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

[Lou Burnard]

On 04/10/14 23:19, Sebastian Rahtz wrote:
> On 4 Oct 2014, at 18:07, Lou Burnard <lou.burnard at retired.ox.ac.uk> wrote:
>> 1. we decided that all datatype declarations must be indirect, so <textNode>  has to have a content model of macro.something, for symmetry.
>>
>> 2. macro.something could be either macro.xText , maps to (text|model.gLike)* or (as yet nonexistent) macro.Text  which maps to text, or indeed something else we haven't thought of yet
>>
>> If someone thinks the TEI has decided the wrong way (for a given application) redefining the macro is not so hard.
> Skating on thin ice here.  <textNode> has no "content model".

Sorry, yes, you are right. It probably is redundant then.
> What it translates into in a given schema language
> is determined by the ODD processor, not by the ODD designer.  They do _not_ have control over whether
> textNode is rng:text, rng:data, xsd:token, or macro.xText.

If it's a macro, they can redefine it, of course.