[R] Problems for 13 year old

[James Marca]

At that age, I recall trying to program a spaceship fighting game in
basic on the apple ][e (hopeless spaghetti that never ran, aside from
the "hyperspace" effect at the start), and later getting very involved
in programming robots in the game Robot War, which had a maximum
number of instructions and a limited command set, which required
programming in the pseudo assembly language to really milk the most
out of your code.

In short, I enjoyed programming games and moving pixels around.  

As to scientific computing, in a similar vein I can see a teenager
becoming interested in generating fractal images, getting chaotic
behavior from a simple iterated function, or simply plotting 3-d
surfaces of sines and cosines.  These all fall under "changing colored
bits on the screen as a result of some programmatic command."  Some of
the clustering code is also pretty cool, I recall toying with the EM
code and being pretty impressed by its ability to pick out similar
pieces in noisy data.  The inverse of clustering is of course,
generating noisy data, so that could become a game of sorts---try to
stump the clustering code by generating data from more than one
distribution.

I use a GPS a lot with my work, and put that data into R for
analysis.  Perhaps a handheld GPS unit can do the same thing, thus
offering a source of data collection to dump into R for plotting and
analysis.  If of course you have a gps unit.

James
At approximately Fri, Jan 24, 2003 at 08:20:20AM -0500, ggrothendieck at yifan.net wrote:
>       
> I would like to teach some scientific/statistical computing to my 13 
> year old nephew and was considering using R for this.  He has a Mac G3 
> OS 9.1.
> 
> I am looking for ideas for problems that would be interesting and 
> motivating for someone that age.  I recently taught him the basics of 
> HTML and noticed that he particularly was intrigued by the ability to 
> change colors; thus, perhaps problems that involve flashy color plots 
> would keep his attention.
> 
> Thanks for any ideas.
>