Re: [apps-discuss] Robert Sparks' Discuss on draft-leiba-urnbis-ietf-namespace-02: (with DISCUSS)

[Barry Leiba <[e-mail]>]

[Adding Peter Saint-Andre and apps-discuss, to include Peter's AppsDir
review, which raised the same issue.]

> As far as I can tell, this document doesn't Update the two RFCs it points
> to. These look like the "See Also" use of Updates we've agreed not to
> use. Am I missing something it changes in the earlier RFCs? If so, could
> that be written more clearly?

I believe that the IESG hasn't decided what "updates" means (and I
know I owe the IESG a draft proposal for this).  I also believe this
document qualifies, by criteria we have chatted about, though it's all
still up in the air.

The reason here matches what we've done in Apps in the past: when a
document creates a new registry that people reading the old document
will need to know about, we use "updates" to create that forward
reference.  Otherwise, people reading the old document will not be
aware that a registry exists, which they have to pay attention to.

In this case, RFC 2648 says this:

            If the IESG (or it successor) adds a new document series,
            this ABNF specification will need to be updated.

We aren't adding "a new document series," but we did, in RFC 3553, add
a new value ("params") for the "NSS" production in that ABNF, and 3553
should have updated the ABNF in 2648... but it didn't.

In the case of RFC 3553, it defined "params" but didn't register it
anywhere.  This document creates the registry and registers all the
items from 2648 and 3553.

I'm quite certain that this is valid use of "updates" for 2648, and
strongly suggest keeping that.  I will happily remove the "updates"
for 3553, because readers of that document probably don't need to know
that "params" is now registered anywhere.

Barry