Re: a path towards implementing the way forward

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

The semantics in [2] is not adequate to handle sh:hasShape, as far as I can
tell, as the semantics only covers positive recursion and sh:hasShape
permits negation in recursive loops.

peter

On 06/11/2015 02:56 PM, Arthur Ryman wrote:
> Peter,
> 
> I added a link to the wiki [1]. The document is at [2].
> 
> [1]
> https://www.w3.org/2014/data-shapes/wiki/Main_Page#Other_Input_Documents 
> [2] http://arxiv.org/abs/1505.04972
> 
> -- Arthur
> 
> On Thu, Jun 11, 2015 at 2:12 PM, Peter F. Patel-Schneider 
> <pfpschneider@gmail.com> wrote: Could you send the pointer to your
> well-founded semantics again?
> 
> Putting a link on the WG Wiki page would also be helpful.
> 
> peter
> 
> 
> On 06/11/2015 11:01 AM, Arthur Ryman wrote:
>>>> On Wed, Jun 10, 2015 at 7:52 PM, Peter F. Patel-Schneider 
>>>> <pfpschneider@gmail.com> wrote:
>>>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256
>>>>> 
>>>>> Which limited forms of recursion do you mean?
>>>>> 
>>>>> Where are these limited forms useful (and better than other 
>>>>> approaches)?
>>>> 
>>>> Peter,
>>>> 
>>>> I mean that there is no need to require that sh:valueShape is
>>>> acyclic. This is form of recursion is useful (based on OSLC use
>>>> cases) and has a well-founded semantics which I spelled out
>>>> recently. However, this means that you cannot translate the shape
>>>> into a single SPARQL query, but that is not a requirement.
>>>> 
>>>> -- Arthur
>>>> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBCAAGBQJVeg+YAAoJECjN6+QThfjz1zYH/1+pQ6NJRllHSd2Uv8KyGGFt
y6TXglu2ol1qB1a8FR0sjybcTLkHTm8MRguzidZpJWvPkGUlbB2MsUHQISc7etVC
hAmjBNl0a21YIthgtoxv2kKVKZQj7NeJdpd/Qz/8Rxn58ZtL9QqyVwrCtXdx7HxJ
Kntl3tpflP9kMWA95uXZcSpPj575nOhcWf5kcqRPtL+/IjYsVtFDs8qpVeWsfSTL
0aXvqkfqRFz7MmnlLwpi3cocOfQ321Q8Kha/LiErG+pljMUCRwAMLiwCIkH+KUfa
rNSxBOQ0alWlzulBK6i+dpbu6wOo0dkX4sLv9ue3Vl//ODhbWzBTU8Mzaodbh7A=
=ZWs0
-----END PGP SIGNATURE-----

Received on Thursday, 11 June 2015 22:46:15 UTC