🚀 go-pugleaf

Brian Harvey, in response to a discussion on visual programming environments,
says:
>I *still* hate GUIs.  Let's see you do "rm *.o" on a Macintoy.  They trade off
>making certain things really simple against making *everything but* those few
>things completely impossible.

Interesting comment.  He's right.  But the important thing about it is how much
it reveals about how he thinks.  Go back and read C. P. Snows lectures on "The
Two Cultures" and then review the research on left-brain/right-brain dominance.
People really do think differently.  And it's interesting to engage them in a
conversation on a test topic (my litmus test used to be: What do you think 
about Macs vs. PCs? -- Now a bit dated by the arrival of Windows), and then
rank them on the left-brain/right-brain scale.

But here's the important point:  It's not where you're centered on that scale,
it's how broadly based you are.  How many different points of view can you 
handle.  Alan Kay and Marvin Minsky are right.

Back to the issue of Logo futures.  Here the real issue is NOT whether visual
or analytic/text-based environments are better, it's: How do we integrate them?
As Peter Wegner(?), the physicist/mathematician said: All the interesting stuff
is in the cracks.  It's the interdisciplinary areas where the real payoffs are.

So I'm very interested in attempts to integrate the visual and (text) language
based paradigms.  Here are some comments on a few attempts thus far:
* Logo: A wonderful, conceptually clean programming language which can be used
	to produce programs which are often wonderfully visual and spatially
	interactive applications.  But what about visual programming?  Visual
	traces of program execution? What about encapsulation and abstraction
	of state and behavior, i.e., objects?
* HyperCard, the "programming environment for the masses": What a joke. A 
	brain-dead, trivial programming language (I probably wouldn't have
	given one of my MS students a passing grade on a language design
	project for it.)  As Ward Cunningham said: Yeh, it's "object-oriented,"
	but you don't get to pick the objects.  BUT, it does have the
	beginnings of an interesting interaction betweeen visual and text-based
	programming environments.  If we could only have HyperCard with a real
	MOP (meta object protocol) and a real programming language.
* Prograph: As noted here before, this is a VERY interesting visual programming
	environment.  I'm interested in it as much for its data flow approach
	to programming, as its visual aspects.  Because it's so different, it
	helps stretch the mind and force you to take that different point of
	view.

So as a system builder/former physicist/father, I'd be very interested in an
environment that allows me to encapsulate and deal with many levels of
abstraction, to experiment and iterate (which probably means "dynamic 
languages"), and to both program and interact with the results in both visual
and text-based modes.  I want a tool that takes advantage of both sides of
my brain.

Comments?

Mike Wirth
mcw@cs.rice.edu
(a guest at Rice.  "my MS students" refers to a former life of mine elsewhere)