Peter J. Holzer wrote:
STD 10 clearly demanded to add the host to the reverse path.
[...]
The reverse path _was_ modified before 1123.
Was it? Just because RFC 822 demanded it, doesn't necessarily
mean it was ever widely implemented.
s/822/821/ I can't tell, when I tried to figure out sendmail
it was for a UUCP route to some kind of "emergency smart host",
and some years _after_ 1123 was published.
Today my general impression is that Postel wrote good stuff
not limited to STD 10. He already had the very same concept
in RfC 788 (obsoleted by 821), so I guess it's no nonsense.
RfC 821 was published in August 1982, 788 in November 1981.
In RfC 780's "MTP" (May 1981, obsoleted by 788) it was also
the same concept (apparently, better check this). In RfC 772
(September 1980, obsoleted by 788) also about "MTP" I find no
reverse path / error handling concept (again please check it -
I never had any reasons to look into this old stuff before ;-)
So _maybe_ he first tried to get away without reverse path and
error handling in 772. Nine months later he found that this
is not good enough, and added reverse paths and error handling
in 780 to MTP.
The he sticked to this concept up to STD 10 as we know it, for
16 months. "Nobody implemented it" doesn't sound plausible
for a full Internet Standard, there was enough time to modify
this "reverse path" concept if it didn't work out in practice.
Okay, seven years later "they" pulled this stunt, in RfC 1123,
after discussing it for 18 months (see chapter 1.4, the list
of credits is very interesting, many folks that are Internet
legends today, 2822 has a similar "who-is-who in the IETF" :-)
Bye, Frank
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg