[tei-council] <content> vs <mixedContent>

[Sebastian Rahtz]

On 4 Oct 2014, at 11:34, Lou Burnard <lou.burnard at retired.ox.ac.uk> wrote:

>> The semantics of <textNode> == data/@type=‘string’, no?
> 
> rng:text or rng:string ? presumably the latter.  

eh? rng:string doesn’t exist, does it?

> Calling our macro "data.text" (even though it has the value rng:string) is going to return to bite us several times, I foresee. For example, the macro macro.xtext is defined as "text | model.glike" not "string | model.gLike".
> 
> How do you feel about giving <textNode> macro.xText as content? It's more consistent with the TEI semantic model!

Yes, I was going to raise this point. But then how do people say they really _do_ want just text? 

The more I think about this in real life, the more I worry about the biting. Using <mixedContent> 
at least has the benefit that we avoid saying what the “mixed” consists of, but <textNode>
exposes us to the “text” question which seems to haunt us like Europe haunts the Conservative Party.

 * if <textNode> == rng:text, people have to be taught to add <macroRef key=“model.gLike”/> as well
 * if <textNode> == data.text, then people can’t specify just the equivalent of <rng:text>

since we’re saying that <mixedContent> is for newbies, and <content> is for experts, the experts
might as just use the full power of data.*; why bother with the shortcut element for the
(actually very unusual) case of #PCDATA (we use that 2 or 3 times in the TEI)? note also
the lack of symmetry with <datatype> content.

as I said, promoting use of <textNode> seems like we’re teaching people to make
old-skool XML DTDs where that was the only datatype available.
--
Sebastian Rahtz      
Director (Research) of Academic IT
University of Oxford IT Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431

Não sou nada.
Nunca serei nada.
Não posso querer ser nada.
À parte isso, tenho em mim todos os sonhos do mundo.