Re: [Suit] [COSE] draft-atkins-suit-cose-walnutdsa

[Michael Richardson <[e-mail]>]

Derek Atkins <[e-mail]> wrote:
    >> And the other two are Expert Review anyway, no specification required!
    >> You are even listed as one of the experts for COSE Key Types :-)

    > Yes, I know I am one of the experts for these registries.  However, I
    > feel I must recuse myself from approval of these.

Yes, obviously, but the point is that you will know how to ask nicely :-)

    >> You could also ask for early allocation via the COSE WG, but that
    >> probably requires the document to be adopted first!

    > Does it?  Or does the WG just need to approve of the document going
    > through any of the publication processes?

If the document isn't adopted, then I don't think the WG chairs can ask the
AD to approve.  I think that adoption is the only required step; your
document could take awhile to progress.  IANA will check back on progress for
the early allocation, to make sure the document is still of interest.

    >> That's what I would do: just ask for the values via Expert Review, and
    >> live for now, with the a five-bye (encoded) COSE Algorithm code, if
    >> your market need is urgent.

    > Yes, I could just request the entries in the registry.  However, I am
    > trying to be a good netizen; I want to have a document that people can
    > look at to understand what the registry items mean.

While you don't have to provide a specification for Expert Review, I think
that if you *do* that it will get recorded.  If you later on get a 1-byte
code allocated then, it's code complexity, but it's not too terrible, I think.
At some point, your constrained devices will only use the 1-byte code,
and your non-constrained devices will continue to be tolerant of both.

    >> Derek Atkins <[e-mail]> wrote: > We have customers who are
    >> looking to use this technology today, so we > would like to do it in a
    >> standard way that others could understand.  > Personally, I'm fine
    >> with an Informational (instead of Standards-track) > publication if
    >> that would make people happier.  Even so, the RFC Editor > would still
    >> require approval from this WG as it is within their (the > WG's) area.
    >> 
    >> yes, if you went ISE, which would still take awhile.

    > I am fine going through the ISE, however I suspect that the IESG in
    > their conflict review would then go and ask this WG about it.  So I
    > wanted to bring it up here, first.  What is your definition of
    > "awhile"?

Given the conflict review and the typical length of the RFC editor Production
queue, I would say you are looking at a year before AUTH48, but the early
allocation could occur much sooner.

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