BEST Best Practices of the IETF
Christopher Allen
ChristopherA at AlacrityManagement.com
Sat Mar 22 11:11:21 CET 2003
There was insufficient time at the Problem-Statement WG in San
Francisco for my presentation on BEST Best Practices of the IETF to be
on the agenda. So instead, here is the information from the results of
my informal survey asked in
http://www.alvestrand.no/pipermail/problem-statement/2003-March/001016.html
so far.
-- Christopher Allen
----------------------------------------------------------------------
.. Christopher Allen Alacrity Management ..
.. <ChristopherA at AlacrityManagement.com> 1563 Solano Ave. #353 ..
.. o510/649-4030 f510/649-4034 Berkeley, CA 94707 ..
----
Problem-Statement WG
BEST Best Practices of the IETF
Christopher Allen <ChristopherA at AlacrityManagement.com>
----
The Problem-Statement WG Charter says:
"In advance of trying to change the IETF procedures and rules to
deal with these problems, the IETF should have a clear, agreed
description of what problems we are trying to solve."
----
I believe that we should add to the WG Charter...
"... and we should have a clear understanding of existing
procedures and rules which are effective, so that we may preserve
them."
For we don't want to throw the baby out with the bath water.
----
First Step...
What Works?
----
Questions Asked...
o What is "Best" about the IETF?
o What do I get out of the IETF that is the most valuable to me?
o What "ideals" does the IETF attempt to exemplify that I'd like
not to be lost?
o What excites me about participating in the IETF?
o What makes IETF better then other standards organizations?
o What part of the IETF "process" do I feel is the key to why the
IETF works?
o What one "story" from my experience with the IETF exemplifies
what is best about the IETF?
----
Some early answers
o Mailed to problem-statement list, ietf list, and handouts at
IETF-56 (San Francisco)
o Dozen responses first day
o Most answered all questions
o Many are similar
o Some are interestingly unique
o Need more responses
----
We Have Some Shared Values!
One thing that is clear from early responses is that there are some
things we agree on.
----
Culture & People
o Almost every response included aspects of culture and personal
values...
o "People with clue working toward the common goal of
a better internet."
o "Working in a group of world-class engineers on
innovations that do change the world."
o "Humming"
o "The hacker spirit (no suits, please!)"
----
Individual Work
o Most mentioned they value the "individual" nature of the IETF
work...
o "An organization of individuals, NOT companies."
o "An organization which judges people and ideas on their
merits, not affiliation."
o "The ability for any individual, regardless of affiliation,
to submit drafts individually, and participate in
working groups of interest."
----
Standards
o Of course, our Standards themselves are mentioned by most...
o "Easy publishing of proposals."
o "Free access to RFCs."
o "Care for the Internet."
o "An operational Internet."
----
Rough Consensus & Working Code
o Our unofficial motto is mentioned often...
o " 'Running code and rough consensus' - particularly
'running code' and interoperability testing as a
requirement for advancement to Standard (and Draft
Standard)."
o "Running code. If we don't have that, we'll just be
professional out-of-touch standardizers who make plans
for complicated architectures."
----
Openness
o Many people point out our ideal of openness...
o "Open standards"
o "Openness in general"
o "Enough openness in resulting products to tend to minimize
customer dependence on proprietary lock-ins."
o "Openness, in several respects. (But without too much loss of
focus.)"
----
Process
o Half the the responses point out various part of our process...
o " No final decisions at meetings, all decisions must be
confirmed on list."
o "Organization that to at least some degree is able to
accommodate active participants who can't always attend
meetings."
o " No need to attend meetings."
o "[unlike other standards organizations, we don't] come out with
a first version of a 'standard' that hardly supports
interoperability and then requires several 'versions' to be
'standardized' before products interoperate.
o "We have tended to avoid imposition of undue levels of process
and procedures."
o "IESG quality control, or, more generally, meritocracy."
----
Interoperability
o Some people point specifically to our emphasis on interoperability...
o "Considering which features of a specification have actually
been implemented and shown to work interoperably is
important."
o "I know that a number of IETF specifications such as MIBs seem
to require somewhat different "interoperability testing"
than protocols, but it's good that the IETF seems to figure
this out and modify procedures accordingly."
o "We have done some interop testing during the IETF but I think
that is a bit extraordinary."
----
Technology Pragmatism
o Some people mention our technology focus...
o "The quest for the *right* solution (within the constraints of
the real world)."
o "The IETF community has strived to create an environment where
standards are evaluated based on technical merit, not
corporate sponsorship or politics."
o "Someone may disagree with an IESG decision, but everyone would
likely still agree that the decision had been made on
technical grounds."
o "A lot of cross-pollination across the protocols for the
Internet."
----
Quality
o Some mention the quality of our standards...
o "The ability to generate technically excellent standards under
a serious quality control regime."
o "Quality control, integration/close collaboration between the
various areas."
----
Misc.
o There were a few items that were only mentioned once, but are
deserving of mention...
o "The IETF preserves a place for "creative ambiguity" in the
specs."
o "Focusing (when appropriate, and usually) on scalability."
o "Be lenient in what you accept and strict on what you
transmit."
o "The End-to-End principle."
o An interesting point is that these were only mentioned once, thus if
these are important, they are not important to many, or we have
not taught these values very effectively.
----
What's Next?
o How do we rate if a process or shared value is "effective"?
o How do we eliminate self-selection bias?
o Does the Problem-Statement WG find this type of information useful?
o Should the charter be amended to add this as a work item?
o Meanwhile, you may send more responses to:
<ChristopherA at AlacrityManagement.com>
----
More information about the Problem-statement
mailing list