Java is just a four letter word

[From Rick Marken (970626.0830 PDT)]

Bill Powers (970625.1754 MDT) --

Java seems designed mainly to get flashy effects on the screen,
at least as I've seen it used.

This is not really true. Java is very much like C++; it's a regular
programming language. It just has some capabilities (threading,
network related stuff) that are relevant to network operations.
But, in principle, if you can write it in Pascal, C, or C++ you can
write it in Java. I believe that in the near future many
applications programs (word processors, spreadsheets, etc) will
be available in Java. I believe that the difficulties I am having
with Java are not due to Java itself but, rather, to the fact that these
programs are being run via a browser in a networked
environment. When you run Java code as a separate application it's
much more stable (in terms of timing and stuff).

I think it's [Java] worth pursuing further, too. But I really
don't want to see PCT exposed to the world looking kludgy and
klunky, with nothing working really well, and lots of irritating
difficulties of use. We can be forgiving of each others' interim
results, but other people aren't going to be so generous.

I've run my Java demos on many different platforms in many
different venues (I try them wherever there is a computer
connected to the Net). My experience is that all of the demos
have worked rather well wherever I have tried them. They don't
seem kludgy or clunky to me. I have encountered difficulties
(such as networks that have protection and won't let the Applets
download). But I have never had difficulties once an Applets was
loaded. The basic demos may not be perfect, in some sense, but they
certainly work in the most important ways. The "Nature of Control"
demo, for example, shows very clearly and consistently that, when there
is control, the correlation between invisible disturbance
and mouse is .99 while that between visible cursor and mouse is
near 0.0. The "Levels" demo shows that the subject and model
runaway after reversal looks nearly the same.

I have never had the experience of people getting a bad impression
of my demos (and, by implication, of PCT) because they were kludgy,
klunky or difficult. In fact, the only criticism of the demos that
I've gotten had nothing to do with how well they worked; it had to
do with what one of the demos (the "Selection of consequences"
demo, which works fine technically) claimed to show. I agree that
any demos of PCT that we make available to the public should be as
non-kludgy, non-clunky and difficulty-free as possible. But I
think that the criticisms we get of our demos have and will always
have more to do with what those demos _show_ than with things like
how many samples of data are being collected per sec.

I wasn't referring to the lag, but to the trick Bruce Gregory
found for continuing to track by using the mouse arrow instead
of the cursor that the computer draws. This is the cheat that
renders the proportional case meaningless.

Since I can't make the mouse arrow disappear, I considered the
"Proportional" control case to be relatively meaningless anyway.
I'll probably just eliminate it from the demo. The paydirt (in terms of
discriminating model-based from control of perception models of
purposeful behavior) is the "Integral" control condition.

Have you tried this to see what happens?

Sure. The RMS errors I get in the closed and open case using just the
arrow are just about the same as those when I use the cursor in the
closed and the arrow in the open. For me, the main determinant of the
size of tehe RMS error is the nature of the disturbance. To get a
new disturbance you have to leave the demo page and return to it;
"reload" doesn't do it, apparently.

Net browsers have the ability to use "helper" programs while on
line. Would it be possible to write one that downloads source-code,
compiles and runs it, and then returns control to Netscape?

I think you would essentially be reinventing Java if you did this.

Alternative: A program like Wolfgang Zocher's "Simcon" allows one
to send very brief scripts in ASCII, which are then interpreted

This sounds good, too. But I think Java itself might be a good
medium for something like Simcon. Remember, the main problem with
Java is real-time _interactive_ applications. A JIT compiled
version of Java could probably run and animate the output of a
Simcon model at close to the same rate as a version of Simcon
running in native compiled C.




Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: