[relax-ng] [Fwd: Ann: Xml Validation Interoperability Framework (xvif)]

[Eric van der Vlist <vdv@dyomedea.com>]

I'd like to insist here on the level of integration of xvif within Relax
NG.

All the examples in this strawman have been tested against a Relax NG
implementation following the derivative algorithm kindly described by
James and xvif is just about associating a transformation "T" with a
pattern P so that for each node "n" the derivation is applied on T(n)
instead of being applied on "n".

Relax NG has proven to be the perfect technolgy to host such a framework
and I am convinced that it can bring almost unlimited new possibilities
to Relax NG

Eric

PS: What about a compact syntax for xvif?
-- 
See you in San Diego.
                               http://conferences.oreillynet.com/os2002/
------------------------------------------------------------------------
Eric van der Vlist       http://xmlfr.org            http://dyomedea.com
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------

--- Begin Message ---

• From: Eric van der Vlist <vdv@dyomedea.com>
• To: Xml-Dev <xml-dev@lists.xml.org>
• Date: 20 Jun 2002 11:11:37 +0200

Hi,

I am happy to announce a first version of Xml Validation
Interoperability Framework (xvif), a proposal for embedding pipes of
transformations and validations within grammar based schema languages.

While the common approach is to define pipes of transformations and
validations operating on complete XML documents, my purpose with xvif is
to explore the possibility of defining "micro-pipes" transforming
information items (ie attributes, text and element nodes) during the
transformation itself.

The main advantage in doing so is to bring the transformation closer to
the place where it is needed. This makes it more easy to maintain, but
also allows it to interoperate more closely with the schema language:
with Relax NG and xvif I can say for instance that an element is valid
if it gives a result 1 matching a pattern1 after a transformation 1 *or*
if it gives a result 2 matching a pattern2 after a transformation 2.

In this proposal, I have given the main focus to defining the
interoperability framework itself, relying on existing technologies for
the transformations and validations defined in my micro-pipes (XPath,
XSLT, Regular Fragmentations and regular expressions are supported in
the current implementation).

Some of these pipes are rapidly getting quite complex which IMO
shouldn't be seen as a problem of the framework itself but rather as an
indication that it might be needed to define a lightweight
transformation language specific to such applications.

This prototype is based on a partial implementation of Relax NG written
in Python (I have focussed on the features which I have found most
important for my proof of concept) using the derivative algorithm
described by James Clark.

The following resources are available in this first version:

- Strawman: http://downloads.xmlschemata.org/python/xvif/xvif.html
- On line demo: http://downloads.xmlschemata.org/python/xvif/tryMe.cgi
- Download: http://downloads.xmlschemata.org/python/xvif/

I assume that discussion about xvif can be carried on xml-dev, but I
have also setup a more specific mailing list for XML schema related
discussions at xmlschemata@xmlschemata.org:

- http://lists.xmlschemata.org/xmlschemata/

Enjoy!

Eric
-- 
See you in San Diego.
                               http://conferences.oreillynet.com/os2002/
------------------------------------------------------------------------
Eric van der Vlist       http://xmlfr.org            http://dyomedea.com
(W3C) XML Schema ISBN:0-596-00252-1 http://oreilly.com/catalog/xmlschema
------------------------------------------------------------------------

--- End Message ---