Re: RDF Data Shapes WG agenda for 19 May 2016

Regrets for missing today's telco

All my votes are in the proposals page, regarding the agenda issues

issue 141: I would prefer to keep the current design (sh:class /
sh:datatype), if this is not possible, I would favor 3c
the behavior of 3a is non-deterministic and depends on data on the data
graph
e.g.
ex:A ex:p "asdf"^^ex:myDT
might or might not pass validation depending on if
ex:myDT a rdfs:Literal exists in the data graph
and here we also have the inference issue for subclasses of rdfs:Literal
which complicates the constraint
e.g. one might have ex:myDT a ex:Literal in the data graph and ex:Literal
rdfs:subClassOf rdfs:Literal somewhere else

3b cannot express non-xsd datatypes which, imho is out of the question

The behavior of 3c is consistent and depends only on shape definitions
the equivalent shapes definition would be
ex:S sh:property [
  sh:predicate ex:p
  sh:type ex:myDT
  sh:nodeKind sh:Literal
]

3c needs 2 constraints (sh:type / sh:nodeKind) but so does the existing
design (sh:class / sh:datatype) we cannot simplify definitions without
loosing something

issue 133: I would favor my proposal of course
I also have a draft revision ready on a separate branch if accepted
https://github.com/w3c/data-shapes/compare/editorial-dk

Regarding the draft publication
I think the resolution of issue 133 needs to be integrated in the spec
before publication
we also need to discuss where to position the current section 3 (shapes &
validation) or how this needs to change
if sections 1-3 become consistent and flow well, the rest of the spec will
be easy to manage


On Wed, May 18, 2016 at 10:46 PM, Arnaud Le Hors <lehors@us.ibm.com> wrote:

> Here is the agenda for tomorrow. FYI, I put ISSUE-133 on it as a reference
> for a syntax related discussion although it's clear that there are several
> related issues.
> https://www.w3.org/2014/data-shapes/wiki/Meetings:Telecon2016.05.19
> --
> Arnaud  Le Hors - Senior Technical Staff Member, Open Web Technologies -
> IBM Cloud
>
>


-- 
Dimitris Kontokostas
Department of Computer Science, University of Leipzig & DBpedia Association
Projects: http://dbpedia.org, http://rdfunit.aksw.org,
http://aligned-project.eu
Homepage: http://aksw.org/DimitrisKontokostas
Research Group: AKSW/KILT http://aksw.org/Groups/KILT