[Top] [All Lists]

Re: I-D Action:draft-klensin-rfc2821bis-10.txt

2008-04-18 13:08:55

Lisa Dusseault wrote:

My serious offer here is that if somebody sees a way towards consensus on the implicit MX issue, please help find it or build it. The result of that consensus may not go into this document, but I will help to get the consensus documented and published.

What are looking for here?

All I see here is an "Enforcement" argument.

    Thou shall not do implicit MX lookups anymore.
    Thou shall enforce MX records.

Even if this was written into 2821bis, I fail to grasp how any commercial vendor is their right mind is going to heed that without being concern of the potential harm to their customer base, and worst - increase support issues!

Maybe if there was some NEW identifier marker that would put the client into a whole new category, then we can say

       "ok, now that DUDE must have MX records"

Can we say this for an IPv6 layer connection? Well, there might be opportunity to do this with minimal damage. But we really don't know, do we?

We can't say that for IPv4 connections. The long term autonomous nature of the SMTP system will broken.

Paul is off base with a presumption this has no cost impact.  It isn't
just 2-4 lines of code change.

At best, it can probably be mentioned as a DIRECTION so the contemporary vendors can assert a deprecated option. But to completely
turn it off without any other information to work with?  I don't think so.

And off course, none of this stops DNS managers / SMTP operators from running a proper setup too.

Anyway, Tony decided. When will it end? or isn't that sacred any more either?


Hector Santos, CTO