On Fri, 2005-07-08 at 11:42 +0000, Julian Mehnle wrote:
You're missing the point of SPF. We _want_ forwarding without sender
rewriting to break in the general case (that is, unless the receiver
whitelists his trusted forwarders). We consider forwarding without
sender rewriting a broken legacy feature.
Not even all of the SPF supporters would claim that. Many are happy to
admit that the only thing 'wrong' with the long-standing practice of
forwarding is that SPF can't handle it.
I said this to Dick St Peters on Wednesday:
"There is a technical incompatibility between the flawed initial
assumptions of SPF and the long-established practice of forwarding mail,
but that doesn't mean that mail forwarding is being abused. Why do you
say that it's being abused?"
Stuart jumped in with the response I'd already given as part of the
question: that forwarding doesn't work with SPF. Dick phrased it
slightly differently but said basically the same thing.
But nobody volunteered any reasons why, in the _absence_ of SPF (which
we know has technical problems with forwarding) there is any other
reason to consider it to be a bad thing. I'd heard nobody say that
forwarding was wrong, or 'forgery', before SPF was invented. What,
_other_ than the technical problems of SPF, has changed?
We know the caveats about content checking causing bounces -- that kind
of thing happens anyway with backup MX (and in fact it's worse with
those backup MX hosts which don't know which addresses are even valid).
That's not a reason to consider forwarding to be evil, unless you're
also going to abolish backup MX too, and then you might as well just
take us all the way to IM2000.
Yes, many, like you, don't agree with that. I guess it is going to be a
struggle for which side gets backed by more people. Your best chance at
winning this struggle is to explain to people why forwarding without
sender rewriting is worth to be preserved.
I don't often claim that forwarding without sender rewriting _should_ be
preserved. I merely opine that it _will_ be preserved. Feel free to get
a replacement for RFC2821 published which mandates SRS, to show me that
it's actually been obeyed in the wild by a large number of sites, and
prove me wrong.
Forwarding without sender rewriting is the status quo. You aren't going
to achieve a massive change to existing practice merely by showing that
there's no _overriding_ reason why it should stay as it is. Again, if
your head is that far into the clouds you might as well go all the way
to IM2000. I'll _definitely_ not be arguing on SMTP's behalf, if you
want to throw it away and start again. Roll _that_ one on, by all means.
--
dwmw2