Re: IUTF8 pseudo-terminal mode

[Mouse <mouse%Rodents-Montreal.ORG@localhost>]

>> Or, anyone who's as fed up as I am with the "rough consensus,
>> running code, and a whole lot of stupid hoop-jumping" way the IETF
>> runs today and who holds a domain can define an extension, with no
>> coordination with anyone else needed, and have it working in the
>> time it takes to implement it.
> However, no one did that, either.

Well, except for the "other people implement it" part, I did.  To the
extent I can tell, at least; as I wrote, I can't really test it because
I don't run any systems that have IUTF8.

> I'm sorry if you feel that writing and submitting an internet-draft
> to get number assigned in a finite namespace is "stupid
> hoop-jumping".

It's more that writing an I-D at all these days demands it.

RFCs were originally supposed to be "float an idea".  Then they got
ringed about with bureaucracy.  So I-Ds were invented to be the "just
float an idea" mechanism.  Now _they_ have been ringed about with
bureaucracy - if I have to include a hundred lines of boilerplate in
order to float a five-line technical idea, something is broken.  That
is the hoop-jumping I refer to.

I would be happy to write up a document outlining the proposed
assignment for IUTF8 - though there's really no need, as someone
already did that upthread.  I not, however, willing to jump through the
stupid "include this boilerplate on this condition, that boilerplate
always, this other boilerplate if your age is prime or it's a
Wednesday" hoops which are necessary (or were last I checked) for it to
be accepted as a valid I-D.

> Someone asked about getting a number allocated, and Bill and I just
> described what would be required given the policy _for that registry_
> as specified _by the secsh working group_ back when the documents
> were published.  I'm pretty sure you were there at the time.

I was.  There were numerous respects in which I considered the
doucments flawed.  I spoke out about some of the ones I considered most
important; I don't think this was one of them, though if those changes,
the ones I considered important, had been made I may have worked my way
down to items this minor eventually.  (I don't really consider this a
failing of the ssh spec, but rather a failing of IETF procedure in
general.  That is, I blame the current meaning of "IETF consensus", not
the ssh spec, for IETF consensus being too high a bar for this.)

> Then again, given the number of times people have spoken up in the
> last 7 years about wanting a number, maybe it just doesn't matter.

Well, I would have liked to have a number or four myself, for the four
pieces of tty state that I put into
missing-pty-modes%rodents.montreal.qc.ca@localhost.  But I didn't bother saying
anything _because_ I knew the procedure would be ridiculously
bureaucratic; it was far easier to just use an extension and be done
with it.  (I also had no will to push something like that, especially
after just having done my best to get a much more important problem
with the spec fixed and gotten nowhere.)

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		mouse%rodents-montreal.org@localhost
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B