RJ Atkinson wrote:
Rather than quibble about the details of this, I'd
urge folks to support the move to 2-track.
If it becomes clear later, after experience with 2-track,
that 2-track needs to be further refined later, then
the community can always do that. In the meantime, it
is quite clear the 3-track system is not working.
I'd rather redefine the qualification criteria for the
3rd maturity level than getting rid of it.
I think it would be ridiculous if the IETF declares
a specification a "full standard" if >>90% of the
installed base is still one or more protocol revisions
behind.
Take TLS as an example. While I assume it might be
possible to find a few interoperable implementations
of TLSv1.2 (rfc-5248, Aug-2008), it would be ridiculous
to declare this a full IETF standard, because in reality
it is not actually used.
From a recent survey done by Yngve N. Pettersen about
rfc-5746 patch status of public TLS servers on the
internet:
http://www.ietf.org/mail-archive/web/tls/current/msg06432.html
- 99 of 383531 [servers] support TLS 1.1
- 2 of 383531 [servers] support TLS 1.2 (both are known test servers)
A specification only deserves the "full standard" label if
there is a significant amount of usage (IMHO > 20%) of the
installed base actually uses this particular specification/protocol
revision.
If there is a lag in adoption, then this is likely an
indicator of feature creep, i.e. insufficient separation
of non-essential functionality into true options.
Look at IPv6 as another example:
Although there is a significant installed base at least in
theory, this part of the implementation is normally
disabled because it cannot be used anyway.
The full standard label should be reserved for a protocol
or technology that has achieved a significant usage
in the marketplace and by that proven that it is an
adequate technology living up to the market and
consumer requirements as-is.
The current "adoption lag" for several IETF specs is a clear
indicator that there is something wrong with the development
process in the IETF.
An approach of making the demonstration of independent interop
the qualifier for "full standard" is how Management nowadays
creates success -- by definition rather than by achievements.
-Martin
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf