Okay, what it does is this: in a sed regular expression, you
can in fact use any character as an inter field separator -
up to and including alphabetic characters. which means:

saMooaBinga and s/Moo/Bing/

are semantically the same. So, what this means is that english
words can, in fact, be valid sed regular expressions - words like,
say, savanna (s/v/nn/) and segregate (s/gr/gat/). Which is nice.

So what if these word regexps are then applied to other words? Can
we generate valid words by passing a valid word into a word regexp?
The answer is yes, we can. And sick.pl finds them :)

You might like to just look at sickwords.txt, as sick.pl can take
quite a long time to run.