From: Jean-Marc Desperrier (jean-marc.desperrier@certplus.com)
Date: Tue Jan 07 2003 - 15:01:08 CST
Keith Moore a écrit:
>>Well then maybe the IETF needs a reality check too.
>
> there seems to be a lot of need for reality checks.
It's always a good thing to do a reality check.
[...]
> widespread current practice is to do things that work on a local
> level, even if that practice diverges from one user to another,
> or one group to another. [...]
>
> in standards work, it's more important to specify something that can
> be made to scalably interoperate than to specify something that is
> compatible with existing practice.
I fully agree with all of that.
I spend time discussing USEFOR, because I'd love the current
non-scalable situation to change.
My point was that the argument that it's bad and evil to make an
incompatible change that would mean that 8 bit data would circulate
somewhere where it's not allowed in the current standards is weak
because the actual situation is that a lot of 8 bit data is being sent.
And that the current endorsed solution for sending non US-ASCII data
(RFC2047) has so many problems that in many situations, messages are
more likely be handled correctly by sending them in raw 8 bits that by
using RFC2047.
It was not my point to support any solution that does not scale.
I would 100% approve that a change for something that does not scale is
bad and evil.
But by saying that I have misrepresented Usefor which I regret.
The current version is almost 100% compatible with email standard.
Maybe it can be made 100% compatible by enhancing the part that is still
problematic.
I just intended to voice why usefor is favoring an 8 bit solution in
usenet, not to say change incompatible with email are a good thing if
there's any way to avoid them.