[tei-council] constraint again : content

[Sebastian Rahtz]

Lou Burnard wrote:
> Who are these people
us :-}
> and why are they redefining the content model of <content>? 
to say that it contains RELAX NG elements.
> More exactly, why would they want to redefine it one way when it is 
> used to express a constraint on the content model of an element, and 
> another when it is used to express a constrainty on err... the content 
> of an element?
cos you can be precise about the elements allowed and required.
>
> I think that one of the many the nice things about our new 
> <constraint> element is precisely that it is a generalization of the 
> rather strange pre-existing <content> element; I regard the latter as 
> a degenerate form, or very specific specialisation, of the former: in 
> a subsequent release maybe we should insist on wrapping it round with 
> a <constraint scheme="relaxng" type="model"> ; in this version maybe 
> we can  insist that those who do wish to redefine <content> should do 
> so: anyone for <constraint scheme="dtd" type="model"> (NASS)?
I have some sympathy with this. It's just a pain that it would rule out 
schema-based syntactic checking
of the content model, and would mean relying on NVDL. Maybe no bad thing.
>
> And, for what it's worth, my personal verbosity-vs-clarity meter comes 
> down firnmly on the side of requiring <content< within <constraint>
>
and on that I'd concur. this is not stuff you write by hand every day.

-- 
Sebastian Rahtz      
Information Manager, Oxford University Computing Services
13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431