But wait, if you print the value all at once, then why do you need the complicated incremental radix conversion routine? Why not just generate the whole binary representation at once, then convert it to decimal all at once, using the built in bigfloat or bigint operations? (Sorry, I won't write that in code now.)