Also, when you figure out reentry, you get a supply of human-flavored jam and jelly.

Wouldn't the other limitation of a computer powerful enough to simulate all of the particles in a universe be that it would have to be as big or at least a significant fraction of the universe itself?

It's not clear that we know the answer to this question. In terms of processing speed, there's no requirement for simulating at full speed, so this is not an issue. In terms of precision, a bit- (or word-) serial approach can achieve any finite precision with merely a reduction in speed, so also not an issue. So the remaining questions are (a) is a simple finite precision Turing machine sufficient for simulating the universe and (b) how much space do we need for information? (a) comes down to a strong form of the Church-Turing thesis, which Is we're not sure; (b) is a function of both maximum density of information and the actual information in the universe (taking into account redundancy), which are closely linked (see also the holographic principle).

In summary, who knows?

If you're not good enough at arithmetic to understand that this isn't an issue, should you really be developing software?

s/no/new. My bad.

The student version is cheap (free at most decent universities). The Wolfram folk are great if you need a deviation on the license for student stuff (running on a multi-processor machine before multiple kernel executions were included in the default license); just ask. As a long-time student, Mathematica is the greatest tool out there, and is the only software out there where I'm consistently excited about no versions, and /always/ find ways to incorporate at least a few of the new features in my existing notebooks. With Mathematica 6, Manipulate[] was an absolute game changer. With Mathematica 7, I'm betting ParallelTable[] and the new charting features will be just as big a deal, for me.