Re: question about http://w3c.github.io/data-shapes/semantics/

Le mar. 19 mai 2015 22:26:36 CEST, Peter F. Patel-Schneider a écrit :
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
>
> I don't understand why Well defined schemata are defined so closely.
>
>
> Definition [Well defined schema]
>
> A shape expression schema S is called well defined if for all negated shape
> label !T in negshapes(S), the corresponding dependency sub-graph
> dep-subgraph(T, S) is a directed acyclic graph.
>
>
> This excludes any schema where a negated shape references a
> recursively-defined shape, even if that shape does not involve negation in
> any way.

Indeed.


> Is this required?
It is a sufficient condition for soundness, maybe not necessary.


> Would it be possible to instead just say that in a
> well-defined schema there can be no loops involving negation?

It might be possible, however (for now) I do not know how to prove it.
Also, I do not know how the validation algorithm would go, whereas I do 
have a validation algorithm for my proposal.

--
Iovka Boneva
Associate professor (MdC) Université de Lille
http://www.cristal.univ-lille.fr/~boneva/
+33 6 95 75 70 25

Received on Wednesday, 20 May 2015 07:25:32 UTC