[Rats] Re: draft-fv-rats-ear-05.html

["Smith, Ned" <[e-mail]>]

[nms] not as chair.

On 2/9/25, 07:53, "Michael Richardson" <[e-mail]> wrote:


Thomas Fossati <[e-mail]> wrote:
    >>
    >> is this URI going to live into WG adoption and publication?

    > No, this is bound to change if the draft is adopted.

    >> (I understand that for running code, like veraison, that you need
    >> something now) While EAT does not require (and goes out of its way to
    >> avoid) a registry for EAT profiles, for core infrastructure things
    >> like EAR, I wonder if we shouldn't have something.

    > The rationale for employing URI/OID as profile types was to bypass the
    > need for maintaining a registry and instead leverage established
    > governed namespaces.  I see your point though: it'd be handy to have
    > the EAT profiles map in one place.

Good. So, what do others in the WG think?
[nms] The spec defines a base profile essentially already. Populating the profile field seems redundant. The content-format / media-type / other(?) content labeling scheme describes the payload contents. Since ear is a special kind of EAT (the ar in ear means Attestation Result), it should have its own media-type / C-F such as “application/ear-cwt”. The profile parameter might still be used for vendor/implementation specific considerations.

    >> IANA Considerations section should probably back-reference section 4
    >> (EAR Extensions).
    >>
    >> I see that this document is somewhat TEEP oriented/inspired.  That's
    >> fine, even very desireable.  It seems to me to be TEEP critical path!

    > We used TEEP as a playground to test the extensibility story.  On
    > paper, it works.  However, we'd need to double-check with the TEEP
    > people to see what they think.

Yes, I think it's really good that you used it.
I'm concerned that TEEP needs this to complete.  Am I wrong here?

--
]               Never tell me the odds!                 | ipv6 mesh networks [
]   Michael Richardson, Sandelman Software Works        | network architect  [
]     [e-mail]  http://www.sandelman.ca/        |   ruby on rails    [