I voted against a prefix.
I did this not just because of the "_" character support issue, but also
because:
-implementing a prefix adds complexity (comprehension complexity if nothing
else)
-many dns providers do not support users adding subdomains (or not through the
UI anyway)
-prefix does not solve the wildcard issue
-one can use redirect in a small record to a subdomain (simulating prefix) if
one really wants
prefix
Anything that hinders deployment needs to be considered for benefit and lack of
alternatives.
Terry Fielder
Manager Software Development and Deployment
Great Gulf Homes / Ashton Woods Homes
terry(_at_)greatgulfhomes(_dot_)com
Fax: (416) 441-9085
-----Original Message-----
From: owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org
[mailto:owner-ietf-mxcomp(_at_)mail(_dot_)imc(_dot_)org]On Behalf Of
william(at)elan.net
Sent: Sunday, September 05, 2004 6:31 AM
To: Michael R. Brumm
Cc: IETF MARID WG
Subject: RE: consensus call of RR prefix
On Sun, 5 Sep 2004, Michael R. Brumm wrote:
This is a good point, however:
- MARID has an immediate (emergency-like) need for
widespread deployment
like few other standards, and thus the precedent is
limited. Most (if not
all) other standards can depend on RFC 3597 (unknown RR
types) and/or
underscored prefixes.
The "immediate need" or emergency is in the eye of the
beholder. What may
seem urgent for us may not seem so urgent for somebody else
and similarly
what may seem like not so important matter to us maybe very important
and seemingly an emergency for somebody else.
The precident will be there, that one is certainly right, so don't be
surprised if somebody else wants to use it for their "emergency".
As such we better show good understanding of such emergency situation
and good example for that to follow in our technical solution.
- Requiring an underscored prefix to prevent this (likely
non-existent)
future problem prevents and slows adoption of MARID by
those whose DNS
providers do not allow underscores. It would seem strange if MARID
provided TXT publishing to improve deployment, but then
placed it in an
underscored prefix which impedes deployment.
I have a question for Michael here:
Is "_" as part of the prefix name the only problem you see (or the
primarily one) as far as what you see wrong about introduced idea of
using prefixes for placement of SPF TXT records?
Similarly if others here who have so far spoken against using
prefix also
see underscore as the main problem, please also speak up as I
believe it
is important issue for chairs to recongnize as far as
arguments against it
and how strong technically these arguments are.
--
William Leibzon
Elan Networks
william(_at_)elan(_dot_)net