Login | Register
My pages Projects Community openCollabNet

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

RE: [Catacomb] Search Engine



See below:

> - When can we generate index ?
There are 2 cases here:
a) A search engine DOES support incremental indexing. 
   Then the answer is simple: During each PUT and PROPSET.
b) A search engine DOES NOT support incremental indexing. 
   Then we can do the following: to have 2 indexes. Let's call first one
temporary index
   and another one - permanent index. So, during each PUT and PROPSET we
regenerate temporary index.
   Once per day (or more often if needed) we regenerate the permanent index
(and then the temporary one
   is cleared). Any query that comes in will be evaluated against both of
those indexes.

> - Where can we put the index?
I personally prefer DB, but filesystem will do as well.

> - The index support multi languages?
Ideally - yes, but it could be next step.

> - Should we use stemming or dictionary?
In my opinion, stemming is better option.

> I think it is really cool, if we could support the embedded search engine.
So do I. There many real life scenarios where Catacomb could be used
especially if it had really good 
search support. Otherwise what people will end up with is intercepting
SEARCH calls and talk to other search engine.
This is actually what I was thinking of before originating the email :-)


-----Original Message-----
From: Sung Kim [mailto:hunkim@cse.ucsc.edu]
Sent: woensdag 5 maart 2003 20:53
To: Yuriy Pasichnyk
Cc: 'catacomb@webdav.org'
Subject: RE: [Catacomb] Search Engine


On Wed, 5 Mar 2003, Yuriy Pasichnyk wrote:

> The candidates for the embedded search engine could be:
> 1. http://www.htdig.org/
> 2. http://swish-e.org/
> 3. http://mnogosearch.org/
>
> What do you mean with "many issues with indexing" ?

- When can we generate index?
- Where can we put the index?
- The index support multi languages?
- Should we use stemming or dictionary?
- etc.

I think it is really cool, if we could support the embedded search engine.

> -----Original Message-----
> From: Sung Kim [mailto:hunkim@cse.ucsc.edu]
> Sent: woensdag 5 maart 2003 19:15
> To: Yuriy Pasichnyk
> Cc: 'catacomb@webdav.org'
> Subject: Re: [Catacomb] Search Engine
>
>
> On Wed, 5 Mar 2003, Yuriy Pasichnyk wrote:
>
> > Dear Sirs,
> >
> > I am working on a project that will use search functionality of Catacomb
> > (DASL) to full extent.
> > But, after looking into details of how Catcomb implemets it I have a
> > question:
> > 	Wouldn't be better to allow plugging-in real search engines instead
> > of using full-text search capabilities of MySQL ?
>
> I agree, but which search engine could we use? Also there are many issues
in
> indexing.
>
> > Thanks,
> >
> > Yuriy
> >
> > P.S.
> > The TODO lists among other things the following:
> >
> > - 409 conflict when we get directory
> > - Set right content-type when we PUT
> >
> > Wasn't it already implemented ?
>
> Yes. This is implemented in CVS development version. The version is not
> released yet.
>
> --
> Sung Kim
>