- From: Peter F. Patel-Schneider <pfpschneider@gmail.com>
- Date: Mon, 27 Apr 2015 16:50:33 -0700
- To: Arthur Ryman <arthur.ryman@gmail.com>, "public-data-shapes-wg@w3.org" <public-data-shapes-wg@w3.org>
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256
On 04/27/2015 12:50 PM, Arthur Ryman wrote:
> I reviewed Peter's spec [1] and have some feedback. First, I think this
> is a positive contribution and is complementary to Holger's spec.
>
> This spec is a great example of how to define the semantics of a language
> by translating that language into SPARQL. We can use this approach with
> other language proposals.
>
> The semantics of the language is not completely defined using SPARQL.
> There is an upper control layer that is described informally. If we make
> this upper control layer a little more flexible then we should be able to
> support recursion.
I would like to see how this would work.
> The Extended Condition Language is somewhat low level and it would be
> difficult to use by anyone who does not understand how it is translated
> into SPARQL. I believe we need a higher level language for the non-SPARQL
> user.
The non-SPARQL portion of the Extended Control Language can be used without
understanding the translation into SPARQL, I believe.
For example, understanding
[ rdf:type sh:Condition ;
sh:severity sh:fatalError ;
sh:classScope "http://example.org/Person"^^xsd:anyURI ;
sh:shape [ rdf:type sh:Shape ;
sh:predicate "http://example.org/offspring"^^xsd:anyURI ;
sh:valueType "http://example.org/Person"^^xsd:anyURI ]
] .
should not require any knowledge of how SPARQL is involved. If the
representational purity is relaxed, one gets
[ rdf:type sh:Condition ;
sh:severity sh:fatalError ;
sh:classScope ex:Person;
sh:shape [ rdf:type sh:Shape ;
sh:predicate ex:offspring;
sh:valueType ex:Person ]
] .
> A more serious limitation is that there is no template mechanism, which
> means that non-SPARQL users contain use extensions defined by SPARQL
> users. We need a mechanism that allows either Core or Extended conditions
> to be packaged and invoked like macros.
Agreed. The proposal does not have a template mechanism. As the proposal
works via a translation to SPARQL, adding a template mechanism is certainly
possible.
> I have other comments and will put them in additional messages.
>
> [1] https://www.w3.org/2014/data-shapes/wiki/Shacl-sparql
peter
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iQEcBAEBCAAGBQJVPstJAAoJECjN6+QThfjzrsYIAMI+EoD2060tHPFbYkUFvzl2
vQ1lAt7LTrXG2R8xB3+R2pk8L6Lq2/IyrEGg8kJSlWUtq2Lcg+6JamddX15oPOXk
Qr8egiLOHyjcBPXKkEhy0jdsvky+m9ONNRpDqC2+ViL9Vk74oIKuSIDAHbGfGY7W
lqRl68kP+cz3avvB8YkVjt8cxJZKyClgVDYP5e6I5Gd6/kXssHeUDpJzg9voscch
Hj35REFmJH/HauYeO+ihp8xkP84LSpnxIWt4ylPsXHGKfG1GxseDxr1JprPvyem9
hU7ettA/PEdlGOaNtkTe0s6x9Cq0tpQNyjhI/EICLLAvuO7cRjeeJt2N5A3ngk4=
=ty21
-----END PGP SIGNATURE-----
Received on Monday, 27 April 2015 23:51:02 UTC