefix match character in the configuration
  file, only words that end in the match character are processed.

- Otherwise, all words that are at least minimum_prefix_length
  characters long are automatically prefix matched.

What do you think?

I ended up introducing the following configuration variables and
their defaults:

    minimum_prefix_length: 1
    prefix_match_character: "*"
    max_prefix_matches 1000

I'd welcome some review and comment of the diffs, and a fellow in
the Netherlands has asked for a copy.

-- 
Esa Ahola
esa@cyclone.mindspring.com



This archive was generated by hypermail 2.0b3 on Sat Jan 02 1999 - 16:25:25 PST