[Edu-sig] The keyhole problem and learning environments

kirby urner kirby.urner at gmail.com
Fri Jul 14 01:15:06 CEST 2006


On 7/13/06, Andre Roberge <andre.roberge at gmail.com> wrote:

> This is the philosophy that was behind rur-ple.  Users starts writing
> simple programs in the robot world, with very few instructions.
> However, unlike with Guido van Robot which uses a Python-like syntax,
> rur-ple uses Python through and through ... thus no "learning step"
> required when going from the robot world into the full python world.

This is my own preference, that if we're writing in Python, we just go
with Python syntax, and not try to use Python as a parser-interpreter
of yet another language (even though Python can do that -- I think it
was Danny Yoo who'd implemented Scheme in Python).

Per my excerpt from a letter to Jason (TSF), Alan relayed Seymour
Papert's feeling OK about "exposing the receiver" or some such jargon,
by which he meant objectifying the turtle as the message receiver.

In the original LOGO you'll recall, there was never ambiguity about
which "self" was meant, as the FD 30 was more like "inside a turtle's
head" (and so no mystery, like "it's me, duh").

But by objectfying each turtle instance, by constructing from a Turtle
class, we get this greater freedom of multiple turtles.  Plus why just
turtles?  Robots, entire 2D ecosystems emerge.  Pandas.

As to the "keyhole problem" -- that's almost to general a metaphor to
be useful.  Like, *of course* we don't want unnecessarily limiting
assumptions built-in by mistake, but until we're countered with true
counter-examples to our assumptions, we just won't think of them.
That's what the word assumption means.

Put another way, I think we should focus on the capabilities and
features of actually running projects, not hypothetical "could be"
software.  That's one aspect of this group that I like:  most of us
have running code to bring to the table.  Keeps things more concrete
that way (so many ivory tower lists I lurk on have no makers thereon
-- people who actually make, don't just wax eloquent about this that
and the other).

Kirby


More information about the Edu-sig mailing list