RelaxNG et al. [was: DocBook Technical Committee Meeting Minutes: 17 September 2008]

["Grosso, Paul" <pgrosso@ptc.com>]

> -----Original Message-----
> From: Bob Stayton [mailto:bobs@sagehill.net] 
> Sent: Wednesday, 2008 September 17 13:35
> To: DocBook Technical Committee
> Cc: docbook@lists.oasis-open.org
> Subject: [docbook] DocBook Technical Committee Meeting 
> Minutes: 17 September 2008
> 
> DocBook Technical Committee Meeting Minute: 17 September 2008
> =============================================================

> There was discussion about the lack of support in tools
> for RelaxNG with Schematron rules, despite them both
> being international standards.  Since DocBook 5's
> official OASIS standard will use such a schema, there was
> concern that this will hinder deployment.
> 
> Some workarounds were suggested:
>  - using the DTD and adhering to authoring style rules.
>  - validating documents outside the editor.
>  - using XML Schema and ACL rules in Arbortext.
> 
> One question that came up was what features would be lost
> if RelaxNG were supported but not Schematron rules.
> 
> ACTION: Jirka to analyze RelaxNG without Schematron.

I'd be interested to know what constraints can only
be validated by Schematron, and what constraints can 
be validated by RelaxNG but not XSD.

If someone needs to use the DocBook 5.0 XSD to validate 
a document, what validation will they be giving up 
(unless they use other validation techniques such as
ACL or other scripting)?  And if a tool supported
RelaxNG, how much validation (due to Schematron) would
they still be giving up and have to do via other methods?

I note that DocBook 4.0 allowed both a fileref and an
entityref attribute on graphic elements as well as 
allowing content, and the prose said "element content 
should be used in favor of either EntityRef or FileRef 
and EntityRef should be used in favor of FileRef".  I 
wonder, if this was acceptable, how unacceptable is it 
not to be able to validate all DocBook 5.0 constraints 
via a DTD or XSD.  [But this last bit isn't part of
Jirka's action--just my editorial musing.]

paul