As for the duration of this process, I am NOT sympathetic to
the objection you raised. I said this when you filed the revised
draft just prior to the last IETF meeting, I reiterated it in my most
recent message to Amanda (recapping the history of the MIME-PEM
specs), and I'll say it one more time. When the authors took over 4
months to respond to the first set of comments provided on the March
draft, they forfieted all right to demand a speedy resolution to this
process, period.
Steve, up until now I have let comments along these lines pass as being so
completely irrelevent that they are not worth the time to reply to. This one,
however, goes too far. You are spouting gross inaccuracies here and you know
it, and I cannot let this continue any longer without presenting the true facts
of the matter.
Here is a rough cronology of the MIME/PEM work. This cronology can easily
be verified by studying the archives of the PEM-DEV list.
The MIME/PEM document began as an idea of Dave Crocker's during the November,
1992 IETF meeting. Marshall Rose quickly drafted a specification describing the
idea and it was first circulated on November 12, 1992. It was subsequently
posted to the PEM-DEV list on November 16, 1992, along with a request that time
be allocated to discuss the specification at the following meeting.
There was some discussion at the meeting and subsequently on the list, but that
discussion petered out before the end of the year. It had been generally agreed
on that the overall approach would be reviewed again at the next IETF meeting
in April. Everyone seemed to be waiting for that meeting to occur.
Marshall found himself too busy to continue as editor so I assumed that duty. I
published a revised version of the specification on March 9, 1993. I note in
passing that the revised specification was published almost three weeks prior
to the next IETF meeting in this case, over a week before anyone responded in
any way on the mailing list, and over two weeks elapsed before any substantial
comments appeared on the list.
The meeting that took place on April 1, 1993 was fairly unpleasant. The
discussions were fairly acrimonious, but worse, there was little in the way of
substantive comment on the specifications. However, Steve Kent had provided
some comments on the specification on the list, and I used some of those
comments to produce a revised specification. A revised specification was in
fact published on April 20, 1993, almost three months prior to the IETF meeting
in April.
I was unable to attend the meeting that took place in Amsterdam on July 14,
1993. (A lot of other regular participants were also absent.) I was surprised
to learn from the minutes, however, that there were supposedly issues that
hadn't been dealt with in the April version of the specification. No such
issues were ever effectively communicated to me. I did not object to this
characterization in the minutes at the time, which was a serious mistake on my
part.
However, it was during the period around the Amsterdam meeting that the editing
of the document shifted again, this time to the folks at TIS. They wanted to
work on the underlying PEM mechanisms as well as the specifics of MIME
encapsulation, and I agreed that this was a good idea.
These revisions took a considerable amount of time, as did the shift of
responsibility (which did not go as smoothly as I would have liked -- it was
mostly my fault that it didn't) and the resulting specifications were quite
different from the previous versions. They were first released on October 15,
1993. This was over a month before the November IETF meeting in Houston. This
specification came out as an Internet Draft on October 27, 1993.
There was no little if any comment on the specifications at this point on the
list. There was considerable discussion of the specifications at the meeting,
and that discussion was fairly supportive of the overall approach (this is
reflected in the minutes).
The next revision of the specification came out on March 11, 1994. This was
almost three weeks before the March IETF meeting. Again, there was little if
any comment or discussion on the mailing list prior to that meeting.
The next IETF meeting took place on March 28, 1993. The only substantive
comment made concerned MIME canonicalization issues, which the authors agreed
to revise. (Again, this is straight from the minutes of the meeting.)
I don't have a record of when the next draft was released, other than it was
some time in June. In any case, another draft was released on July 21, 1994.
This was the only draft I'm aware of that came out in extremely close proxmity
to the actual meeting -- five days or thereabouts.
The next IETF meeing took place on July 26, 1994. The new documents were
discussed. A number of specific changes were requested in the document at this
time. There were some additional comments and exchanges on the list the
following month, but not many.
Revised specifications were again released on November 23, 1994. I note in
passing that for the first time there was a deadline by which documents had to
be submitted in order to be distributed for consideration at the subsequent
IETF meeting. The MIME/PEM specifications came in under that deadline.
The MIME/PEM specifications were approved by the Working Group at the December
6, 1994 IETF meeting. The documents then entered last call. And for the
first time, there was substantive discussion on the mailing list.
This is as accurate a cronology as I can arrive at from my records and archives
of this work. A number of conclusions can be drawn from this cronology:
(1) Steve Kent's assertion that these specifications were routinely issued just
prior to an IETF meeting is simply untrue.
(2) For whatever reason, serious Working Group discussion and perusal of the
MIME/PEM specifications has only occurred in conjunction with meetings, not
on the mailing list. This is an unusual model for Working Groups to adopt,
but nevertheless it seems to be the model this group has used up until
recently.
(3) The release of revisions of the documents has grown closer to actual
meetings over time. The reasons for this gradual change are probably
implicit in (2) above -- if nobody is going to say or do anything with the
documents before the meeting, why bother to release them until there's a
meeting where they can be discussed? This seems to be a natural outcome
of how this group functions. I'm not saying it is good or bad -- it is
neither -- but it is how things have been done here.
In conclusion, your childish attempt to somehow penalize the authors for not
producing specifications on your schedule is simply preposterous. Not only are
your assertions in this matter substantially at odds with reality, you are also
not empowered to enforce any judgements of this sort. Even assuming the authors
were unspeakably tardy in timely production of revisions, this doesn't
translate into any sort of forfeiture of the right to demand speedy resolution
of remaining issues in these specifications and advancement by the Working
Group to proposed standard.
The one right you do have as Working Group Chair is to demand the timely
release of documents, and if documents don't appear in a reasonable time you
have the right to ask for a change in editorship to effect timely release. (I
have used this right to great advantage in halding other groups.) You could
have done this in the past at any time, subject of course to being fair and
reasonable. However, you chose not to do so, and thus your right to criticize
the timing of these matters has for all intents and purposes disappeared.
In addition, it is the Working Group, not the authors, that you are penalizing
here. Speaking as an author I could not care less when these specifications
reach closure -- as long as there isn't anything on my to-do list to edit or
revise or whatever I'm a happy document author. It is as a Working Group member
and as an implementor that I desire closure, quite obviously along many other
non-author members of this group.
Ned