On 06/02/15 17:42, Sebastian Rahtz wrote:
>> On 6 Feb 2015, at 16:33, Martin Holmes <[log in to unmask]> wrote:
>> Thanks indeed for this, Louis. I had no idea this would work; I always think of attribute specification as being entirely separate from content specification, and resort to Schematron for this sort of thing.
>> Now the obvious question is: is this possible in PureODD, and if not, should it be, and how?
> it’s an interesting question. TEI has always kept a distinction between attributes and element content,
> so the PureODD stuff hasn’t gone back to examine attributes enough.
> I’d be interested to hear if anyone else wants this form of power in their attributes. Personally, I think
> its angels dancing on pinheads :-}
I've used this a lot for modelling corpus texts and their metadata, but
not within the TEI, just straight in RNG. Anyway, my message is: it's
done and it can be useful, but I'm afraid that my use cases may fall out
of the "90% of users, 90% of the time", or whatever the slogan is.