[CrackMonkey] [schoen@loyalty.org: Pill-weighing]

[Mr. Bad]

>>>>> "B" == Bad  <mr.bad at pigdog.org> writes:

    B> This gets mucked up if you have 100 or more bottles,
    B> though. Another way you could do it is for bottle X, put 100 *
    B> (10 ^ ((X - 1) * 3) pills on.  In other words, for bottle 1,
    B> put 100 pills on. For bottle 2, put 100,000 pills on. For
    B> bottle 3, put 100,000,000 pills on. Then you'll get a reading
    B> like 100101100. You can tell its the 2nd bottle that has the
    B> bad pills, because it's where the "101" shows up.

Actually, now that I look at it, that's kind of lazy. You can do it
with a lot fewer pills. For each bottle X, put on 10 ^ (X + 1)
pills. So, for bottle 1, put on 100 pills. For bottle 2, put on 1000
pills, for bottle 3, put on 10000 pills. If all the pills are honest,
you'd get a number like:

	1111100

But if there's a bunch of 101 pills in any of the bottles, you'll get
a number like:

	1121110

You can see that a "bad" bottle causes an extra value 2 places
down. In this example, bottles 2 and 5 are bad.

~Mr. Bad