Incorrect type used for CoRIM profile in `veraison/corim`

[yogeshbdeshpande]

The IETF RATS CoRIM document Corim Version 07, defines Profile as either a URI or a Tagged OID Type.

It is not exactly EAT Profile, as the EAT profile is not tagged.

Veraison/CoRIM Repo uses profile as EAT profile.

This issue tracks the changes needed to align the CoRIM Profile implementation to exactly align with the CoRIM Spec.

The changes are needed at multiple places.

Inside the corim/corim package

Any other packages using this ID, example, comid/tdx-profile and newly work in progress coev package!

[thomas-fossati]

Another way to look at this is to ask why CoRIM deviates from EAT with regard to encoding.

After all, it adopted the entire concept and semantics from EAT.

In other words, isn’t this something that CoRIM needs to fix instead?

[yogeshbdeshpande]

No where in the CoRIM spec we say that the profile is the same as EAT.profile. They do not need to be the same as well,

[thomas-fossati]

There is no reason to diverge. It's just more work with no benefit: they have the same semantics and allowed types (URI and OID). The only difference is an insignificant detail, which complicates implementation and prevents code reuse.
Since EAT is already standardised and CoRIM isn’t, it makes sense for CoRIM to just embrace the EAT encoding convention.

[yogeshbdeshpande]

There is no tagging in EAT profile I guess, but that is a minor point!

[thomas-fossati]

Exactly, it’s trivial to reconcile: ~uri / ~oid instead of uri / oid.

[yogeshbdeshpande]

Will discuss this in the CoRIM Meeting. Remember this has a wider impact on CoEV also.!

[yogeshbdeshpande]

And there might be implementations out there in the fields which will break, making this change to untagged oid

[thomas-fossati]

As has been repeated far too many times, the fact that someone has committed their implementations to a draft document, i.e., documents that the IETF defines as:

hav[ing] no formal status, and are subject to change or removal at any time [and] therefore should not be cited or quoted in any formal document.

must not prevent us from doing the right thing.

[yogeshbdeshpande]

is the EAT Profile, extensible, I guess yes, so should be ok?

[yogeshbdeshpande]

have a look at ietf-rats-wg/draft-ietf-rats-corim#90 and the conclusions!

[thomas-fossati]

have a look at ietf-rats-wg/draft-ietf-rats-corim#90 and the conclusions!

I read it. The arguments are not particularly compelling.