If the URIs do end up looking drastically different, it should come with improvements in size, and readability.
If the URI is just different, not smaller or easier to understand, it's not an improvement. I'm not attached to one registry over the other, but whichever one gets picked will probably need to be updated. URIs are typically meant to encode readable strings, the COSE algorithms registries primary purpose is to give integer identifiers to those strings. Ideally they would be layered when they can be, so COSE registry would point to the names hash algorithms registry when possible. It does seem strange to use oauth URNs when OAuth generally doesn't use COSE. ... What are the alternatives? OS On Sat, Dec 23, 2023, 2:34 PM Carsten Bormann <[email protected]> wrote: > On Dec 23, 2023, at 21:11, Michael Jones <[email protected]> > wrote: > > > > Like Hannes, I disagree with the assertion that the IANA "Named > Information Hash Algorithm Registry" is not well-maintained. Like any > other IANA registry, if you want a new entry, create a specification and > register it. I've created many specification to just that, such as RFC > 8230 and RFC 8812. If you have an itch, scratch it yourself! It's the > IETF way! > > Sure. > It just hasn’t happened as much as it has for the COSE algorithms > registry, which is the dose of reality I would like to add. > > But my point really is: > > > Finally, like Hannes, I would prefer that we continue use the same hash > algorithms registry as the JWK Thumbprint URI spec [RFC 9728] does for the > same purpose. Unnecessary differences when doing the same thing should be > avoided. > > And I would prefer that the thumbprints use the same Hash registry that > all other COSE specs use. > > Grüße, Carsten > >
_______________________________________________ COSE mailing list [email protected] https://www.ietf.org/mailman/listinfo/cose
