[tei-council] rowing back on <mixedContent>

[Martin Holmes]

On 14-07-04 12:18 PM, Sebastian Rahtz wrote:
>
> On 4 Jul 2014, at 20:12, Martin Holmes <mholmes at uvic.ca> wrote:
>>> sure. and they can shoot themselves in the foot now, using Roma, and
>>> make an invalid DTD. It is not hard at all.
>>
>> But that's not a good situation. We should have a plan to get away from it.
>
> er, yes. I don’t have one to hand, I am afraid.

Well I was suggesting one: we could deprecate DTDs and remove support 
from them at the same time that we transition to Pure ODD. Then we 
wouldn't have to worry about DTDs when thinking about how Pure ODD 
should work.

>> My point is that we shouldn't be designing Pure ODD (which is what
>> you're doing) on the basis that it has to take account of the
>> limitations of DTDs.  If Pure ODD can be simpler and easier to
>> understand if we don't think about DTDs, then why not use this as an
>> opportunity to move beyond them?
> sure. you’re agreeing with me, then, that we should eschew <mixedContent>
> and go straight to Pure ODD in all its raging glory and not worry about
> whether you can DTDize it

Definitely. Isn't Pure ODD supposed to enable us to look beyond the 
limitations of current schema-languages anyway?

>>> sure. no change there. and its very very hard to guarantee --
>>
>> But much easier with DTDs out of the picture. Just sayin.
>>
> not much but slightly, yes.
>
> writing an deterministic content model using TEI’s very lax
> content is far from easy.

That's a more important question, and something we should be thinking 
about. Should Pure ODD attempt to constrain or warn about 
non-deterministic content models? Although the issue is embodied in 
Relax vs XML Schema, it's not inherent to specific schema languages; any 
language or schema processor has to decide whether it will tolerate 
non-deterministic content models or not. Since it's such a core issue 
for TEI, it's worth some careful thought. Pure ODD might for instance 
allow a NDCM if you add @allowNonDeterministicModel="true", and if 
that's not there, an ODD processor would attempt to detect and warn, 
even if all you're creating is RelaxNG.

I'm sure Lou's given this a lot of thought already.

>
> … P6 … drop all schemas.. drone….allow everything everywhere…. express constraints
> in schematron… drone....

Not a P6 issue at all, really. I don't see why we shouldn't drop DTDs in P5.

Cheers,
Martin

> --
> 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.
>