[tei-council] rendition, rend, and style

[James Cummings]

John A. Walsh wrote:
> On May 16, 2007, at 8:53 AM, Sebastian Rahtz wrote:
>> John A. Walsh wrote:
>>> The problem is with a fixed value list is that you can't combine 
>>> rendition styles.
>> but  the datatype of rend is "multiple tokens"[1], so you can say 
>> rend="a b c" and each of a, b and
>> c will be checked against the list. honest.
> Thanks for the clarification, Sebastian. 
 > The value list approach provides better validation but still
> does not provide the benefit of linking to explicit and formally defined 
> styles in the <rendition> element in the header, but I'm probably 
> beating a horse (whether it's a dead horse or not remains to be seen :-).

I still agree that there is a benefit of being able to point back to a 
<rendition> in the header to document it more fully than one does with an 
attribute value.  Yes, I like the validation that a closed value list provides, 
and I think I'd seriously recommend that approach for most uses, but being able 
to point to a place in the header which has even a prose description is in many 
ways much more powerful.  Because doing rend="bigWobbly" where I define that by 
a paragraph in my ODD necessitates the ODD to accompany the instance (not a bad 
thing!), but point to a <rendition> element in the header where I have a 
paragraph describing it means my document instance stands on its own in terms of 
documenting the meaning of the values it contains.

Is there an easy clear way to have something be multiple tokens and/or multiple 
pointers? Or is that just getting icky?

-James
-- 
Dr James Cummings, Oxford Text Archive, University of Oxford
James dot Cummings at oucs dot ox dot ac dot uk