I discovered today, interestingly enough, that the exponentiation operator ("^") in the TI-83/84+ calculator series is actually left-associative: the expression a^b^c^d in fact evaluates to (((a^b)^c)^d), or abcd.  This is also the case with the older TI-36X II scientific calculator.

Ordinarily, one would expect the operator to be right-associative: a^b^c^d evaluates to (a^(b^(c^d))), or abcd, in common programming languages.

IMHO, this discrepancy could be considered a bug, but apparently the behavior has always varied from model to model.

Still, a part of me is left wondering how that little detail managed to escape my attention for several years.

The homework today for AP Statistics was to toss a coin a hundred times and then record the result.  Hmm ... what busy work (for a computer).

#!/usr/bin/perl -w
# Usage: ./$0 $sample_size $event_name $probability [$event_name $probability ..]
use strict;

my @tmp = @ARGV;
die "Not enough arguments" if scalar(@tmp) < 3;
my $total = shift @tmp;
my %events = @tmp;
my $sum = 0;
for (values %events) {
$sum += $_;
}
die "Sum of probabilities $sum != 1" unless $sum == 1;

my @keys = sort keys(%events);
for(my $count = 1; $count <= $total; $count++) {
my $result = rand;
my $lower = 0;
my $name = undef;
foreach (@keys) {
my $upper = $lower + $events{$_};
if($lower <= $result && $result < $upper) {
$name = $_;
last;
}
else {
$lower = $upper;
}
}
die "Out of range" unless defined $name;
print "${count}: ${name}\n";
}
exit 0;

In all: ./randsamp 100 heads 0.5 tails 0.5
The Perl script can work with a variable number of arguments, assuming that the first denotes the sample size and the rest are pairs of event names and probabilities.  Of course, the sum of the probabilities must exactly equal 1 or else the script dies.
Remember to adjust the shebang line to point to the path of your Perl installation.