[tei-council] issues on Schematron and deprecation

[Syd Bauman]

> >   * classSpec
> 
> without itemising every member of the class, I can’t see how you
> _can_ get this right. is that what you plan to do?

Exactly. 

For the attributes I special-cased classes containing 'global', so as
not to have a test for 500+ elements. As for model classes, I think
model.common is probably the biggest at 113 members. That doesn't
seem so bad.


> I can see that it could be desirable to add a constraint to one of
> the data.* macros, checking on the date in some way. for example,
> would you use this to check that the date was always the first
> Monday of the month?

Yes, perfectly reasonable thing to want to do. All I'm saying is you'd
have to figure out the darn context on your own!


> this seems to me the sort of thing that makes James Clark look
> gloomy and say he wishes namespace prefixes to the devil. I’d be
> quite surprised if you could ever really sort this out
> automatically.

Me too. I think it is probably pretty easy to generate an error
message, but pretty hard to know what to do. To be honest, I really
don't have a problem with telling ODD customization writers they
shouldn't do this. (Heck, I don't mind telling *everybody* they
shouldn't do this. :-)