Re: shapes-ISSUE-18 (S35 examples): S35 needs to state what constraints are required from Holger Knublauch on 2014-12-20 (public-data-shapes-wg@w3.org from December 2014)

From: Holger Knublauch <holger@topquadrant.com>
Date: Sun, 21 Dec 2014 08:35:00 +1000
To: public-data-shapes-wg@w3.org
Message-ID: <5495F994.8070409@topquadrant.com>

On 12/21/14, 5:29 AM, Eric Prud'hommeaux wrote:
> There are likely to be many ways that folks may want to select terms 
> for verification, by no means limited to: by type (as in conventional 
> SPIN), query (Axel's proposal), rooted graph (one option in most ShEx 
> interfaces), every subject node (another option is most ShEx 
> interfaces). Conventional SPIN and OWL provide a couple. 

As discussed previously, if we want comprehensive coverage of all the 
different ways to select terms for verification, then we likely end up 
with a language as complex as SPARQL. SPIN's global constraints 
(currently attached to rdfs:Resource) are covering all those cases with 
arbitrary patterns at the discretion of the user, without having to 
create a new meta-vocabulary. These queries can be made quite readable 
with the help of user-defined SPIN functions.

The selection of which constraints to execute happens via the usual 
named graph mechanism.

Holger

Received on Saturday, 20 December 2014 22:35:32 UTC