[tei-council] where to put schematron constraints

[James Cummings]

Sebastian Rahtz wrote:
> James Cummings wrote:
>>
>> Would the ability to stick the schematron inside <content> still be 
>> kept? Or would it only be allowed in the new place?  Will it break any 
>> existing user-created ODDs?
> there are two sorts of breakage:
> 
> a) the XSL scriptts would no longer do what they expect
> b) the ODD would no longer be valid against p5odds.odd
> 
> one would have to treat these separately.

Ok, but equally the cost of them upgrading to putting the schematron in 
the better place (which seems to make more semantic sense rather than 
being inside 'content') is a fairly trivial task.  (CutnPaste or a quick 
XSLT to move it from one place to the other?)

I'd suggest you do a quick show of hands on TEI-L to see if there are 
many people using schematron inside their ODDs.  If there are, then 
maybe allow it both places, if not, then just do the change but write a 
quick example conversion transformation and also some examples of using 
co-occurrence constraints in the new way.  I can see this as something 
that potentially a lot of people will want to do, but probably haven't 
done yet.


> since this is not core TEI, we _can_ break if we have to

I'm not sure those are precisely the rules we have set for 
ourselves...but I certainly can't imagine this affecting more than a 
small minority of people all with only a very few ODDs at most.

-James

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