Re: http://w3c.github.io/data-shapes/semantics with negated recursion

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

I wrote the syntax by hand, so maybe it is not completely correct.

I can see that the closed might be ambiguous.

Aah, the 'and' should be a ',', which is (roughly) conjunction for shape
expressions.

peter


On 06/24/2015 01:39 PM, Arthur Ryman wrote:
> Peter,
> 
> I am having trouble parsing your example against the abstract syntax 
> given in [1]. Perhaps you have abbreviated it or is the grammar wrong?
> 
> Your example is: S = { s closed ex:r !s [1;1] and ^ex:r iri [1,1] } Here
> I assume S is the schema, s is a ShapeLabel, and we are dealing with a
> ClosedShape. Therefore the following must be a ShapeExpr: ex:r !s [1;1]
> and ^ex:r iri [1,1]
> 
> However, the 'and' operator only appears in ConjShapeConstraint where it
> combines other ShapeLabel's. How does one form conjunctions of 
> constraints?
> 
> [1] http://w3c.github.io/data-shapes/semantics/
> 
> On Wed, Jun 24, 2015 at 1:22 PM, Peter F. Patel-Schneider 
> <pfpschneider@gmail.com> wrote: The behaviour of
> http://w3c.github.io/data-shapes/semantics on simple negated recursive
> shapes is quite bizarre.
> 
> Consider S = { s closed ex:r !s [1;1] and ^ex:r iri [1,1] } on graphs of
> the form Gn = { ex:i1 ex:r ex:i2 . ex:i2 ex:r ex:i3 . ... . ex:in ex:r
> ex:i1 . }
> 
> The valid typings for these graphs appear quite random and also quite 
> computationally difficult to determine.
> 
> peter
>> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJVi2X4AAoJECjN6+QThfjzRpoH/2Ak3NUrV/0bqRfjjGYAlKUS
rTuaNzj8CmLikZtXMNtTiAjSMAxR3rcpIQNiySEOdqxqElP5m/HM1Xzn10JT0P9A
8vZeJhWZ7ZSj2aWGq7xg35eIvY63ez8OkpguVVf8j6DWkg9vOYVm/AJG3UU/JRMd
lshMLf8M8rJcwJN/69mlWO/Kv0uDhRpvEdYkVOr8CF/oNqThpJmhbDKkbg/eBrwt
CCB6dkXcrK+CChd/1Ovryiw4wPC5comm0X9NUEGSi+Fk2HdeZtya9wRFbwNJ6Uxt
mU22s3iJzmUIl8xxOGhpqiht+tiP3yZcTYFaWwwDJnwoXfT3AhPnkpWSGcYwebY=
=cHSH
-----END PGP SIGNATURE-----

Received on Thursday, 25 June 2015 02:23:32 UTC