Keith Moore writes:
Dan, you're simply wrong. Header fields have been designed that
modify essentially every aspect of email operation
No. New header fields are OPTIONAL protocol extensions. Unrecognized
fields are, as a rule, skipped, so programs using the unextended
protocol are not affected by the header fields.
Is it possible to create a message that simultaneously (1) uses new
fields and (2) causes failures in programs unaware of the new fields?
Of course. But this failure is _not_ caused by the presence of the new
fields; it is caused by the misuse of the old fields.
When you mischaracterize extensions as ``arbitrary modifications,'' you
are ignoring the simple fact that extensions preserve interoperability,
while ``arbitrary modifications'' can break interoperability.
You say that HTTP caused ``headaches'' for network operators. You say
that new header fields deployed without your infinitely wise guidance
often ``disrupt'' the email system. I bet it horrifies you that millions
of users are chatting on Internet telephones without waiting for the
IETF to finish coming to ``consensus'' on the monster known as SIP.
---Dan