Re: cyclic rules

But <Foo> has exactly one :patron, namely <Foo>, and <Foo> is of shape <FooShape>.

peter


On 07/10/2014 01:31 PM, Solbrig, Harold R. wrote:
> Don¹t think so ‹ <FooShape> says that every subject has exactly one patron
> of shape Foo.  The shape below would pass your example:
>
> prefix : <http://example.org/>
>
> <FooShape> {
>      :patron @<FooShape>+
> }
>
> As would the instance
> <X> :patron <Y>.
> <Y> :patron <X>.
>
> (Note that the the ³<Foo>² in your example is *not* related to the shape
> named <Foo>.   <X> :patron <X>,<X>. Would be an equally valid test)
>
>
> On 7/10/14, 2:11 PM, "Peter F. Patel-Schneider" <pfpschneider@gmail.com>
> wrote:
>
>> I played around with Eric's implementation, and it doesn't recognize
>> <Foo> :patron <Foo>, <Foo> .
>> as a <FooShape>.
>>
>> Isn't that a bug?
>>
>> peter
>>
>>
>> On 07/10/2014 12:53 PM, Jose Emilio Labra Gayo wrote:
>>> I have tried your example in both implementations and it passes without
>>> infinite loops.
>>>
>>> You can check it here:
>>>
>>> Eric's implementaion: http://goo.gl/kBbgCZ
>>> Shexcala implementation: http://goo.gl/rUJ8OS
>>>
>>>
>>> Best regards, Jose Labra
>>>
>>>
>>> On Thu, Jul 10, 2014 at 9:24 PM, Peter F. Patel-Schneider
>>> <pfpschneider@gmail.com <mailto:pfpschneider@gmail.com>> wrote:
>>>
>>>      Doesn't this cause an infinite loop on cyclic rules when there are
>>> cycles
>>>      in the RDF graph?
>>>
>>>      Consider, for example
>>>
>>>      schema:
>>>         <FooShape> {
>>>           :patron <FooShape>
>>>         }
>>>
>>>      data:
>>>         <Foo> :patron <Foo> .
>>>
>>>
>>>      peter
>>>
>>>
>>>
>

Received on Thursday, 10 July 2014 20:35:05 UTC