[Twilight Suzuka]

You are wrong about the sizes of bytes, nybbles, and word sizes (how much a CPU can process/store at one time).

The size of a byte is non-standardized. Many archetectures use 4 bit bytes (stores one hexidecimal value) and 6 bit bytes (makes it easier to do some operations). The nybble, which is based on the byte, is also non-standardized.

Any code relying on the byte being 8 bits isn't portable to CPU's that use 4/6 bit bytes, which a lot of smaller appliances (cell phones for instance) and such use.

On the topic of word sizes (a word is how big a register is in the CPU, and in effect, the largest peice of data that the CPU can process in one OP code), you are mistaken, though your assumption is not nessasarily a bad one. A 32-bit processor can work on more than 32-bits at a time; it just does it in 32-bit chunks. The defining quantity than, is the size of the registers, which are generally the same size as the processor, but which can be larger or smaller.

For instance, some of the game consoles like Saturn, N64, and PSX had variable word sizes, and could thus process big floating points with as much ease, though a little less speed, than if they had a bigger, more expensive processor size.

No offense or anything; the assumption that a byte is 8 bits is pretty much standard with PC's, and the other part wasn't too far off. Just technicalities really, nice job.