solid
Solid OIDC Issuer Discovery
In #oidc-issuer-discovery:
A WebID Profile lists the OpenID Providers..
Mention the property solid:oidcIssuer somewhere up front - in brackets at the very least.
Consider changing code after #example-75c9f7b4 :
?webid <http://www.w3.org/ns/solid/terms#oidcIssuer> ?iss .
to an Example.
Mention the property solid:oidcIssuer somewhere up front - in brackets at the very least.
Currently, we don't define any namespaces in this document, it doesn't really need it. Instead of using CURIE, provided triple pattern uses full IRI for the predicate.
If you suggesting to mark the triple pattern as an example, this will not work since it is part of a normative text. We clearly specify what statement needs to be checked, including what needs to be substituted for both variables.
To discover a list of valid issuers, the WebID Profile MUST be checked for the existence of statements matching
?webid <http://www.w3.org/ns/solid/terms#oidcIssuer> ?iss .where ?webid is set to WebID. The ?iss will result in an IRI denoting valid issuer for that WebID.
-
Namespace is not the point. The first paragraph (and even the first sentence) dives into describing providers/issuers. This is a good place to tie the notion to
http://www.w3.org/ns/solid/terms#oidcIssuerwhich will be picked up shortly after. -
That code is an example resembling SPARQL. Secondly, it neither says it is SPARQL or requires it. Reconsider and clarify what that line of code is intended to achieve. Thirdly, the text explanation around the code is not an issue but state using RDF terminology (as opposed to SPARQL variables) - along the lines of "where the subject of the statement is... and the object is.."
If knowledge of SPARQL is required to understand the Solid-OIDC spec, see https://github.com/solid/solid-oidc/issues/23
Effectively, we are representing a triple pattern in normative text. This is something SPARQL 1.1 defines, so we should add a reference to the relevant section from the SPARQL spec when describing that triple pattern. There might be other specs that define this sort of triple pattern, but SPARQL is likely the canonical reference here.
Thirdly, the text explanation around the code is not an issue but state using RDF terminology (as opposed to SPARQL variables) - along the lines of "where the subject of the statement is... and the object is.."
I believe in this case we will still need to add RDF 1.1 Concepts and Abstract Syntax as a normative reference.
I think we could go either way, SPARQL Syntax for Triple Patterns seems clearer for me even implementations will not use SPARQL to query the graph of WebID Profile.