🚀 go-pugleaf

In article <davecb.724685064@yorku.ca> davecb@nexus.yorku.ca (David Collier-Brown) writes:
>  SAML and SOML are underused on the internet, yet the equivalent
>on BITnet is heavily used....

The only problem it that BITNET is dying... (not exactly a problem, i'm very
glad than RJE and all that stuff becomes a matter of folklore).

talk is much better than SMTP's SEND anyway.

>  Don't take them out of the spec: you may want them later,and a standard
>way of doing them is desirable.

Nobody calls to change SMTP commands. All we need is to get rid of the
7-bit restriction.

>  VRFY is utterly necessary to postmasters, and is therefor mandated by the
>Hosts Requirements RFC: finger is shut down on all too many sites, and so
>vrfy/expn are required to verify adressees when mail goes astray.

MAIL/RCPT followed by QUIT works just as fine. VRFY and EXPN practically
never issued by software (though it can save some typing).

>  And HELO is the hook for stronger authentication: even now it and thge
>recieved lines allow me to spot (poorly-)forged mail at a glance.

It you need something even remotely resembling autentication you'd better
use DNS to get the name from peer's IP address. Any forger with IQ higher
than 50 can say HELO with anything he wants. As it is it's nothing more
than wasted bandwidth.

>(small flame)  If you've implemented smtp, you should really know all that...

Somehow my software (uumail/batchmail/newsserver/uucp line disc.) handles mail
and news on all backbones in ex-Soviet countries. And as i already said
the 8-bit problem was solved without adding ugly encodings to already
bloated software. Flame somebody else.

BTW, the central RELCOM backbone handling the number of links comparable with
UUNET's currently runs on a couple of uVAX-es III and few 486s. The "vanilla"
e-mail software is uttelrly sloppy and inefficient and we had to hack
it a lot (or sometimes to throw away and rewrite from scratch).

--vadim