Re: Looking at the current proposals for SHACL

On 3/20/2015 21:47, Peter F. Patel-Schneider wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Thinking further on
>
> An alternative design would be to hard-code the sh:valueShape property into
> the engine, and do the recursion in the outer layer, outside of SPARQL.
>
> I don't see how it would work.

I believe I had explained this before, although these things can easily 
get lost in the many emails. It would basically be a new bullet item in

http://w3c.github.io/data-shapes/shacl/#operation-validateConstraint

that would hard-code the sh:valueShape template by recursively calling 
validateNodeAgainstShape for each value of the property. This solution 
does have the same limitations as the approach based on sh:hasShape.

Holger


>
>
> peter
>
>
> On 03/19/2015 09:46 PM, Holger Knublauch wrote:
>>
>> On 3/20/15 1:30 PM, Peter F. Patel-Schneider wrote:
>>> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
>>>
>>> So you are saying that you can either do recursive shapes inside
>>> SPARQL, which requires an extension to SPARQL, or outside of SPARQL,
>>> which requires something more than SPARQL.  OK, I'll change the wording
>>> to "It provides a formal specification for SHACL in terms of SPARQL
>>> plus something extra to handle recursive shapes."
>> Yes the critical difference is that option 2 could still be executed
>> with off-the-shelf SPARQL processors.
>>
>> I would also be fine with the same solution that you have, and simply
>> disallow recursion at sh:valueShape.
>>
>> Thanks, Holger
>>
>>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1
>
> iQEcBAEBAgAGBQJVDAjHAAoJECjN6+QThfjzj3gIAJOs8jHDN1NruSNt85meW4Mz
> ubqO0IojW0jHhrf2ATKKLkOR+KnMGQdx5PoP38UCWVFFkITQu2niMMOKl6LMtaoO
> zur7k1yyrdrAE5HnS1dBipIyN6WdAzMwvcaPCGozN36u1GZRpK9smmQx6rFZBOSu
> 0a81FZc3qUovPlNxkdRKDo+h/Dl/3D90E85hKgqSbnNLna0WYtY44iffZAp2boWj
> TrOr8HGXj6MwZMSJfdo9AqnzpHO7Kpkw952O3QJ1qwylDDYC2zRX/JrJ2l1nlb+g
> FpqXhnQYonFJdtdweVaPUpI2HP5cHgSSb+x4+Dh55tU7SaPSIdqPOT6+PJPm2ps=
> =VZ1w
> -----END PGP SIGNATURE-----

Received on Sunday, 22 March 2015 23:00:38 UTC