Ruben Verborgh
Ruben Verborgh
The current ACL system accepts agents and groups thereof. One way is indeed to start building the above into extension vocabularies. However, the open-world assumption of RDF can byte us...
Not if it’s a one-time thing though. Sort of like git repo access versus pull request.
> @rubensworks Can you please elaborate? Perhaps he wants to, but I certainly can 🙂 So you wrote: > I think this use case can be generalised to: > >...
Both are a PATCH to a document: 1 to a non-ACL, 2 to an ACL document. But since both are a collection of triples, there's not much difference.
We should do the former; no need to prescribe what exact HTTP request Bob should be making (who might be running a different user agent). We could just use the...
Why 202? If that action is posting an LDN, just 201 is correct.
As far as the server is concerned, it is just a request to add an item to a container (an LDN to an inbox), and that request has succeeded. The...
While 202 indeed avoids the Location header, it also has different semantics, which do not seem to be met. 204 is better for that case.
I’m still missing an example of an unprocessed request that warrants a 202. As I wrote, a posted LDN is processed when added to the container, so 200/201/204 are appropriate....
It was just a sketch, copy/paste from elsewhere. Needs to be attached to an LDN and POSTed as usual.