>> Decimal floating point math can only ever be approximated by computers, so
>> it's fairly likely for "25.8" via one process to not equal "25.8" via some
>> other process, given that 0.8 is not easily representable in binary (it
>> seems to be an irrational number, checking to only 10 places).
>> 
>> 0.8 = 2^-1 + 2^-2 + 2^-5 + 2^-6 + 2^-9 + 2^-10...

> Wouldn't 25.8 simply be 258 x 10^-1.
> I was under the impression that floating decimal was an integer with a
> 'power' value that was based on powers of ten, not powers of 2.
> What you've described seems to be some odd 'binary fraction' system, that
> seems interesting yet very complex and (I would have thought) very
> inaccurate.

On reflection, I'm inadvertently arguing the case as to WHY it's done the
way you describe, as opposed to how it's actually done.  Oops :-/

That said, it doesn't explain how things like this 25.8 <> 25.8 crop up
rather frequently.  It's a fairly easy situation to reproduce (I shall try
to find an example).
-- 
Adam

---
You are currently subscribed to cfaussie as: [email protected]
To unsubscribe send a blank email to [EMAIL PROTECTED]
Aussie Macromedia Developers: http://lists.daemon.com.au/

Reply via email to