initial shaclc-grammars

[VladimirAlexiev]

@jeswr @HolgerKnublauch

• get 4 grammars
• try to convert to EBNF (2 worked, 2 had problems)
• convert to Railroad Diagrams in html

You can see the links at the bottom of https://github.com/VladimirAlexiev/shacl/blob/shaclc-grammars/shacl-compact-syntax/grammar/README.md:

• shaclc-ANTLR.html : this link will work after the PR is merged
• (or in branch): this link works now

[jeswr]

LGTM - the detailed README of steps you've taken is greatly valued.

I'll address the jison errors in a follow up PR, unfortunately my schedule has not allowed me to look into this yet

[afs]

In order to get to being a working group, the IP and copyright of material needs to be handled.

The different grammars are under different licenses or no license mentioned and have different copyright owners.
They are not under the W3C Software License.

This should be written down somewhere with attribution for copyright (README may be).

The Community Contributor License Agreement does not apply if the copyright owner isn't the one contributing the work.

One possibility is to to link to resources rather than copy them, then have a clearly labelled "snapshot" with date and explain this is to protect against link rot.

[VladimirAlexiev]

Hi @afs thanks for the "Copyright" comment!

• I think it's best to keep them somewhere outside of the w3c tree, and only put the synthesis (merged) version in here
• But where should that be? I could do it in my repo and rework it as its own "project" shaclc-candidate-grammars rather than a PR against this repo?
• One downside is that if we want to process these files in some automatic way, people would have to copy them locally or deal with my repo...

@HolgerKnublauch and @jeswr what do you think?

[jeswr]

The different grammars are under different licenses or no license mentioned and have different copyright owners. They are not under the W3C Software License.

Given that it is grammars coming from @HolgerKnublauch, @afs, @tw-osthus and myself; would it be an option - assuming @afs and @tw-osthus consent - to have those authors sign the CLA and then (1) make a comment on their PR noting that they consent to those grammars being added to this repo under the CLA or (2) create a separate PR adding their grammar to the repo.

One downside is that if we want to process these files in some automatic way, people would have to copy them locally or deal with my repo...

If you were to go with this approach then I would suggest making your shacl-candidate-grammars a submodule of this repo.

[afs]

I am a member of the CG and my organisation has signed the community group agreement.

The Apache Jena grammar is under the Apache License which allows reuse without restriction.

For the WG, the important thing will be the language and W3C-style BNF for the spec. Addition presentation material would be great to have but should not be normative.

Grammars for a variety of parser-generator is nice to have and can be links the source. Copying them into a repo implies they will be maintained though.

I've added the tooling to generate BNF HTML and uploaded the current output to a gist. It's not perfectly presented W3C-BNF but it should show what is possible automatically.

HTML:

2024-08-23-shacl-c

[tw-osthus]

I can't sign any CLAs, but the grammar was released with the Allotrope Shacl-C editor under Apache License.