[
Lists Home |
Date Index |
Thread Index
]
On May 1, 2004, at 7:24 PM, Bob Wyman wrote:
> Currently, the standards business seems to rely a great deal
> on face-to-face meetings in addition to mailing lists and telephone
> conferencing. What might make sense would be to do a concerted effort
> to either develop or select technology that could be used for hosting
> remote discussions and working sessions over the Internet.
Great ideas in this post! I have thought about this over the years
quite a bit but don't have real success stories to point to. A few
observations:
- The problem with any flavor of voice or audio conferencing is that it
is intrinsically synchronous. It's all very well and good to open up
that meeting in Tokyo or Paris to remote participants, but the time
differences are still a killer to real participation by people 9
timezones away. There's got to be some asynchronous component to
capture proposals, allow voting (formal or informal) over a 12-24 hour
period, and get the decision recorded in a useful form.
- Email discussions are asynchronous and could be managed by a
well-organized WG to do all that; the problem is is that email threads
INEVITABLY drift into tangents. In a F2F meeting, the Chair can
presumably cut them off, but if the Chair is asleep in Tokyo while the
flame war rages across the Atlantic, that doesn't work so well.
- In principle, a Wiki-like mechanism could be both focused and
asynchronous -- Comments that are in an inappropriate context can be
deleted or moved, and the norm of having a piece of text that needs
improving (rather than a vague thought that needs to be elaborated,
critiqued, defended, and mutated in an email thread) should focus
discussions in a productive way. The Atom people tried this, with
something between limited success and qualified failure, so I'm not
sure what the lesson from it is. In either case, a group with an
agreed upon charter, a Chair, and a voting system could do things
differently than Atom did/does, so the lesson may not apply.
>
> Other than the "travel" problem, what other significant
> problems exist in the standards world that can be addressed with
> technology?
>
- If the problem ain't the technology, so the solution is unlikely to
be technological. The biggest problem with the "standards process" is
that it's usually an unholy mess of problems chasing solutions,
solutions chasing problems, and both open and hidden agendas being
pursued by all. [1] If you miss Gosling's Window of Standardization
[2], you might as well stay home rather than engage in any of the usual
alternatives, such as open warfare (cf the SOAP reliable messaging
stuff); non-interoperable specs full of "logrolling" deals ("I'll
support your proprietary thing if you support my proprietary thing,
accepting that we'll never implement the others') (cf SQL 3); and specs
written by WG's that inhabit some alternative universe, decoupled in
space and time from the real world (cf more SGML and XML specs than I
can bear to think about).
[1] That's a paraphrase of Michael Cohen et al's Garbage Can Model of
Organizational Choice, a 30-year old paper that is still apparently the
inspiration for fruitful research. My actual degrees are in history and
political science; I had NO IDEA when I gave that up for nerd-dom that
Thucidides, Kautilya, Machiavelli, and von Clausewitz would continue to
be highly relevant to my professional life. Then again, academic
politics is even uglier than standards politics, with even lower
stakes.
[2] Sadly "Phase Relationships in the Standardization Process" doesn't
seem to be available on the Web anymore;
http://java.sun.com/people/jag/StandardsPhases/index.html redirects to
Gosling's Java.net blog (Google's cache too). Oddly enough, my quote
from it at http://lists.xml.org/archives/xml-dev/200204/msg00419.html
is Google's second hit on the title. I guess I'm the article's
biggest fan!
|
