Compact syntax examples in Core document - Upkeep and draft document link

[ajnelson-nist]

On today's working-group call, someone mentioned that the Core document had picked up SHACL-C (Compact) syntax for its examples.

@YoucTagh noted this in the PR adding the SHACL-C snippets:

I am adding my review late, but there are 28 Shapes graph boxes with Turtle and JSON-LD only, with no SHACL-C. Is this intentional?

The SHACL-C snippet mention on today's call was pertaining to getting continuous-updating draft documents started up for several documents, including SHACL-C. As I understood the question on the call, it amounted to:

1. Should the Core document have SHACL-C snippets for its examples?

It seems this was already decided sufficiently in #419 to add the snippets. Do we need to revisit this decision?

If we remain fine with keeping snippets in Core, then we should probably consider this another nudge to get the SHACL-C document up to the same continuous-updating state as Core.

In light of @YoucTagh 's observation, we seem to have a review-rule need, that examples should consistently provide alternative syntaxes, maybe unless they self-excuse because we hit something SHACL-C can't currently support.

I suggest these resolution steps for this Issue:

• Affirm, or revisit, SHACL-C inclusion in Core. (Please discuss in Affirm decision to keep SHACL-C snippets in Core #462 .)
• Devise a system (whether programmed, or social (checkboxes in PR template?)) to keep SHACL-C consistent with JSON-LD on Core.
• See whether that SHACL-C presence system applies to any SHACL documents aside from Core.

[ajnelson-nist]

I posted this Issue in response to two other people raising points, so we didn't lose track of unresolved matters.

I'm personally not inclined to revisit decisions of #419 . I see it as a good opportunity to continually exercise one of the new outputs of this working group.

I'm curious what automated review systems are available and palatable to us for SHACL-C example-presence consistency.

[jeswr]

Also responded to this in #421 (comment)

All SHACL-C in #419 is auto generated. The cases where no SHACL-C is provided is because the serializer errored on features it could not handle. As @VladimirAlexiev and I work on SHACL-C 1.2 I suggest we make it a goal to extend the support of the syntax to ensure these cases are covered.

Affirm, or revisit, SHACL-C inclusion in Core.

I am in favour of keeping this

Devise a system (whether programmed, or social (checkboxes in PR template?)) to keep SHACL-C consistent with JSON-LD on Core.

The scripts for this are in #412 #421 - I suggest we simply add a CI check to make sure they - and the JSON-LD examples are up-to-date. @ajnelson-nist can you open an issue to track this if we have consensus on keeping SHACL-C snippets in core.

See whether that SHACL-C presence system applies to any SHACL documents aside from Core.

I propose yes - and the existing scripts can support this.

[ajnelson-nist]

[snip]
@ajnelson-nist can you open an issue to track this if we have consensus on keeping SHACL-C snippets in core.

Opened #462 .

[ajnelson-nist]

@jeswr - On following the link to 412, I suspect a typo. Which issue/PR did you mean?

[ajnelson-nist]

@jeswr - On following the link to 412, I suspect a typo. Which issue/PR did you mean?

Nevermind, must be #421. Carry on.