Ah. If the support is already present in the indices, then it
should not be all that difficult to add support for near/within/
before/after operators to the parser, which will only require an
efficient means to iterate through the term occurrences in a file
based on proximity constraints.
Do you have a pointer to the data structures used to encode term
position information in the 3.2 indicies?
-bryan
-----Original Message-----
From: Geoff Hutchison [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, March 28, 2001 12:05 PM
To: Bryan Thompson
Cc: [EMAIL PROTECTED]
Subject: Re: [htdig-dev] Query: htDig and proximity queries.
On Wed, 28 Mar 2001, Bryan Thompson wrote:
> I realize that htDig does not currently index work position information
> so there is no basis for computing the answers to such queries.
It does in 3.2. The only thing lacking at the moment is the search side of
things.
> The client is interested in contracting someone to build this feature
> into htDig -- if the price is reasonable. Rather than license a
commercial
> engine, they would like to leverage the existing work on htDig and use
this
> as an opportunity to incorporate a new feature (and one that has been on
the
> wish list for a long time!).
I think the bigger question is one of timescale. At the moment, I think
most of the developers are quite busy. I will not be able to do much
development for about a month due to some new projects.
Granted, I suspect Quim could enhance his query parser to deal with a
"near" operator, but I can't speak for him.
--
-Geoff Hutchison
Williams Students Online
http://wso.williams.edu/
_______________________________________________
htdig-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/htdig-dev
