--On Wednesday, 19 September, 2007 19:42 +0200 Eliot Lear
<lear(_at_)cisco(_dot_)com> wrote:
Here our views of history differ. I can't say that I had a
gray beard at the time, nor that I was at all involved in the
design, but I was there, and I think there was something of a
view that the whole point of MXes was for the MX relay to
bridge between online and offline devices.
As well as to intermittently or poorly-connected Internet
devices (the notion of backup via a host that was more able to
dependably reach the destination was well-understood too).
It was well
understood at the time that MANY more systems were in fact on
the networks that you mentioned than were on the ARPANET. And
I would even argue that there MXes solved a major problem,
which was that there was that sites that sat on both UUCP and
the Internet often times Got It Wrong with regard to the
precedence of "!". The abstraction that MX provided made
relaying behavior much more explicit. I think this was
understood at the time.
Very well understood by just about everyone involved, often by
everyone on both sides of the gateways. It is also the reason
why wild-card MXs ("route to this machine to reach all sites in
that university or country") were so important.
john
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf