Re: SHACL for SHACL

I have meanwhile completed a version of this SHACL-for-SHACL file for now

https://github.com/w3c/data-shapes/blob/gh-pages/shacl/shacl-shacl.ttl

I ran this against various test cases and also temporarily imported this 
into my regular test set to make sure that none of the known SHACL tests 
report syntax errors (a couple actually did because I was using 
sh:targetNode with blank nodes; fixed). It covers almost all of the 
SHACL Core syntax rules.

ACTION: Everyone please review the SHACL file above, and/or try it 
against shape definitions - we should sign this off in the Wednesday 
meeting so that we can swiftly resubmit the CR request.

Thanks
Holger


On 1/04/2017 12:51, Irene Polikoff wrote:
> Yes, if anyone has written any shapes that are syntactically invalid (intentionally or by mistake), please provide them.
>
> Of course, examples of valid shapes that don’t pass this validation would be even more interesting.
>
>> On Mar 31, 2017, at 9:50 PM, Holger Knublauch <holger@topquadrant.com> wrote:
>>
>> Hi all,
>>
>> we did have a quite successful CR transition meeting with the W3C management yesterday. I say "quite" because TimBL wanted us to do one last thing before moving to CR status:
>>
>> We are supposed to create a SHACL shapes graph to validate other SHACL shapes graphs, essentially implementing the syntax rules for SHACL Core, up to what can be reasonably expressed with SHACL itself. We had talked about this topic several time and decided not to do that due to our very short time frame. But given that W3C management is OK with us spending this additional time even beyond the March 31 deadline, we now should do our best to fill this gap.
>>
>> I have started such a file here:
>>
>> https://github.com/w3c/data-shapes/blob/gh-pages/shacl/shacl-shacl.ttl
>>
>> based on the syntax rules in Appendix B
>>
>>     http://w3c.github.io/data-shapes/shacl/#syntax-rules
>>
>> I need everyone's help to finish and check this. This is a great example of using SHACL itself, so a nice exercise for those who want to get their hands dirty.
>>
>> - The file above is completely untested. We should have some counter examples with invalid shape definitions.
>> - The file has two notable TODO gaps: the shape for rdf:Lists and for SHACL paths.
>>
>> While I may eventually get to this next week, in order to create the best possible quality for such a file in the shortest possible time frame, I hope others can volunteer to help.
>>
>> Thanks
>> Holger
>>
>>

Received on Monday, 3 April 2017 02:14:34 UTC