Some misc. thoughts about the problem sets and submissions I saw.

I thought the first 4 problems would be easier than the last 2.
Unfortunately, I'm not sure my writing was all that great for problems
#2 and #3, which may have led some teams to think they were harder
than intended.



Problem 1: License Words
The biggest problem I saw was that some submissions were allowing numeric
(or other) characters into the words.  All non-alphabetic characters were to be ignored,
no matter what they were (punctuation, numbers, etc).

I should have double checked the test inputs.  (I grabbed a copy of
the GPL and an iTunes EULA as the inputs.)  They included some special
characters that I didn't anticipate being a problem.  My solutions
checked explicitly for alphabetic characters.  I didn't consider that
some solutions may check for characters to remove rather than
characters to include.  So some solutions were failing because they
allowed those special characters instead of removing them.  (I adapted
during the contest to allow some of these solutions, as long as they
did not print numbers in the output, which some did.)


Problem 6: Crop Circles

I've posted a PDF that shows a derivation for computing the area.  You
can also find info on computing the intersection areas online.  

There were differences between my output and some of the accepted
solutions.  The method I used to solve the problem follows the
derivation in the PDF, which is different than some of the popular
resources you'll find online.  It caused differences in one test case.
My method uses asin, while popular online resources use acos.  I
suspect the difference is numeric on certain angles, but I haven't
delved into it.  I did allow these submissions since the other test
cases were the same.