[tei-council] Conformance .... the continuing saga

[James Cummings]

Sebastian Rahtz wrote:
> when someone writes the first non-Consortium ODD processor,
> oh frabjous day.

I'm not holding my breath.

> So my critical question is whether the form of the
> generated schema and DTD fragments is normative
> or merely indicative.  Can I break the format
> at P5.1?

I think you can, they are an output and just as we are changing the 
organisation of the guidelines (which will change again if we add a 
chapter in P5.1) then other outputs can change.  We're committed to 
avoiding breaking backwards compatibility in the recommendations the 
guidelines provide (etc.), but if the form of the generated schemas 
change from one version to the next, I think only Wohann would complain.

> In case anyone is wondering, it really is not that obvious
> what to model in the RELAX NG schemas. CUrrently
> I say
> foo = foo.content, foo.attributes
> 
> because that lets you say
> 
> foo.attributes = empty

Which seems like a good thing to me.


> but I could equally say
> 
> foo = foo.content, attribute bar ( text)
> 
> and miss out the intermediate stage.

But loses you the indirection and just doesn't seem to be a good thing. 
  Does not creating this intermediate stage save lots of work?  What is 
gained by it?

-James