Re: Reminder of the workshop and its outcome

On 7/23/14, 12:28 AM, Arnaud Le Hors wrote:
> Hi all,
>
> I've been on vacation and haven't had a chance to follow the 
> discussion in details but I'm a bit surprised by the way the 
> discussion is going. We seem to have lost sight of the fact that the 
> proposed charter was drafted based on the outcome of the workshop that 
> took place last year and provided for a direction to follow.
>
> With that in mind, on behalf of the chairs of the workshop, I'd like 
> to remind everyone that:
>
> TopQuadrant was invited to present at the workshop.
Yes and we should have attended. I guess we assumed that it was clear 
that SPIN would be high on the agenda and that the workshop participants 
would be sufficiently educated about how SPIN works.

Having said this it is IMHO not appropriate for a Consortium about the 
World Wide Web to assume that all key decisions are made in a single 
face-to-face meeting. Of course follow-up discussions on the internet 
should be taken very seriously, because many more people can contribute. 
One of the outcomes of the workshop was the need for human-friendly 
syntaxes, and based on that requirements it is necessary to present SPIN 
differently.

>
> During the workshop, Guoqian Jiang (Mayo) presented SPIN (with SPARQL 
> queries, as it was in the Member Submission):
> http://www.w3.org/2001/sw/wiki/images/d/d6/RDF-Validation-Workshop-GJiang-v1.0.pdf14-17 
>
>
> The workshop participants found this good for enforcement but not 
> high-enough level for definition, like SPARQL.
It has been mentioned by others that people at the workshop apparently 
believed that the only way to define SPIN constraints was via SPARQL 
queries. However, even the W3C member submission already introduced SPIN 
templates, see

http://www.w3.org/Submission/2011/SUBM-spin-modeling-20110222/#spin-meta-modeling

"Templates are reusable "boxed" SPARQL queries that can be used in 
conjunction with properties such as spin:rule or spin:constraint, but 
also in other places."

and then it goes on illustrating a constraint template for a cardinality 
restriction. Given this (apparent) glaring oversight of the workshop I 
am afraid the whole discussion needs to be reopened. Of course SPARQL 
alone does not meet the requirements, but together with templates SPIN 
really combines the best of both worlds.

Holger