re: LC-36 (your comments on XML Schema last-call draft)

• From: David RR Webber <Gnosis_@compuserve.com>
• Date: Thu, 21 Sep 2000 20:25:00 -0400
• To: "C. M. Sperberg-McQueen" <cmsmcq@acm.org>
• Cc: www-xml-schema-comments@w3.org
• Message-ID: <200009212025_MC2-B441-CCC1@compuserve.com>

Message text written by "C. M. Sperberg-McQueen"
>The adoption of a multi-tier design does have a number of advantages;
the two-tier design of the current XML Schema spec, which
distinguishes resolutely between the abstract component level and the
transfer-syntax level, seems to us to be the best available compromise
between simplicity and flexibility.  We may be wrong, of course, but
on the whole the WG felt that moving to three tiers would add
complexity faster than it would add flexibility.
<<<<<<<<<<<<<<
Once again this has been largely overtaken by events.
First the GUIDE initiative is seeking to show how to implement
a multi-tier architecture using XML as part of the ebXML work
 - see https://www.xmlguide.org
And this constrasts from the W3C Schema work that is programmer
centric (audience is Java and C++ programmers) by being 
business and data analyst centric.
The GUIDE approach is based around the ebXML work and
it is significantly simpler to use - something the W3C schema was not
able to achieve - specifically because it is using the ebXML
business transport, registry/repository and Business Process
work to amelorate the delivery and deployment issues.
This should result in a happy marriage between a simple 
and effective use of base XML syntax, while allowing 
particular users to use the extended features in W3C Schema
when appropriate for their specific business need.
The biggest issue therefore remains consistent behaviours
across implementations.  Approaches such as GUIDE that rely
on a narrow sub-set of XML Schema will have significantly
better chance of guaranteeing interoperability compared to
approaches that demand the full complexities of XML Schema.
Ultimately of course the market will decide on value for money
here!
What may make sense is a harmonization project as a follow-on,
and part of the W3C Protocol WG - to select a subset of Schema
for a lightweight basic business processor that can support ebXML.
Thanks, DW.

Received on Thursday, 21 September 2000 20:25:51 UTC