The script works by matching a regular expression against itself (e.g.
$_ =~ /$_/, but this is slightly hidden inside the map for good measure).
During the matching, some parts are executed as code (hence the need for
5.005 and "use re qw(eval)"), while other parts are just there to provide
the correct letters which will eventually find their way towards the standard
output, and are skipped during matching. The code at the end removes some
characters from the accumulator ($^A) and executes them. The first time,
this is the string "format STDOUT=\n."; the second time, it is "write#"
followed by other stuff. When the write is eval()ed, the contents of the
accumulator are the string "The Perl Journal", so this is what you get
on output. It should be obvious how these strings get in $^A; but just
in case it isn't, the code in the middle of the regex will add one byte
to $^A, taking the byte from the regex itself ($_), using the ASCII code
of the n-th character in the regex as an index. Well, more or less, but
the code is so clear it doesn't need more explaining, does it?

