Re: wcard discussions in Minneapolis

[Robert Elz <kre@munnari.OZ.AU>]

    Date:        Tue, 29 Mar 2005 13:57:25 -0500
    From:        Edward Lewis <Ed.Lewis@neustar.biz>
    Message-ID:  <a06200700be6f5083c205@[10.31.32.83]>

  | How about we document that there are inconsistencies in the 
  | definition of DNS which lead to unclear semantics when asterisk 
  | labels are delegation points or DNAME rewrite points.

I'm going to skip DNAME for a minute, since I don't recall its spec
well enough to comment,

But for NS, I don't understand this at all, nor most of the other comments
on this thread (the ones I've just read, which isn't all of them).

Everyone seems to be assuming that there's something undefined about
"* NS".    Nonsense.    What's there now is perfectly well specified.

It's useless, and a bit bizarre, but undefined it isn't.

Much of the "it is undefined" can be traced back to really meaning
"it isn't defined to do what I think it should do" (which is perfectly
possible, but utterly irrelevant), or "when I skimmed the RFC it wasn't
obvious to me what happens" which is also possible, and is (part of)
what the wildcard clarify doc is supposed to be fixing.

I fully understand that lots of people don't like having it defined as
it is currently, and making it be undefined would be a fine first step
in later being able to supply a definition (different than the current one).
But if that's to happen, everyone needs to be very clear, that what is
happening is a change to the DNS - an explicit and deliberate change, not
just a clarification or fixup of what has been there for decades now.

That might be OK (after all, the WG did decide to change the definition
of "* CNAME" - and recognised it as being an actual change).

Just please, when considering this question, don't start out from
"it is undefined" or "it is ambiguous" or anything else like that,
because it isn't.

What is there now is perfectly well defined, and 100% useless for
anything.   In some sense, all the wcard clarifications doc really
needs to say is that - just make it clear to everyone that "* NS"
isn't going to produce any meaningful result (in the wildcard sense).
Even explaining why that's true probably isn't required.

kre



--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>