[tei-council] [Fwd: Re: the new <constraint> element]

[James Cummings]

 From Syd

-------- Original Message --------
Subject: Re: [tei-council] the new <constraint> element
Date: Thu, 9 Apr 2009 11:20:39 -0400
From: Syd Bauman <Syd_Bauman at Brown.edu>

James -- please post to Council.

> An alternative might be to allow <desc> inside <constraintGrp> and
> describe any of the individual <constraint>s contained in that
> group at that location. However that goes against the spirit of
> <desc>, which explicitly mentions its parent element, doesn't it?
> 
> Other solutions?

Stick to functional (or logical) grouping, as opposed to
by-schema-language grouping, and then the <desc> *does* describe the
constraints (and thus the <constraint>s) as a child of
<constraintGrp>.  :-)


BTW, the other way we need to think of how these things get chopped
up is by natural language. Most projects will want their
constraint-based validation messages in a single natural language,
but the TEI doesn't have that luxury. We're going to want to provide
messages in multiple languages. Thus

   <constraintGrp ident="A">
     <desc xml:lang="en"/>
     <desc xml:lang="fr"/>
     <desc xml:lang="zh-TW"/>
     <desc xml:lang="es"/>
     <constraint scheme="isoschematron" xml:lang="en"/>
     <constraint scheme="isoschematron" xml:lang="zh-TW"/>
     <constraint scheme="isoschematron" xml:lang="es"/>
     <constraint scheme="schematron" xml:lang="en"/>
     <constraint scheme="schematron" xml:lang="zh-TW"/>
     <constraint scheme="schematron" xml:lang="es"/>
   </constraintGrp>


-- 
Dr James Cummings, Research Technologies Service, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk