This e-mail
- summarizes some of the responses/suggestions to using R for 13 year olds
from my query on this list
- discusses some additional investigation I did
- the approach I am currently using (which will likely incorporate your
suggestions
for R)
SUGGESTIONS
- find relationship between monthly heating bills and temperature. For data
see http://lwf.ncdc.noaa.gov/cgi-bin/res40.pl?page=climvisgsod.html
- generate fractal images, get chaotic behavior from a simple iterated
function, or simply plot 3-d surfaces of sines and cosines
- rep('Dad is a task master', 5000) and
paste(c('Joe','Sally','Roger'),'is not playing
with a full deck.')
- analyze Bush vs. Gore vote. See http://www.sas.upenn.edu/~baron/policy.html .
- rate autos (or other product) via several criteria and use stat methods to
combine these into a single desirability score
- spam filtering See http://www.paulgraham.com/spam.html and
http://www.mozilla.org/mailnews/spam.html .
LANGUAGES
Suggestions. I don't really want to get too sidetracked into the relative
merits of various languages here as the intention was to focus on science, math
and stats rather than programming, per se, however, for completeness some
people mentioned the following two languages:
- python (www.python.org). www.pygame.org allows you to quickly develop
games in python and I noticed www.pythoncard.org which in the spirit of
Apple's
HyperCard (which is a commercial product at hypercard.apple.com that could
also be considered)
- the kid-friendly lisp variant, logo (el.media.mit.edu/logo-foundation,
education.mit.edu/starlogo)
First Languages. Also my googling revealed that first year university courses
in programming often use:
- the squeak version of smalltalk (www.squeak.org),
- scheme (www.drscheme.org is one possibility of many) and
- java (java.sun.com , www.bluej.org).
RAD. On the theory that rapid application development (RAD) provides better
motivation, the commercial www.realbasic.com could be considered.
Other. Finally, I am not sure where this one fits but I also came across
www.toontalk.com which is a concurrent constraint language aimed at kids.
CURRENT APPROACH
My current approach is to start with HTML and then transition to Javascript
followed by R: HTML --> Javascript --> R
I think motivation is key since the exercise is pointless if the student lacks
interest.
HTML is pretty motivating since it does not take long to teach a basic tag
set and after that you can develop your own web pages which is empowering.
Moving to Javascriptis quite motivating since it allows one to put nifty dynamic
features into your web pages. By that time you are using programming and its a
natural transition from Javascript to R since the syntaxes of both are inspired
by C.
For teaching HTML, which is where I am now, I printed out the tag reference at
werbach.com and both source and web page image of the pages at
www.epcomm.com/webwiz . I am still deciding on how to proceed with Javascript
and the answers here will undoubtedly help me formulate my approach to R.
Variety. This approach means that we will have covered topics in:
- web design via HTML
- programming via Javascript, and
- scientific/math/stats via R
This should give him a broad introduction to three different areas from
which I can assess his relative interest and return to refocus on that area of
greatest appeal to him.
Thanks to those who responded.