Re: The state of DNS support, was Deprecating SPF

On Sun, Aug 25, 2013 at 11:28:47PM +0100, Andras Salamon wrote:

> I don't know whether there are many examples or not, but ISC and
> Nominum seem to add new RRtypes to BIND whenever an RFC is published
> defining them, regardless of who employs the authors of the RFC.

Yes, and virtually every DNS server system actually does in fact cope
with TYPE99 records.  That is nowise an argument that saying, "Use
TYPE99, darnit," will actually change the world.

I find myself disagreeing with Dave way more often than not, but on
this point I think he's quite right.  We have a functioning system
that's actually deployed, as far as anyone has argued.  It is far from
perfect, but it works.  Those who think that the spec ought to be
changed to reflect some other desired behaviour are responsible to
provide evidence that the mere publication of an RFC -- one that
recommends against the overwhelming widely-deployed practice in favour
of something else -- are indeed responsible to show why that
publication would have the requisite effect.  Even one example would
suffice.  Like Dave, I'm having a hard time coming up with one.

As I've said before, I'm mighty unhappy about this.  But the world is
as it is.  Rather than fighting old battles, I'd like us to focus on
making this never happen again.

Best,

A

-- 
Andrew Sullivan
[email protected]
_______________________________________________
dnsext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsext