Re: Advanced notice of proposal to be made on our weekly call to move us forward

Arnaud,

+1.

I'd also like to suggest how we layer the semantics documents.
1. The compact syntax semantics should be given by translation to
verbose RDF SHACL syntax.
2. The verbose RDF SHACL semantics should be given by translation to
SPARQL (to the maximum extent possible), plus some outer control logic
for non-SPARQL aspects, e.g. for recursion.

-- Arthur

On Tue, May 26, 2015 at 7:45 PM, Arnaud Le Hors <lehors@us.ibm.com> wrote:
> At the virtual Face to Face meeting last week I made a proposal that was
> rejected by Jose who had missed the discussion leading to it and that was
> generally deemed too important to resolve immediately. After discussion we
> agreed that I would take it to the next teleconference, with advanced
> notice, giving everyone a chance to prepare for it. So, here is for the
> advanced notification:
>
> My proposal will be as follows:
>
> PROPOSED: Adopt Holger's draft as the basis for the SHACL specification,
> leveraging Peter's proposal to improve it, and define a compact syntax for
> it along the lines of ShEX.
>
> As explained in my previous email, I expect that defining the compact syntax
> on top of the SHACL spec will lead to identification of specific gaps. I
> expect us to discuss what to do about each gap on a case by case basis:
> e.g., drop the feature or extend the base/SHACL (possibly beyond what SPARQL
> alone can do, which is consistent with our resolution about defining the
> semantics with SPARQL as much as possible).
>
> Several WG members have expressed interest in also having an abstract syntax
> in various formats. It is reasonable for them to pursue doing so, because
> people may find them useful in various contexts, as long as it is understood
> that the SHACL specification provides the only governing semantics. In other
> words: consistent with resolution of ISSUE-31 (see
> http://www.w3.org/2015/05/20-shapes-minutes.html#resolution03 ), in case of
> discrepency, the SHACL spec is the definitive reference.
>
> Note that references to "the SHACL spec" are not meant to imply that this
> has to be in the form of a single document. The WG remains free to decide
> how many documents to produce.
>
> If this proposal is accepted, Holger's draft will become a Working Draft
> that is under the governance of the WG. As such it will be subject to
> changes in parts or in all, pending WG's resolutions.
>
> Let me know if you have any questions.
> Regards.
> --
> Arnaud  Le Hors - Senior Technical Staff Member, Open Web Technologies - IBM
> Software Group

Received on Thursday, 28 May 2015 11:41:45 UTC