W3C home > Mailing lists > Public > www-zig@w3.org > June 2000

RE: "Z39.50 Server List" DTD

From: Denis Lynch <DLynch@ACM.org>
Date: Fri, 23 Jun 2000 00:37:05 -0700
To: "Rob Bull" <bull@crxnet.com>, "Z39.50 Implementors Workshop" <www-zig@w3.org>
Message-ID: <01b001bfdce5$d41f3620$0300a8c0@kanisa.com>
Rob's description here is helpful.

As Bob says in another message, everything in Explain is there because
somebody thought they couldn't live without it. Its size and complexity has
meant that many people have decided they could live without Explain all
together.

The question of XML vs. ASN.1 really isn't very interesting. (But if you're
going to use XML, create a schema rather than a DTD. The XML-Schema
processing instruction makes the XML self-describing -- which supports
Lennie's idea of using a generic XML OID.)

What is interesting is Rob's claim that ONE-2 has identified "the _real_
things people want to know about a server." One test of this relates to
another of Bob's comments: everything in Explain is optional. It would be a
Very Good Thing to come up with a non-trivial core set of required
information. If that set doesn't coincide with Explain, it's time for a new
structure.

Even if the core required information set is already a clean subset of
Explain, it might be better to create a new structure rather than rev
Explain. Rob's points about the advantages of using mainstream technology
are very significant.

On the other hand, I don't agree with Rob's implication that an XML version
of Explain has any bearing at all on the effort required to maintain the
descriptive data. The mechanical tools help with the mechanics, not with the
information content. I'm convinced that the barrier to implementing Explain
is the information content, not the data format.

--Denis



-----Original Message-----
From:	www-zig-request@w3.org [mailto:www-zig-request@w3.org] On Behalf Of
Rob Bull
Sent:	Wednesday, June 21, 2000 12:25 PM
To:	John Robert Gardner
Cc:	Robert Waldstein; www-zig@w3.org
Subject:	Re: "Z39.50 Server List" DTD

. . .

Now, in the follow up ONE-2 project, we have re-visited explain again, and
(I guess to your dissapointment) invented "another work round" we call
explain-lite - its a piece of XML passed on Init.

However, this has been created for real reasons -
- it is considered that creating an XML approach is far more in line with
21st century technology, rather than the overhead of the explain PDUs,
- it covers the _real_ things people want to know about a server, and not
the additional baggage that you can get with explain;
- such XML can be used in conjunction with other means of disseminating
information about a server - for instance, the very same XML could be
published on a web page - this saves administration effort etc.
- the effort in creating a tool for managing data in an explain database
is quite significant, you cant assume a Z39.50 expert/programmer to be
available for maintaining such data;
- tools in the XML world are available now that can be used to parse the
validity of such explain data, these tools deal with related
issues such as character sets etc. that would probably be otherwise hand
coded.

. . .

Rob
Received on Friday, 23 June 2000 03:39:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 29 October 2009 06:12:21 GMT