Re: Question on CR issues

On 21/01/2025 11:05, Sasaki, Felix wrote:
>
> Hi all,
>
> I looked at this list
>
> https://github.com/w3c/rdf-concepts/issues?q=is%3Aissue%20label%3Ams%3ACR%20 
> <https://github.com/w3c/rdf-concepts/issues?q=is%3Aissue%20label%3Ams%3ACR%20>
>
> and have a few questions.
>
> 1)
>
> https://github.com/w3c/rdf-concepts/issues/139 (Merge unstar mapping)
>
> The proposed section is marked as non normative, so it would not be 
> relevant for CR. There is a remark in the PR on having this normative 
> or not.
>
> What is the intent?
>
My understanding is that other specs might want to normatively refer to 
this mapping (e.g. a future version of https://www.w3.org/TR/rdf-canon/)
>
> If this would be normative, how would tests for this look like?
>
With the caveat that any test on rdf-concepts would have to rely on a 
concrete syntax,
testing the unstar mapping is easy: provide a number of pairs of graphs 
(resp. dataset): (input, expected output),
which should be tested in both directions (round-tripping)
>
> 2)
>
> https://github.com/w3c/rdf-concepts/issues/138 (triple terms in 
> subject position)
>
> How would a test for this look like?
>
For me, this feature falls under the caveat in the "Conformance" section

 > This specification, /RDF 1.2 Concepts and Abstract Syntax/, defines a 
data model and related terminology for use in other specifications, such 
as concrete RDF syntaxes 
<https://www.w3.org/TR/rdf12-concepts/#dfn-concrete-rdf-syntax>, API 
specifications, and query languages. Implementations cannot directly 
conform to RDF 1.2 Concepts and Abstract Syntax, but can conform to such 
other specifications that normatively reference terms defined here.

If the abstract syntax allows triple terms in the subject position, then 
Concrete would allow them as well, and contain (positive) tests where 
triple terms occur in the subject position.

>
> 3)
>
> https://github.com/w3c/rdf-concepts/issues/129 (distinguish RDF data 
> model from abstract syntax)
>
> How would a test for this look like?
>
That's a good question. This is about clarifying the terminology. I 
don't think that it *needs* testing (a definition can be normative, yet 
you don't test a definition, do you?...)
>
> 4)
>
> https://github.com/w3c/rdf-concepts/issues/128 (upgrade to graph 
> isomorphism)
>
> How would a test for this look like?
>
This one is a bit tricky because, AFAIK, graph isomorphism is only 
indirectly tested. More precisely, a number of tests on the concrete 
syntaxes /rely/ on graph isomorphism: you test that the graph parsed 
from a turtle file (input) is isomorphic to the graph parsed from an NT 
file (expected output).
>
>
> 5)
>
> https://github.com/w3c/rdf-concepts/issues/116 (rdf:json value space 
> incorrect)
>
> This discussion seems to be blocked. Could we clarify in the group if 
> we can move on without unanimity? See the mail from PA at
>
> https://lists.w3.org/Archives/Public/public-rdf-star-wg/2025Jan/0112.html
>
> and
>
> https://www.w3.org/policies/process/#Consensus
>
+1
>
>   * Felix
>