This is a pretty picture of some source code! Isn't it nice-looking?
One really neat thing to do is to convert this image into a raw PNM (using pngtopnm, in the netpbm library, or the pnmtopng Debian package - other programs put in comment headers, and that would be Bad), and then type the source code in, and compile the sourcecode and then run it with the PPM as input and, say, pipe it through tar xzvf. For example:
gcc source.c pngtopnm < fluffy.png | ./a.out | decrypt - see note | tar xzvf -
You can do this with other raw-type PNM files too, using this crappy source, though you need to make sure your
input PNM is in raw (rather than ASCII) format, has no comment header (if
it has one, it can be stripped easily by doing something like
pnmtopng < myfile.ppm | pngtopnm > strippedfile.ppm, or the
truly daring can just use vi) and that it's got at least 8 times the
number of image bytes as the data file (i.e. a 12345-byte datafile needs
at least a 98760-byte image, such as a 200x165 24bit image).
Note: Under the DMCA, ROT13 is considered "use protection," and so the output of this program has been double-encrypted through ROT13 so as to protect its output. Replace the "decrypt - see note" in the command above with some means of decrypting double-ROT13 (also known as ROT26), in order to make the output usable.
One nice NCSU student who was bored with his Scheme homework decided to improve the source code, so I made up a nicer-looking image with his nicer-looking source code. Below is the original one (which doesn't look as nice).
