W3C home > Mailing lists > Public > www-webdav-dasl@w3.org > January to March 1999

Re: DASL / GILS alignment

From: William Fisher <fisher@hollywood.engr.sgi.com>
Date: Wed, 27 Jan 1999 21:15:11 -0800 (PST)
Message-Id: <199901280515.VAA02259@hollywood.engr.sgi.com>
To: www-webdav-dasl@w3.org
Cc: fisher@hollywood.engr.sgi.com (William Fisher), babich@filenet.com
> Today several of us met to discuss alignment between the GILS and DASL
> specifications.
> Global Information Locator Service (GILS) info: http://www.gils.net/
> DASL info: http://www.ics.uci.edu/pub/ietf/dasl/
> Federal Geographic Data Committee (FGDC):
> http://www.fgdc.gov/clearinghouse/clearinghouse.html
> Attending were:
> Alex Hopmann, DASL WG Chair
> Jim Davis, DASL author
> Eliot Christian, Chief GILS Architect
> Archie Warnock, FGDC Clearinghouse and ASF Implementation
> At the meeting, we all agreed that there is enough synergy between the two
> specifications that it should be possible to align them. Archie is rewriting
> the existing web -> Z39.50 gateway, and will attempt to support receiving
> DASL requests. If this experiment is successful it should make available
> thousands of databases and literally terabytes of information to DASL clients
> After a day and a half of intensively working through the design, we did not
> come up with any changes we would have to make to DASL. We are actually
> really happy about this as a test of our design so far for DASL. It's
> possible that as the experiment progresses we might find things we want to
> improve in DASL, and these will be offered as suggestions on the DASL
> mailing list.
> Participation in this experiment from other folks is strongly encouraged-
> Please just post to this list (www-webdav-dasl@w3.org) if you want to get
> involved.
> Alex Hopmann, Jim Davis, Eliot Christian, Archie Warnock
	The one area that I believe DASL protocol should support is
	the ability to handle legacy search engines. The two most
	notable one's are DIALOG and Lexus-Nexus.

	Originally these were command line oriented on main-frame machines
	but now they have some simple WEB user interfaces which can be used
	to access a subset of the capabilities. Both of these are stateful
	search engines and have a very rich set of search operators
	for doing full-text searches on the huge number of data bases
	licensed by DIALOG. They have more than 850 on-line data bases
	including anything from Dunn-BradStreet corporate financial
	information to lots of refereed technical journals used by researchers
	and lawyers.

	I would like to see the "mapping" of an existing search engine
	command set into the DASL protocol as an example that it is
	powerful and sufficient to provide a solution to these "legacy"
	full text database search companies.

	At the IETF working group meeting in Orlando, Florida in December 1998,
	lots of energy was spent on discussing the "requirements" but it was
	still an open question of the "scope" of the protocol. Lots of folks
	talked only in terms of the "server" and forgot that IF the protocol
	didn't allow or have sufficient flexibility to support this then
	the "server" couldn't add anything to overcome this short-fall.

	One basic requirement which it needed to support DIALOG searching
	using this protocol is proximity operators and some method of
	storing state on the server for a particular "session".
	I volunteered to write up a strawman example which would show how
	a relatively complex DIALOG search query could be done in the
	proposed DASL protocol.

	Others mentioned the Z39.50 protocol and the huge library base
	which utilizes that protocol.

	I recently re-read the "draft-reddy-dasl-protocol-04.txt"
	by Reddy, Babich et.al. and the search grammar is NOT sufficient
	to handle this case. I talked with Alan Babich, since we used to work
	together about this proposal to gain more background information.

	I also mentioned this IETF work group to a couple of folks at DIALOG,
	Oracle and Thompson-Thompson and asked them to participate in
	the development of a good search protocol. I don't know if they
	are on the mailing list.

	So my question is: Are you willing to consider extending this
	protocol to cover legacy full-text searching?

	If so I will work on the strawman example of how to do DIALOG
	searching using some extensions to the '04' version of the protocol
	or will wait until this proposal is updated to reflect the
	current thinking.

-- Bill (fisher@sgi.com)
Received on Thursday, 28 January 1999 00:17:30 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:22:41 UTC