[NCLUG] Python Rox

[Stephen Warren]

Chad Perrin wrote:
> On Wed, Oct 15, 2008 at 06:11:22PM -0600, Stephen Warren wrote:
>> What I might suggest is somebody writing a very simple-to-use GUI
>> infrastructure that kids then plug the algorithms into. For example,
>> write a graphics toolkit that creates a window, draws lines, changes
>> line colors etc. with just 3 simple calls. Then, you could use this to
>> teach the children LOGO[1], but with Python syntax. (i.e. teach them to
>> write algorithms to draw square, house, plant, ... but not have to know
>> anything about windows, widgets, graphics contexts, ...)
> 
> What you're talking about is technically "turtle graphics", not "Logo".

Yes, I was being imprecise in order to be brief.

Turtle graphics and Logo are however very closely entwined; Logo and
turtle graphics were developed together as a coherent entity for
teaching purposes.

> Logo is just a language -- in fact, it's an M-expression variant of LISP
> (but not a variant of Common Lisp, really, for those who know the
> difference).  UCBLogo even sports macro capability.  Logo itself is no
> more a "toy" than other Lisp languages, in general.  Turtle graphics is
> kind of a framework for educational use, and that *is* a "toy".  One
> could conceivably create a turtle graphics framework for any language,
> but Logo is where it first appeared, so it's associated with Logo in the
> minds of those who know of its existence.

I'm not sure what the your comments are rebutting; I don't recall
equating Logo/turtle with being a toy, but perhaps I did.

That said, all the Logo implementations I've used didn't really have any
IO beyond turtle graphics, so it's not that unrealistic to call those
implementations a toy. I agree that the Logo language itself is probably
Turing complete, and as useful as any other language, given suitable
libraries/... for system interaction, data access, etc.