> ... If networking is
> political in nature, i.e., in pursuit of personal goals, then I still
> think it is nonvalue-added. However, if it is aimed at developing the
> relationships across the organization needed to get work done, then I
> think it may be good, even very good. In fact, it is consistent with the
> idea of self-organizing systems.
The main thing that bothers me about these discussions is that they
are so fuzzy and unconcrete. I rather like the part of "Fifth
Dimension" that I have read, but it pales by comparison with the
stark utility of Tom Gilb's "Principles of Software Engineering
Management". The software part is really irrelevent.
When you talk of meetings, you don't classify them very closely.
I tend to characterize them by what their goals are. It was in
contemplating the goals of a meeting that I was drawn back to Gilb.
He only makes three major points in his book, but the central one
is that in any project, we have multiple, critical goals. Only when
you recognize them and treat them explicitly do you stand any real
chance of accomplishing what you intend.
Tom Gilb shows ways to confront the multiplicity and ways to be
explicit and measurable about the goals. Does anybody else?
Invention of good goals is quite an art. Kennedy's man-on-the-moon
goal is often held up as one to be emulated. I am even more pleased
by the goal set for the Macintosh: The new owner of a Macintosh shall,
in the first hour, come to feel in charge of the machine (rather than
dominated by it). Both worked pretty well.
A second point of Gilb is that highly structured meetings, such as
work-product inspections, with formal role structures and written
guidelines appropriate to the content being inspected, can be both
highly and measurably effective. It's so hard to train folks to do
this kind of meeting well that Gilb has devoted another entire book
("Software Inspection") to that process. Despite the word software
appearing again, the single most effective place to apply inspection
is in the first elaborations of the one-pager that the CEO blessed.
Any wrong, missing or extra function at the very beginning of a
project will lead to massive waste unless it is caught and corrected
at an early stage. The meetings to accomplish this are a natural
point of study in applying the principles of statistical quality
control to intellectual work. Gilb takes his lead from Fagin at IBM.
Is this making any sense?
Dick
-- Dick Karpinski Special Projects Voice: (415) 476-4529 (11-7) Email: dick@itsa.ucsf.edu FAX: (415) 476-5523 Org: Information Technology Services, UCSF Home: (510) 658-6803 WWW: http://www.ucsf.edu/~dick Messages: (510) 658-3797 USPS: Box 0704 UCSF, San Francisco, CA 94143-0704