Thanks Sandro. it seems good to me. it would be nice to be able to exaplain that although a properly functioning computer is always trustworthy, there are two kinds of behavior "against common sense" in this case: a) just nonlinear chaos in a map, or system of nonlinear equations and b) numerical integration error. b) sort of is "untrustworthy", although the computer is always correct in its computations strictly. It's something we watch for carefully. that's somehow different to chaos, which would yield the same answer (to whatever precision) if the equation set had an analytic solution. but I don't see how all that can be discussed in documentation for a turtle graphics package! so it seems fine to me. thanks. Colin On 3 March 2011 23:14, Sandro Tosi <sandro.tosi@gmail.com> wrote:

Hello Colin, thanks for your email.

On Thu, Jan 13, 2011 at 17:19, Colin Reveley <reveley@gmail.com> wrote:

...

here:

http://docs.python.org/library/turtle.html#demo-scripts

you say that chaotic behavior "proves that you cannot trust a computer's computations"

well, the opposite is the case. it proves you can trust computations absolutely. the computations are deterministic. they are correct. that's how the lorenz attractor was discovered.

you do need to know what a computer does (and to what) when it computes, and what it doesn't, and how to interpret it.

it's misleading to say that computation is untrustworthy. computation is a particular thing, that's all.

it might seem pedantic but I think it's a much more interesting and important thing for a student to think about than "untrustworthy".

which is just false anyway. as you know I'm sure. I'm not trying to be patronising.

In http://bugs.python.org/issue11392 I tried to rephrase the description, feel free to contribute on the issue for additional adjustment.

Regards, -- Sandro Tosi (aka morph, morpheus, matrixhasu) My website: http://matrixhasu.altervista.org/ Me at Debian: http://wiki.debian.org/SandroTosi