[NNTP] Re: Version numbers
Mark Crispin
mrc at CAC.Washington.EDU
Mon Nov 8 10:00:55 PST 2004
I strongly suggest that before even contemplating a version number
mechanism such as Clive proposes, that you look into the experience of
other protocols. In particular:
. Why is the MIME version number forever fixed at 1.0?
. Why is the MIME version number forever fixed at 4rev1? [Hint: read the
next question.]
. Why do the IMAP people shudder whenever they refer to the transition
from 4 to 4rev1? Does much IMAP client software actually work with a 4
server?
. Why do HTTP people shudder when asked about HTTP versioning?
In my experience, a version number facility does nothing except to add
substantial complexity and open the door to mischief.
What we need is One True definition of NNTP, an indicator of compliance
with this One True definition, and perhaps a slight crack left open for
extensions. We need to *abolish* the myriad non-interoperable variants
that currently exist, not bless them. We need to *simplify* the requires
on clients, not make them worse.
NNTP is a dinky stupid little protocol. If it was a car, it would be a
Yugo. It does not need Ferrari gauges. We can try to make this Yugo be
reliable (unlike the car...) and easy to implement. We're not going to
accomplish this with complex versioning.
-- Mark --
http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.
More information about the ietf-nntp
mailing list