No, I don't think that's a stupid way to do it. In fact, it could have some advantages. For one, your not limited by a certain number of bits 8,16,32 Your only limited by the number of digits you can fit in the array. I'll bet the divide routine's going to be a pain in the butt though.

Result Var Word
... and later...
Result = Result * Result ' create 1,000,000 for Div32

a 32 bit word?
I must look at how the decimal value of 1 million is handled.
Actually, 1,000,000 isn't stored in the Result variable. The 32 bit result is stored in PBP system variables for use with DIV32, which then has to be the next statement.