Re: ISSUE-68: Updated definition

On 11/03/2016 14:00, Peter F. Patel-Schneider wrote:
> Is there a shape-in-shape construct in SPIN that needs pre-binding?   That
> would presumably give excellent guidance on how to do shape-in-shape in SHACL.

No, SPIN doesn't have a concept of sh:valueShape, so there was no need 
for this. Functions and magic properties may be calling other functions 
and magic properties though.

>
> Looking at
> https://www.w3.org/Submission/2011/SUBM-spin-modeling-20110222/#spin-rules-thisUnbound
> it appears that SPIN is doing something different from what is needed for
> sh:hasShape, however.
>
> It appears that magic properties are where pre-binding is most needed.  This
> is even outside a FILTER construct, so there is a need to allow multiple
> values for the same variable.

Yes, but these multiple values are never passed in at once, but magic 
properties iterate over all input bindings, just like BGPs do.

>
> Given that pre-binding is in use in SPIN, is there a definition from SPIN that
> can be used in SHACL?

SPIN never defined this, just vaguely hinted at "whatever Jena does".

Holger


>
> peter
>
>
> On 03/10/2016 05:40 PM, Holger Knublauch wrote:
>> Yes, pre-binding has always been used in SPIN.
>>
>> Holger
>>
>>
>> On 11/03/2016 3:38, Peter F. Patel-Schneider wrote:
>>> Is something like pre-binding needed in SPIN?
>>>
>>> peter
>>>
>>> On 03/08/2016 10:06 PM, Holger Knublauch wrote:
>>>> I had updated the definition of pre-binding but forgot to send an email to the
>>>> list:
>>>>
>>>> http://w3c.github.io/data-shapes/shacl/#sparql-constraints-prebound
>>>>
>>>> I believe we need implementer's feedback to see whether that definition is
>>>> clear and precise enough, but that applies to everything in the current draft,
>>>> so I hope we can live with that definition for now and close the ticket.
>>>>
>>>> HTH
>>>> Holger
>>>>
>>>>

Received on Friday, 11 March 2016 04:18:45 UTC