Re: Proposal c)

On 5/6/2015 19:55, Eric Prud'hommeaux wrote:
> I grouped them and struck Jose's (I think he's fine with that but I'll 
> check). now: [[ three documents: primer, core semantics, SPARQL 
> functions SHACL (Shapes Constraint Language) Primer ericP and Holger A 
> Primer for SHACL, concentrating on the high-level language, but also 
> giving examples of constraints directly using SPARQL syntax. Largely 
> compatible with all the other proposals. Core SHACL Semantics ericP 
> SPARQL/SHACL Semantics ericP ]]

Thanks for updating the home page. I believe Primers are usually 
non-normative and therefore cannot be part of the formal deliverables. 
Furthermore, the Primer is orthogonal and could be added to the other 
proposals too.

Regarding the SPARQL/SHACL draft, it is a step in the right direction to 
acknowledge that the SPARQL dimension to SHACL also exists. You have 
basically just copied an outdated snapshot of my chapters 7 onwards into 
a new document, inserted a misleading introduction ("SPIN"?) and put 
your name on it. The result is unfortunately a completely inconsistent 
and incomplete set of unrelated documents. Where did the content of 
chapters 1 - 6 end up? Does your Core Semantics document even have the 
same semantics? How does the RDF vocabulary of the core look like and 
where is its Turtle file? Where did the consistent reading experience 
for users go to? As expected, this splitting of documents is only 
leading to confusion. If we absolutely have to make such a split, we 
should do this once the language has stabilized. Meanwhile I have added 
code that injects Part 1 and Part 2 sub-headers into the TOC:

     http://w3c.github.io/data-shapes/shacl/#h-toc

Having said this, I am always trying to explore possible ways forward 
that everyone could live with. I observe that my draft could be renamed to

             "SHACL RDF Vocabulary and SPARQL-based Semantics"

This is self-contained, covers all requirements and doesn't need 
anything else. The core language elements are defined via templates and 
their SPARQL queries. IMHO elegant, consistent and easy to implement. 
With its many examples it could even serve as a Primer for now.

If enough people believe there is value in the WG publishing the 
alternative semantics document then your draft could be renamed to

             "SHACL Core Abstract Syntax and Declarative (?) Semantics"

and it should be made clear that this is about a small sub-set of SHACL 
only, similar to Part 1 of my draft. (FWIW I still believe something 
like "SHACL Lite" would be a better name.) We would then have two 
semantics documents for the core language. As long as they both define 
the same meaning, I see no problems with that. I personally do not see 
value in the Abstract Syntax document - it neither helps implementers 
nor is it easy to read by users nor is the language properly extensible 
- but I accept that other people may find this interesting from an 
academic point of view, so I would not vote against its publication as 
long as it doesn't mess up the consistency of the other spec.

On the general question of authorship, I guess we can reshuffle editors 
at any time - I am certainly open to adding other people as editors to 
"my" draft.

Cheers,
Holger

Received on Thursday, 7 May 2015 01:21:15 UTC