In <5AB6FF56-14AF-11D9-8E16-000A95B3BA44(_at_)hxr(_dot_)us> Andrew Newton
<andy(_at_)hxr(_dot_)us> writes:
On Oct 2, 2004, at 11:32 AM, Miles Libbey wrote:
is that you believe that hosters will have a tough time managing
DomainKey-like DNS records. If this is indeed true, MARID must have
been a nightmarish proposal to those folks. IPs change frequently and
sometimes without notice, and can't be shared among different domains.
[...]
I didn't follow the MARID discussions -- was there a thread about this?
Does your quote generally summarize the discussion? Was this one of
the primary reasons there wasn't overall consensus in MARID?
The concern was largely ignored, though I think CSV actually addresses
the concern.
I disagree that the problem of maintaining LMAP-type systems (which
includes SPF), was largely ignored. That was one of the difference
between SPF and most of the other LMAP/CallerID proposals, since SPF
allows you to use indirection based off of A RRs and MX RRs. (I
haven't done exact calculations, but it appears around 75% of all SPF
records do not specify any IP addresses directly.) This was
discussed quite a bit early one.
Of course, such indirection has costs and whether those costs outweigh
the increased reliability is an area where I think reasonable people
can disagree on. I also think such discussions are way out of scope
for this WG. I just don't think that such misconceptions as raised by
Miles should go unanswered.
-wayne