Ron Rosenfeld wrote...
On Fri, 19 Nov 2004 23:08:49 -0800, "Harlan Grove"
wrote:
"Ron Rosenfeld" wrote...
...
This UDF written in VBA might help you, at least for positive
numbers.
...

Presumably all it'd take to handle negative numbers would be

IF(x<0,"-","")&SomeUDF(ABS(x),...)

but that should be simple enough to add to the udf.

That's what I do if I need to outside the UDF.

But the OP was talking about binary numbers. My (limited)
understanding is
that there are several methods of representing negative numbers in
binary,
including one's complement, two's complement, and, with the
fractional part,
one can represent it with a signed integer and a positive fractional
part.
....

All depends on whether negatives would be represented with a negative
sign or in some form of binary encoding. Fractional parts would be
trivial if one assumes fixed point representation.

Maybe it's my prejudices, but exponentiation and logarithms are
unnecessary
and inefficient for simple base conversion.

You are a better judge of efficiency than I. But I prefer numeric
approaches.

Unnecessary exponentiation and logarithms are PURE EVIL in numeric
programming, as you should recall if you were doing numeric
programming back in the Dark Ages before FPUs were common in PCs
either as co-processors or included on the same chip as the CPU. It's
not just the speed performance, it's also the precision that suffers.
Take discount calculations, for example. An approach similar to yours

v = 1 / 1.05
dv = 0
For i = 0 To N
dv = dv + cf(i) * v ^ i
Next i

may seem like it does a reasonable job, but an approach similar to
mine

v = 1 / 1.05
dv = 0
For i = N To 0 Step -1
dv = dv * v + cf(i)
Next i

will do a much better job of preserving precision. But precision isn't
really at issue here. Change of radix is a simple procedure. It should
have a simple implementation.