[dsdl-discuss] Re: Concurrent (overlapping) structures

From: "Eric van der Vlist" <vdv@dyomedea.com>

> I'd say that, more generally, any technique such as JITTSs
> (http://sbl-site2.org/Extreme2002/) which can build several XML infosets
> through transformation of a source document should be easy to integrate
> using our Validation Management (ex framework).
> Should we add this as a use case?

First, let me apologize if I was a bit huffy in my previous post. I was frustrated by a bug at the
time.

When I said that concurrent markup was a niche requirement, it was not meant to be dismissive
but to emphasize why a modular approach like DSDL is so important: there are so many niches
or specialist needs. A convenient framework makes it convenient to create specialist validation
languages. Non-nesting structures is a very good example of this.

ISO is in a good position for this: the W3C effort is based on "how can we get the broadest possible
single schema language" which precludes specialist requirements. ISO, on the other hand, can
get agreement for fairly small specialist languages: in particular, we can focus on the requirements
of people who work with text such as publishers of various kinds.

So a more polite response than last time would be:

1) You can validate almost any constraint with Schematron (especially, Schematron + RELAX NG
+ some kind of datatyping) including non-nesting structure constraints. So DSDL starts off *way*
ahead of DTDs or XML Schemas. However, Schematron schemas may be ugly and algorithmic:
good for validation and for human messages but not reusable in the way that a more domain-specific
language would be.

2) Because DSDL provides a framework, systems based on transforming the document in various
ways before validating it are very possible: in fact, that you can view two parts of DSDL as doing
that already (i.e. Part 4 Validation Candidate Selection Language and Part 8 (?) Architectural
Forms(?)). So that is more systematic than XML Schema's too-primitive <appinfo> elements.

3) If you have a specialist need that you think is general, then the best thing is to get some
running code to fit into, in particular, RELAX NG, such as James suggests, and then joining
the ISO effort to get it introduced into ISO DSDL. To say "DSDL falls short because it
doesn't provide X" really translates to "someone else should get X standardized": but DSDL
languages (and ISO SC34 standards in general) are typically user/developer driven and need
a champion: *you* should get it developed and standardized!

Cheers
Rick Jelliffe

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)