Re: shapes-ISSUE-29 (formalism): Formalism for definition of high-level language [SHACL Spec]

I would like to follow up on this ISSUE before yet another week passes. 
Let's try to create a proposal that stands a good chance of being 
acceptable to everyone.

(On this I wonder whether we can do these polls outside of meetings too, 
e.g. on wiki pages, because the meetings seem to be largely occupied by 
administrative topics and then only a small number of vocal people drive 
the discussions.)

I think if we remove the work "partly" from Peter's proposal below we 
may have enough to move forward. Peter stated that the ISSUE text itself 
mentions two SPARQL-based alternatives:

- SPARQL, and dealing with nested high-level expressions by building up 
a query string to be evaluated at once
- SPARQL, and dealing with nested high-level expressions by evaluting 
each part individually and combining the results outside of SPARQL

The only difference here appears in the handling of nested expressions 
(sh:shape formerly aka sh:valueShape). I don't think this difference is 
worth delaying a resolution of this proposal. While I would prefer the 
sh:hasShape solution (pure SPARQL) for the spec, I regard this as an 
implementation detail and could live with an alternative approach if it 
has the same expressivity (i.e. it would need to be able to handle 
arbitrary templates, not just the core ones). Even if the spec uses 
sh:hasShape, many implementation would indeed combine all nested queries 
into a single large query, to improve performance.

Holger


On 4/13/2015 6:21, Peter F. Patel-Schneider wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> That suggests that there should be an easy partial resolution of this issue
> along the lines of:
>
> PROPOSED:  Partly resolve ISSUE-29 stating that the formalism for the
> definition of the high-level language of SHACL will be SPARQL, perhaps with
> some combination of results that will take place outside of SPARQL, based on
> the previous resolution
> http://www.w3.org/2015/02/18-shapes-minutes.html#resolution02
> RESOLUTION: Define semantics using SPARQL as much as possible
>
> peter
>
>
>
> On 04/10/2015 06:49 AM, Karen Coyle wrote:
>> Is this not what we voted on at the f2f?
>>
>> http://www.w3.org/2015/02/18-shapes-minutes.html#resolution02 RESOLUTION:
>> Define semantics using SPARQL as much as possible
>>
>> kc
>>
>> On 3/28/15 1:20 PM, RDF Data Shapes Working Group Issue Tracker wrote:
>>> shapes-ISSUE-29 (formalism): Formalism for definition of high-level
>>> language [SHACL Spec]
>>>
>>> http://www.w3.org/2014/data-shapes/track/issues/29
>>>
>>> Raised by: Richard Cyganiak On product: SHACL Spec
>>>
>>> A formalism must be picked for defining each of the constructs of the
>>> high-level language, as prose is considered insufficient.
>>>
>>> Proposals include:
>>>
>>> - An abstract syntax plus prose - An axiomatic semantics - SPARQL, and
>>> dealing with nested high-level expressions by building up a query
>>> string to be evaluated at once - SPARQL, and dealing with nested
>>> high-level expressions by evaluting each part individually and
>>> combining the results outside of SPARQL - …
>>>
>>>
>>>
>>>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v2
>
> iQEcBAEBAgAGBQJVKtPLAAoJECjN6+QThfjzAIMIAKsTzKq1PwcpSacPPaaJkdim
> ux7BfK0xCSVKjPJEYmnwbSJyFUUZcCcEhuU9R9XscGEFd+YUIXQlFMfkYhF4KBZF
> PzTX14VqXjLSrHJN5t34b6BKKfP3F4RjV7y/GhQDpbQH4QADD1275so7NDEfpDfP
> QBgpfNaAL1wVEIsLEJmRUoG75klkeTl2lQMxloE8nekfxj1a17OuVdm3wyyQRFO9
> +m5F4lT3nF5QwWILEnUn7x/eUc/81QzUK0rozQgXu94DOhr82YaIUcVNQXeuV9PL
> RCktUxXmVwxybFQrdlAXsMy9fwVRe5v3k22i2EnCeXh/BvR9bOIv1c0D6GLVZUw=
> =6++/
> -----END PGP SIGNATURE-----
>