One system that seems to be work pretty well, without the legal
liabilities, is that of interoperability test events (or b*k*o*fs,
before Pillsbury made that word illegal and threatened to sue us...).
For SIP and RTP, there have been nine so far, taking place at four-month
intervals. While this is no guarantee of interoperability and
robustness, it has greatly helped, with upwards of 50 companies
participating each time. (See http://www.sipforum.org and
http://www.cs.columbia.edu/sip/sipit)
These events also benefit the IETF: The interop events have helped
ferret out a number of ambiguities and corner cases in the spec that
were fed back into the working group. Having a close coupling between
the community that participates in the events and the working group has
been very helpful. Also, there's hope that this activity will greatly
simplify identifying interoperable implementations when elevation to
Draft standard becomes an issue. As is well-known, the Internet mostly
runs on Proposed Standards, with part of the blame probably belonging to
the pain and suffering required to get vendors to cough up interop
statements.
I don't think the IETF needs to be or should be directly involved in
these events, but I think they are a good tool that ADs and WG chairs
can avail themselves of during the later stages of maturing a standard.
The SIPit events have been done on a not-for-profit basis, with "ticket
prices" of around $100/person to cover direct host expenses, making it
easy for both big and small companies to play.
Keith Moore wrote:
can you read this mail? consider what it took to get it to you. how did
that happen?
and then consider the number of messages that fail to be delivered,
or messages that cannot be read properly (because they're not properly
tagged by the sender's MUA, or because the recipient's UA doesn't like
one of the headers on the message), or messages that deliver
viruses to their recipients computer systems, or messages that
resulted in mail loops or sorcerer's apprentice syndrome - because
some MTA or MUA vendor failed to conform to the standards.
as useful as email is, the reliability of email is not even close to
that of snail-mail. and that's pretty sad.
Keith
p.s. though mis-configuration of MTAs and DNS servers appears to be
a far larger source of errors than mis-implementation.
-
This message was passed through
ietf_censored(_at_)carmen(_dot_)ipv6(_dot_)cselt(_dot_)it, which
is a sublist of ietf(_at_)ietf(_dot_)org(_dot_) Not all messages are passed.
Decisions on what to pass are made solely by Raffaele D'Albenzio.
--
Henning Schulzrinne http://www.cs.columbia.edu/~hgs