Re: Content-MD5

> Look, Ned.  Without thinking much I came up with four examples where the
> IETF has RFCs on "competing" standards, and without thinking hard I named
> several.  Your followup was non-responsive.  One isn't your area of
> expertise but you challenged me anyway, one you got wrong (Usenet, not
> UUCP mail), one you don't undersatnd the history of (SNMP), and one where
> I was sloppy in that I talked about character sets without enough context,
> apparently, for you to see that I was talking about all the charset
> definitions and transport issues that are floating around.

I still disagree with your assessments on all counts here, but hey, since it
seems like its vital for you to believe that the IETF endorses competing
standards all the time, then by all means have it your way.

What matters now is the present proposal. All you have to do is finish it and
take it to the applications area directorate and ask to have it put on the
standards track. I'll object to it then on the grounds that it duplicates
existing functionality in MIME and we'll see how well having competing
specifications with identical capabilities sits with the area directors and
then with the IESG. Maybe you're right and I'm wrong and they will all just
love having two of these. Maybe they won't settle for two and will want five or
six. All you have to do is try it and see.

> > In other words, you don't care enough to bother to try and reconcile the
> > two different schemes. That's fine with me, but surely you see it is this
> > sort of attitude that has led to the present situation?

> I don't understand how you could come to such an understanding.  I know
> that you saw my question on convergence statements and subsequent response.

I come to such an understanding directly from statements you have made --
including statements in your last message. You cared enough to bring this up in
private with an RFC author, but not enough to bring it up with the IETF and
IESG during not one but two last calls. This is why we have IETF last calls!

I never said you didn't want convergence. I said that you don't want to
actually have to do anything as tough as either using content-md5 in HTTP or
else moving content-md5 to historical to get it. It seems you'd rather have two
parallel specifications than face either of these alternatives. This is my
understanding, and everything you've said seems to support it.

				Ned

Received on Sunday, 5 November 1995 23:40:09 UTC