- From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
- Date: Thu, 18 Jun 2015 10:26:45 -0700
- To: Arnaud Le Hors <lehors@us.ibm.com>, public-data-shapes-wg@w3.org
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 Some initial comments on http://w3c.github.io/data-shapes/shacl-ref/ What is the intent of this document? Is it supposed to provide the full syntax and semantics of SHACL? (Can SHACL actually express the syntax of SHACL?) There are lots of words and phrases that need to be fixed up or defined. Just in 1.1.1 there are "applies", "expected", "have this shape", "relevant", "tested nodes", "fulfill", "arbitrary", "matching", and "resources". There are as well "abstract", "superclass", "evaluated", "default", "constraint violation", "library", and "properties" of constraint violations", in other places. This is probably not an exhaustive list. The SPARQL definitions require that the shape graph is accessible to the SPARQL engine performing shape validation. The QCR constraint only accepts shapes and there is no analogue for classes. Treating blank nodes as special for typing is a bad idea. Treating xsd:string as special is also a bad idea. The structural constraints on sh:PropertyConstraint are far too constraining. Inverse property constraints should not require a separate definition. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAEBCAAGBQJVgv9RAAoJECjN6+QThfjzrLoH/iDBvj3ymCLx6bUHgpWQVtMJ 3p0jPyS2hURmem8Det5ED4XB6922tCUSAZ2GuZgHxsDSWPK4qm6ku/R6k1181qgx s++6MG2UmsNgy4E763C5368lb22mBxVQL/BYFb8+GgKEBZW7GYlz7S5wMC8BzD5B RHKM/AwiCPrP5NlOIerALild6mxMoFz8JOGG7qCsYLS25Fu8/abRi324NQS4xbqD xMJnth5T0C549vmaDBAN/hqMSyOSkYp5mR3Q6pjnyJFNSVHOy7389V0r+ebmpcCm cbXj/3KuQvo6gIV8Y5od6cdbObZoICT2QRRWNb+K1RmAYChQHikGvIeynozEIqk= =uz8I -----END PGP SIGNATURE-----
Received on Thursday, 18 June 2015 17:27:17 UTC