[tei-council] surfaces, surfaceGrps, etc. [was : minutes/release deadline]

[Piotr Bański]

On 12/11/11 19:14, Sebastian Rahtz wrote:
> ooh, I love it. For a start, I don't think <surfaceGrp> solves your fundament problem, 
> which arises from nested <surface>. You _could_  add an attribute "relative=true|false",
> but that would mean making urx etc optional. Hmm.  How about simply saying that
> if any of urx, ury, llx, llr are missing, they are inherited from up the chain of <surface>?

Maybe one thing to consider is whether all you may want is cases where
you either inherit the entire bunch of 4 numbers, or not, and then you
would use something that functions somewhat like @decls, pointing at
some parent, mutually exclusive (by schematron) with the four attributes?

Considering four vs. nothing may be cleaner than trying to cope with
inheriting single attributes. Or safer to begin with, at least. Or it
may exhaust 90% of use cases.

  P.