> What's going on in here?

Well, the code is, initially, trying to draw a Julia-set quadratic fractal,
a very well-known fractal which only exists in computers and does not model 
real-world phenomena. And the code is correct, if you take the algorithm as 
a pure mathematical object.

Problem is, the numbers it uses are way too large (and integers, which 
is true for every number on a computer), and the computation fails because 
they quickly exceed the memory allocated. So the code is trying to draw 
something which suggests infinity, but stumbles upon the material limitations 
of a computer. Surprisingly, though, the image which results, albeit noisy,
and completely dissimilar to the "healthy" behaviour of the algorithm, is an 
extreme case of aliasing which presents fascinating quasi-fractal patterns 
itself, for an unrelated reason: thanks to, rather than despite of, the 
offending material limitations.

> Why is this a "code poem"?

Because you can't just run the code and watch: you must know what it is trying 
to do and why it fails. The code itself is important. Of course, you couldn't 
just read the code or about the code, either, you also have to run it. This is 
a consequence of the triple nature of code, as descriptive of a process to a 
human reader, as a set of instructions for a machine, and as an executable
application with an output for a user. The three are implicated here.

Note that this is not text which appropriates elements of code, but actual code, 
as I want to show how code can signify through strategies of code - a real "code 
poetics", as opposed to code processing text or text appropriating elements of 
code. And it's run-of-the-mill coding at that: do not look for anything fancy 
or elegant in it.

> And what does all this mean?

In media art and the digital arts it is usual for the artist to explain his
work and the significance of it; in literature, it is expected that the author
will remain silent and let the reader interpret his work. I align myself more 
with the latter tradition.

> Why exactly does that pattern emerge from the error?

I am working on this one (really!), when I feel I have the math straightened 
out I will update it here.

> What about the second version?

The error and the images it generates have elicited some interest (and I
am intrigued myself), so I have decided to put up an applet that lets one 
explore the images better. I have found that the more interesting images 
appear by the periphery of the "fractal."