[tei-council] description of <constraint>

[James Cummings]

Sebastian Rahtz wrote:
> things work more or less with the new <constraint> and <constraintList>
> elements inside elementSpec and schemaSpec. I have not yet implemented
> their use in <attDef>, will do that gradually.

I thought there had been general agreement that constraintGrp was a 
better name than constraintList?

> However, the issue I'd like a decision on is about a description.
> there are four choices:
>   a) do nothing yet. group <constraint>s by implementation language 
> inside <constraintList>s,
>       that's it
>   b) embed constraint description in the language itself, by whatever 
> means come naturally.
>   c) add <desc>/<gloss>/<equiv> set to <constraint> content model
>   d) add <desc>/<gloss>/<equiv> set to <constraintList> content model
> (or e) propose a more complex documentation structure)

I think I'd prefer c) myself.  How does schematron (for example) do b)? 
    Obviously I think people should be allowed to do b) if they like, 
but what if they are using a really badly-designed constraint language 
that has no method for commenting or somesuch. (None occur to me, mind 
you...)  Although I want the ability for *some* documentation, I don't 
really want anything more complex than a <desc> or <gloss>.

-James

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