Re: RDF Data Shapes WG agenda for 4 October 2016

Sorry, I must have caught a summer flu but feel sick and will likely not 
participate today.

On the agenda items:

For ISSUE-22 if we make this change then we'll need to very precisely 
define what it means for a shape to be recursive. Doing this test 
statically may be hard, and is basically impossible for SPARQL queries 
involving sh:hasShape (since the ?shape may be a dynamically computed 
parameter). So I believe detecting recursion at runtime is the easiest 
option and this would need to be reflected in the wording before we can 
consider approving this change.

On ISSUE-131 I have deleted the shapesGraph argument from sh:hasShape, 
and believe this has significantly improved the situation because there 
is no longer a need for passing named graph references around, nor a 
dependency on the concept of data sets.

On ISSUE-177 I would be unhappy if we are building extra dependencies 
between the documents. Making the AS normative will only expose us to 
even more critical feedback on the mailing lists. We are at the risk of 
running out of time already. If we lack precision in the definitions, 
then these need to be lifted into the spec. I am also not sure about the 
general role of the AS document. If we want more examples, then these 
could also just as well be added to the main spec - why do they go into 
the AS? Overall this could make the AS much shorter than it currently 
is, and focused on the Abstract *Syntax*, not some kind of competing 
shadow specification of the semantics and terminology.

Regards,
Holger


On 4/10/2016 3:07, Arnaud Le Hors wrote:
> Now available:
> https://www.w3.org/2014/data-shapes/wiki/Meetings:Telecon2016.10.04
> --
> Arnaud  Le Hors - Senior Technical Staff Member, Open Web Technologies 
> - IBM Cloud
>