W3C home > Mailing lists > Public > www-webdav-dasl@w3.org > April to June 2004

DASL for 'web searching' ?

From: Mario Salzer <mario@erphesfurt.de>
Date: Tue, 29 Jun 2004 11:28:29 +0200
To: www-webdav-dasl@w3.org
Message-ID: <40E1363D.nailTP11SCUD@yoco.erphesfurt.de>

Hi there,

I've lately been wondering about some way to merge different (or
internally divorced) web sites by unifying or negotiating on a
site search standard interface. That's what my current idea was:
http://erfurtwiki.sourceforge.net/OpenSearch

However, as I didn't found it myself, someone only short time after
pointed me to DASL (seriously you should simply promote your effort
as "HTTP SEARCH extension", then it was easier to find). Of course
having the search mechanism built into the transportation protocol
makes a lot of sense, and I clearly see the advantages (at least if
your server is already used as WebDAV file repository), but still
I'm wondering if the application of DASL isn't pretty limited to
just that.
Or does anybody think, that DASL could be used some way or
another for implementing a (hopefully lightweight) cross web site
searching feature (sort of mini-Google over just a few sites)?  Has
that ever been a project goal?

When reading over the latest draft, I feel that your <basicsearch/>
grammar is already rather complicated (complex boolean search
conditions like with that ugly old Z39.50 standard I've been pointed
to by someone else), yet the <orderby/> and <score/> ranking is not
useful to merge result lists from multiple SEARCHed sources. So my
overall impression is, that DASL is meant to scan trough a list
of files of that are just made available via HTTP+WebDAV as access
protocol; and that the very exact (exclusive, SQL like) matching
of search conditions also looks rather file oriented. (So if there
was a WebDAV file manager I'd probably rather copied down all files
to my local hard disk and could let ordinary file tools scan through
it...)
This matches with my impressions of Apache mod_dav merely being
used to allow remote file accesses. I'm sure any more advanced CMS
easily could provide its services through standardized and powerful
interfaces like WebDAV and DASL. But then it is probably not
applicable to the majority of web site background software in use
nowadays? I knew where I'd like to have WebDAV inside and where it
really made sense, but that's more a five-years-from-now story.


No, this post is not meant to annoy anyone, I'm just investigating
if I should carry on with that meta-search-API or better stick to
something existing and standardized. Complexity aside, my only
problem is, that DASL seems not applicable in all points. Sure
another grammar could do it (btw, is there already something else
than <basicsearch/>?), but then I could simply convert OpenSearch
from RPC into <opensearch/> XML... ;)

mario


-- 
Get your free warez from ftp://127.0.0.1/
Received on Tuesday, 29 June 2004 05:29:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 22 March 2009 03:38:10 GMT