On 2/5/2010 9:27 PM, John Levine wrote:
I think the latest round of discussion has eliminated any interest in the
message retrieval mechanism.
Um, this discussion has been about keying the ARF report to the name
of the POP or IMAP server, remember?
name of.
not protocols details of.
On 2/5/2010 9:30 PM, John Levine wrote:
>> The right hand side of a SRV record is a hostname, not an email address.
>> (Trivially worked around in at least two obvious ways).
>
> Right. I suggested using a fixed mailbox name, and there's always
> the SOA hack
That's very Procrustean of you.
Given that the right hand side can be variable, why proscribe the left hand side
when that's not necessary, when using a TXT record.
> I don't see any a priori reason to expect MUAs to be any worse at
> looking up SRV than looking up TXT.
What about: TXT is older and pervasive?
The 'older' means exactly that its, ummmm, a priori...
> What does an MUA do now in normal
> operation, as opposed to funky plugins, that needs TXT records?
Is there some reason this line of design debate is being based on theory rather
than empirical knowledge?
Let's choose an inferior RR because it's just as likely to have been implemented
as a more useful RR (that's been around longer)?
d/
--
Dave Crocker
Brandenburg InternetWorking
bbiw.net
_______________________________________________
Asrg mailing list
Asrg(_at_)irtf(_dot_)org
http://www.irtf.org/mailman/listinfo/asrg