W3C home > Mailing lists > Public > public-shacl@w3.org > September 2018

Suggested additions to SHACL-AF (sh:values keyword)

From: Holger Knublauch <holger@topquadrant.com>
Date: Mon, 10 Sep 2018 17:50:59 +1000
To: public-shacl@w3.org
Message-ID: <2a9f0c54-e911-a8d3-bf16-a0d9f317ea3a@topquadrant.com>
Dear SHACL Community Group,

while this group is currently rather dormant, it is still the best place 
to discuss the future evolution of SHACL. I have recently worked on a 
handful of relatively minor extensions to the SHACL rules mechanism that 
was published in the SHACL Advanced Features note. I have written these 
extensions up in what could become a future Community Group note and 
eventually part of a SHACL-AF 1.1 document. See


In particular, I suggest the addition of a new property sh:values that 
can be used at SHACL property shapes to specify inference rules to 
derive the values of the property. Our initial experiments with this 
property are promising. In order to support this feature, I also propose 
a number of relatively straight-forward extensions to SHACL node 
expressions (e.g. to express aggregations and to use SPARQL queries in 
node expressions).

An introduction to how we are supporting these new capabilities (from 
TopBraid 6.1 onwards and in the current SHACL API) is here:


(Readers may want to skip the paragraphs on GraphQL).

Note that the new features are not related to the SHACL validation topic 
but on its lesser-known inferencing aspects. Users familiar with SPIN 
may see the connections between sh:values and SPIN's "Magic Properties" 
which have been used for many years. SHACL property value rules can be 
implemented and used with a similar backward chaining engine.

If anyone here is interested, I welcome further discussion on this 
topic. I am offering this as input to the Community Group.

Received on Monday, 10 September 2018 07:51:26 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:23:12 UTC