Re: mentions of recursive shapes

On 07/09/2015 08:03 PM, Holger Knublauch wrote:
> On 7/10/2015 9:04, Peter F. Patel-Schneider wrote:
>> Here are some mentions of recursive shapes, most from the WG mailing list,
>> but a few from other places.
>>
>
>>
>> Private email to Jose:
>>    I don't know how a multi-valued approach would work.  It appears to me that
>>    recursion over negation splits the choice space.
>>    Consider, for example,
>>      sh = not p sh
>>      sha = or ( q sh ) ( r sh )
>>    on
>>      a q x .
>>      a r y .
>>      x p y .
>>      y p x .
>>    I think that a should belong to sha, but I don't see how to achieve this
>>    without considering multiple models.
>>    peter
> 
> I have turned the above example into
> 
> https://github.com/w3c/data-shapes/blob/ISSUE-62/data-shapes-test-suite/tests/features/core/recursive-005.ttl
> 
> 
> I assume the test is to verify that ex:a has shape ex:sha?

Correct, but also not verifying that ex:x has shape ex:sh nor that ex:y has
shape ex:sh.

peter

Received on Friday, 10 July 2015 05:43:09 UTC