[tei-council] Fwd: style lang definition

[Kevin Hawkins]

On 10/6/12 11:58 AM, Lou Burnard wrote:
> Nevertheless, I would like to  add some
> prose to the effect that using more than one lang in a given document is
> Not A Good Idea, since it makes the combination of rendition information
> (already possible factored across three different places) needlessly
> complicated.

I agree that we should have a statement to this effect in the 
Guidelines.  But we should also be precise, as James keeps reminding us. 
  Specifically, we should say that you should use only one (or as a few 
as possible) syntaxes for describing rendering in the source document so 
that processing the TEI document, such as for rendering of the digital 
text, is not needlessly complicated.

> And is there any support for my suggestion that in the absence of a
> <styleDefDecl> we assume @style values (and <rendition> contents) are in
> CSS?

I'm not sure that we want to say that if @rendition is used and yet the 
<rendition> element is empty (or entirely missing), you should assume 
that CSS is used.  Might this cause backwards compatibility for people 
with previosly encoded P5 documents?  If this isn't a concern and we 
decide to do this, I think we want scheme="css" to be the default value 
(which it currently is not).

As for assuming values of @style use CSS if there's no <styleDefDecl>, I 
think we need to specify which version of CSS to assume.  But I'm not 
sure which version to suggest.

--Kevin