jacobian; 14df

[From Rick Marken (930310.0800)]

Avery.Andrews (930310.2036) --

It might be an idea to wait a bit till Bill's 14df project is more
consolidated - I think there are a lot of practical lessons in the
application of PCT to complicated situations there (e.g. why Bill's
design works so much better than `naive cartesian transpose

Jacobianism'), and it might make a better impression on any new recruits

to have a reasonably extensive & coherent writeup of this.

I heartily disagree; my experience is that you just can't tell
what will make an impression on new "recruits". It is very likely
that when the 14df project is finished, the nay sayers will simply
yawn and explain how it "could be done" with the appropriate output
generation model. Look at what happened to Bill's ARM demo; you may
think this is because there were not enough df in that model; but
when there are plenty of df in the model the "ho hums" will be
based on something else. I've gone through this for years with my
attempts to produce convincing demos and experimental tests of PCT;
I try to anticipate every complaint and misubderstanding -- and then
when I submit the paper or show the demo there is always some completely
unanticipated kvetch. The "number of df" complaint about the ARM demo
is just a roose -- anyone who understands the principles of the
HPCT model would know that it works with an arbitrarily large number
of relatively orthogonal df (see my spreadsheet model, for example) in
a highly nonlinear (even slightly non-monotonic) enviroment.

So the 14 df model may mean a lot to you (and me) -- and I think
it's a wonderful project and demonstration. But if we (PCT) had to
wait for every demo that we thought would "clear up" things for
the uninitiated (and usually unconvincable) we would wait forever
to publicize PCT. Believe me, supporters of other theories -- with
FAR LESS (how about virtually zero) experimental and working model
support -- have absolutely no hesitation about touting the virtues
of their brilliant models. PCT does not have to be so shy -- especially
given its remarkably strong experimental and modeling base.

I believe that the ad still could be improved a bit (I'll try to
suggest some changes if I get a chance today or this evening). I
suggest waiting until the end of the week (Sunday) for all
input on the ad and then shoot for posting the final version of
the ad sometime next week (3/15).