BPSS - general comments, finalization process

[James Bryce Clark <jbc@lawyer.com>]

I received Karsten's revised BPSS draft v1.00A yesterday (4-23) and am
reviewing it.   We are scheduled to discuss some aspects of the draft, and
other open issues, in this morning's conference call.   I will send
additional messages this morning on other issues,  However, as a general
matter, please note that we will NOT be able to complete our group review
today.  

I am pleased with the BPSS generally.   I am not pleased with the process,
and apologize for having to point out its shortcomings.  However, I need to
mention these so as to ask for, and advocate, process rules that help us
finalize it (this week) in a manner that preserves its useful content, and
reduces the chances that it will be gutted like a fish by our continuing
process flaws.   

THE PROPOSED SOLUTION

1.  In view of Paul's report that our actual press date is Friday, I
suggest we absolutely DISCARD any new suggestions of ANY KIND for any model
changes, as of the close of today's morning conference call, and confine
ourselves to text refinement.   Should have happened earlier, but we can at
least do so now.  

2.   Just in case there is any doubt, I think we need another conference
call Thursday (same as slot as before) in order to work any final text
issues that need attention.  As a group member I am not able to confidently
have additional textual changes made by any single author, without any
opportunity for group discussion.  

THE PROBLEM
(This is optional reading, for more morbid readers who care to know why I
am making the above suggestions.)  

1.     Most turns of the draft since January have embedded multiple,
complicated, and unmarked changes to fundamental assumptions of the
document.   I expect this one to be no different.   

2.    Observers, including casual participants and QRT, might assume that
since January we have been narrowing the open issues and converging on a
final form.   Instead, with each draft and each conference call, we re-open
and re-visit basic questions such as the use of documents, the use of roles
and UMM compliance.  

3.    As a team, the metamodel group is frequently unable to log, accept
and enforce decisions in any reliable or evenhanded way.   Our design churn
has been severe.  Other groups who assume and rely on the relative
stability of our work,  including the ccbp-analysis authors, QRT and the TP
team, have been whipsawed and confused by its continual change -- and the
varying interpretations of it they receive from one or more self-serving
advocates of a position -- and bad text apparently caused by reluctant
drafting from holders of minority views which half-heartedly or ambiguously
characterize a majority consensus.   None of this means to take away from
the fact that some of us -- particularly Karsten, Cory, Kurt, Brian, Nita,
Bob, Jim and Paul -- have been working like dogs, without pay, to do this
work.  The BPSS reflects a great deal of hard work and expert effort.
Unfortunately, where we can't agree, it also reflects a muddled vagarie
that tips the model in the direction of the bias of whichever writer last
gets their hands on it.    Thus, different sections tip in different ways.  

4.    In addition to losing and confusing our EXTERNAL constituencies, the
speed and opacity of design churn has discouraged INTERNAL participation
from anyone who can't be a full time metamodel analyst.  At least five
members -- all bright folks, with varying views -- have told me they've
effectively dropped out because there is too much new stuff sneaking into
each draft, and they can't possible follow it all effectively.    For my
own part, the level of review that is required to parse about 50 pages of
text and review the disposition of about 120 comments is not small.  Last
time I did that it took about two days -- during which I did little else.
This was in connection with the v0.90 draft, and set out in the attachment
to my March 6, 2001 listserv comments.   I am doing it again now.  But even
if I have no life, no career and no sleep, I can't do it in the 23 hours
between yesterday's 1.00A distribution and this morning's call.  

Thus my request that we stop futzing around with the model.   Sorry for the
tone.  That's what I get for writing at 5 am.  

Jamie