| From: | G.A.Griggs |
| Date: | 29 Sep 2000 at 00:50:09 |
| Subject: | Re: To sum it up |
Kree Andrew
On 28-Sep-00, you gnashed on about [amigactive] Re: To sum it up:
>AC Hi Everybody,
>AC>> AC If the basic you're using supports double precision numbers then
>AC>> your AC problems are solved :)
>AC> You seam to be missing the point no matter how precise your FP
>AC> system is you will still suffer from the .1 problem the simple fact
>AC> is that in binary .1 is the same as PI in Decimal you can't get an
>AC> exact representation. You can write your maths routines to handle
>AC> the problem within a certain window but if there supplied equation
>AC> generates a result somewhere during the procedure which falls
>AC> outside that window WHAM there's that problem again. If you don't
>AC> believe me just sit down with pencil and paper and calculate .1 as a
>AC Well, I've forgotten how to do division in binary, apart from it uses
>AC powers, so I can see it needs a lot of digits...
>AC Although, there is one way to fix it, BCD :) (Binary Coded Decimal
>AC 'case you didn't know). Bit wastedful on memory, but fine if you
>AC /need/ the accuracy :)
>AC See ya :)
Oh looks like I'll have to look up the 68k op-codes then, last time I looked
I did not see any for handling BCD, maybe you could refresh my memory.
Draylan tru
Everything should be made as simple as possible, but not simpler.
-- Albert Einstein
Quote carefully and read all ADMIN:README mails