I suppose, that this is a misunderstanding. In the logo world of computing with beginners, this means to decompose somewhat complex tastks in tiny peaces and write functions for those, wich then can be assembled to solve the complex task. This is the ordinary meaning of bottum-up development.
OK, so like test-driven development. Write a test. Write a function that passes the test. Get a bunch of little pieces that work well. Then (and only then) start using these pieces to pass other tests (more complicated tasks). In the world of structured programming, this meant: def f: def g: def h: def main: f; g; h But today it might be more in terms of subclassing generic classes i.e. in place of main call sequences, we have more specialized forms of composed and inherited object.
It worked very well for our students, and the fact that turtle graphics is a toy is important: there's less anxiety playing with
Right! But that doesn't meen that the toy should be poor. IMO it should offer a rich set of means to solve different tasks. Or - as another example, concerning turtle graphics - it should have good means for animationa as are, for example, possibilities to set the speed of the turtle movements or to set different turtle-shapes, maybe even gifs. This makes programming much more interesting for learners.
Those weren't my words you were agreeing with, but I guess you knew. I guess I don't see Python as currently very strong in the turtle graphics department. Maybe someday. Maybe even soon. I like spatial turtles, i.e. turtles that swim in XYZ, not just in XY. I played with doing that using Python + POV-Ray a long time ago, i.e. do your movements, then publish a rendering of where you've been. But that's all pretty esoteric. Too advanced for most beginners. http://www.4dsolutions.net/ocn/numeracy3.html (see 'mywalk.pov' -- ironically planar)
By "toy" I meant something more like "broken toy" -- thinking specifically of turtle.py on Windows.
There are easy ways to overcome this and they are still mentioned several times on this list. (It's not necessary to blame windows)
You mean with the -n switch? Why not just use some other language that's better at Turtles than Python? is my main question. If you're going to use Turtles as your primary pedagogy, I suggest Logo or some more sophisticated 3D turtle implementation (there's a 3D Logo no?). Python doesn't have a strong literature. Logo does. This may change in future, but for now, that's just the way it is. If you forced me to use Turtles today, I'd use Logo.
There are a few important arguments why not to underestimate the importance of a good turtle module in the Python distribution (out of the box as vern ceder says).
1. Many of us are trying to further the use of Python as a first programming language. For them and for their students this would make a great advantage.
Again, I don't think turtle graphics is an essential first step. But IF it is, then why not use something other than Python? Like Logo?
2. At least in Europe there ar many teachers who have a good Logo-experience. But now Logo seems to be largely out of date and many are searching for a substitute.
Why is Logo out of date? Use it for 3 classes (?) to set the stage, get through with the turtle stuff, then turn to Python or something else. This idea that we should pick just one language and use it for everything, regardless, is to be avoided. That's not the real world. As Python has an outstanding clear syntax (imho much clearer than
Logo, when it comes to programs beyond turtle-graphics), Python would be a great cadidate. Had it a good turtle module, Python were an even more seductive offer to all those teachers.
I don't disagree with this. I just don't see it as a high priority. I'm fine with the idea of using Logo for awhile, moving to Python when the students are more sophisticated and don't need a turtle to motivate their interest. Many students I work with wouldn't agree to use a turtle. That's something for little kids. They're more hungry to become adults. 3. Compared to its presence in the software world *and* compared
to its adequacy as a teaching language, the presence of Python in the educational area is relative is reatively weak.
Which doesn't concern me, per above. If Python *never* had a strong turtle graphics capability, I would shed no tears. But if it does, so much the better. I have *no* reason to get in your way on improving Python's turtle abilities. I'm just not seeing these efforts as critical to my own plans i.e. as we say in idiomatic language "I'm not holding my breath". So (many different) things should be done to improve the situation.
One of these could definitely be the construction of a better turtle module, which has apparently been started by now. That's a good thing.
Everybody hwo wants (and has time to do so ;-( ), should support it. Those who prefer different ways, will (an already do) support those, fortunately.
Regards, Gregor
Yes, I think we're in agreement. I plan to stop saying anything on this topic, as I'm afraid I will be cast as "anti- turtle". For political reasons, I want Seymour Papert to like me. Alan Kay too. I hope that's OK with Arthur (whom I like a great deal). Kirby