JdeBP> Gresham's Law trumps Postel's Principle, [...]

MC> Actually, Postel's principle is misstated. I knew Jon
MC> personally. He *never* advocated forcing implementations
MC> to accept bogus protocol, and it is a perverson of Jon's
MC> legacy to claim otherwise.
MC>
MC> "Be liberal in what you accept" referred to accepting a
MC> wide range of valid protocol, as opposed to accepting only
MC> the most common form. [...]

Well, if "be liberal in what you accept" _is_ a mis-statement of his
principle, then he himself - as the editor of RFC 793 where those are the very
words - was responsible for that. And I've read other people claiming that
they spoke to him about RFC 791 and RFC 1122 and that he was actually quite
happy with the various wordings. So I'm not entirely convinced that that
wording _is_ a mis-statement of his principle.

Of course, even if it isn't a mis-statement, that doesn't mean that it
supports the contention that an SMTP Relay server should recognize bare
linefeeds as line terminators. (And it would certainly be good to say that it
actually didn't.) It all depends from what his definition of being "liberal"
actually was. If being liberal was intended to only extend only up to and no
further than the boundary of validity, then its use to support the contention
is, indeed, a perversion of the principle.

It's a pity that there's little to no explanation of the principle next to
where it is stated in the RFCs.

MC> The only negative thing about qmail insisting upon proper CRLF
MC> newlines is that the spammers may fix their SMTP engines and
MC> thus this natural immunity to spam will become less potent over
MC> time.

Ah well. No technical defence mechanism against unsolicited bulk mail that
relies upon an incidental, that is unrelated to the "unsolicited" and "bulk"
qualities of the mail, remains effective forever.