Define/clarify handling of recursive shapes

[HolgerKnublauch]

https://www.w3.org/TR/shacl12-core/#shapes-recursion still mentions that recursive SHACL are undefined. It would be better if the behavior is made consistent and predictable.

CC: @osavkovic

[robert-david]

I think this should also be aligned with the rules.

[afs]

I think this should also be aligned with the rules.

What are the implications of that?

I think shape-validation-recursion is not related to rule-recursion because they happen in different processes.

[robert-david]

I think shape-validation-recursion is not related to rule-recursion because they happen in different processes.

I agree. Technically they are unrelated.
My thought was to make it easier to understand and implement if both follow the same notion of recursion, e.g. stratified negation.

[osavkovic]

Thanks Holger for creating the issue. As I mentioned during one of our meetings, I would be interested in working on this, since I have been working on SHACL recursion for the past few years.

My idea would be to write a document that summarizes some of the existing proposals and discusses their pros and cons. Alternatively, I could also give a short presentation if there's a broader audience interested. I am open for a feedback on this.

Just to add that over the next three weeks I will be a bit overloaded with a new course I am teaching, but after that I plan to be more active on this issue.