Well, even though I don't like "obfuscation" schemes that
include calls to crypt, unpack or chr,
I have to admit I don't know the ASCII table by heart. :-)
So, I noticed all those 44s, decided I could at least
change the eval for a print, and see what it
actually did (mind you, I didn't guess... but the way to
do it is still, er, poor).
If the main point in an obfuscation is based on s///, crypt,
or chr to make it hard to understand, you can be sure I
won't like it. Try to do the same thing in a completely unnatural
and complicated way, and there you will have me try to understand
what it does, parsing it by hand, and so on...