Scope of SHACL Compact Syntax

[afs]

SHACL Compact Syntax is an RDF syntax that generates RDF for a shapes graph.

What is the SHACL coverage of SHACL 1.1 Compact Syntax?

If it is not "all of SHACL 1.1", then what is the subset and is it useful (and worth the WG effort) to have that?

If it is all SHACL, then:

• Syntax for node expressions
• Syntax for SHACL Rules
• Syntax for SPARQL-based constraint components
• Syntax for SPARQL validators, inc. a prefix mechanism
• Syntax for Non-Validating Property Shape Characteristics (including formatting messages)
• Syntax for profiles (?)

[jeswr]

Is this now a focus / topic of discussion in the WG - will return to attending calls if so. My view is that we should have the ability to escape into turtle to support the expression of statements that don't have a special syntax defined.

[afs]

At the moment, we need to see how all the pieces all fit together. We don't need a final SHACL-C.

The key one is node expressions, and writing node expression declarations.

It seems a good idea to have Turtle available (at any point between top-level syntax elements, e.g. turtle { .... } ).

[VladimirAlexiev]

These are closely related to "scope":

• consider embedding SPARQL syntax in SHACL-C #201
• consider Extra SHACL features #199

"Escape to Turtle" is

• Extending SHACL-C to enable the expression of any RDF statements #186

@jeswr if you don't mind, I've assigned this to you?

[afs]

There are scope issue other than node expressions so please do not focus this issue solely on NE.
I think the first step is to establish the intent of SHACL-C.

Is it all of SHACL 1.2?

[afs]

Simply allowing Turtle might give theoretical coverage but how does compact syntax shape (target, sh:values, etc) refer to a node expression?

[VladimirAlexiev]

@afs: I agree with you: NEs are important, but not the only feature to be addressed.

And are you saying it's time for me and @jeswr to get busy on a draft grammar?
Btw, it should allow named prop shapes, so they can be reused.

I'd say the scope of shacl-c is all or most of shacl. If an important feature is missing, people just won't use it.
Of course, it can have profiles, to match shacl's profiles.
I think the escape to Turtle is only for non-standard (non- sh) triples.

I think we should allow Turtle at two levels:

• top level as s,p,o
• inside a shape as p,o

[afs]

And are you saying it's time for me and @jeswr to get busy on a draft grammar?

At the moment, we need to see how all the pieces all fit together. We don't need a final SHACL-C..

I think the first step is to establish the intent of SHACL-C.

Do you agree with the list in the description? Is anything missing?

@VladimirAlexiev Is there anything that you think will not / should not / can not be covered?