NOT SML but XML-for-data and business

[Didier PH Martin]

Hi Shaw,

Shaw said:
1) how are the (real) XML developers approaching XML Schema? Will it be
transparent to applications, and just constrain the data as per DTD's or
will the meta-data be available - if so, in what form?

Didier reply:
Actually we have a problem. There is no recommendations about how to link a
document/fragment with its schema and nor any recommendation on about the
validation rules for documents aggregating fragment (each fragment having
its own schema). I an anxious to see what the XML schema group will present
at XML 99 as an answer to these question or to these unfulfilled needs.

Shaw said:
2) given the fragmentation of DTD definitions in any given market (mine
particularly with FpML, FIXML, FinML, BizTalk etc.) is anyone addressing
general tools for mapping between these - and what are the issues that are
being tackled?

Didier reply:
We are using XSLT to translation from one to the other. We are currently
working on a meta model that could be mapped to these different languages.
However, this is a lot of work even for a specific domain (the finance
domain). But if we reach our goal. It will be easier to translate from this
meta model (or meta language which is XML based) to any other particular XML
based language with XSLT.


Shaw said:
3) are there any discussions going on about how to map between different
formats programatically? I can imagine RDF being used, but then there needs
to be agreement on meta-meta-data(!) - and I can't imagine (see 2 above)
people agreeing on anything much when it's a core business differentiator
(read tie-in/revenue opportunity) to have a proprietary format.

Didier reply:
We tried RDF but RDF is a completely different data model and we discovered
that to create a meta data model in RDF is simply not practical. We
discovered that it is a lot easier if the meta domain language is simply an
element hierarchy. Not necessarily easier for machines but definitively
easier for humans. And believe me, to reduce error is quite important. The
problem is that errors will introduced by humans. If the system is too
complex, we have errors. So we re-discovered what Ben Schederman discovered
several years ago. Thus, from the software psychology point of view, we
discovered that using RDF for translation is error prone and that using a
master schema translated into other schemas is less error-prone. However,
this experience is limited to our group and it would be interesting to see
what others got as result. Of course, this is possible only if they keep
track of their process and have in place a learning mechanism to fine tune
these processes.

Cheers
Didier PH Martin
mailto:martind at netfolder.com
http://www.netfolder.com


xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/ and on CD-ROM/ISBN 981-02-3594-1
To unsubscribe, mailto:majordomo at ic.ac.uk the following message;
unsubscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)