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

[Lou Burnard]

On 03/10/14 17:03, Sebastian Rahtz wrote:
> On 3 Oct 2014, at 16:36, Lou Burnard <lou.burnard at retired.ox.ac.uk> wrote:
>> We invented this element precisely because it has a different content model from <content> -- mixed content is always exactly the same as an alternation of a bunch of elements with text. So it would be nonsensical to permit it to contain sequence, and redundant to permit alternate.
> Clearly I have been in some parallel universe for the last 6 months

You also seem unwilling to accept council decisions agreed to twice 
(both times in your presence). Is this just your evil twin speaking?

>
> if mixedContent is solely,  in your definition
>
>      <content>
>         <rng:oneOrMore>
>      	    <rng:ref name="elementRef"/>
>          </rng:oneOrMore>
>       </content>

Yes, that is what I suggested back in June as the content model for 
mixedContent. Please look at the tagdoc for it which I added at that time.

> then it is incomplete, as it must contain classRef too.

Yes, I agree that the content model should in fact be

<content>
        <rng:oneOrMore>
<rng:choice>
<rng:ref name="classRef"/>
     	    <rng:ref name="elementRef"/>
   </rng:choice>
       </rng:oneOrMore>

  </content>

probably with some other constraints  (e.g. @expand with any value other 
than "alternate" doesnt make sense here)
>   But leaving that aside,
> this is shorthand for:
>
>    <alternate maxOccurs=“unbounded” minOccurs=“0”>
> 	<textNode/>
> 	   … refs…
>    </alternate>
>
> and, in the cold light of dawn etc, it seems to me an over complication
>

You're entitled to your view. But (I repeat) Council has now agreed 
twice that this would be a useful shortcut. Of course you're free to 
refuse to implement it, if you want to be really bolshy.,