Don't have to go to 'exotic' things like PI and e, simple things like 1/3 do
not have an exact representation as a decimal fraction
(0.33333333...333333....33333).
Similarly some somple fractions (do not have an exact binary representation
e.g. 0.1 is a repeating binary fraction. Others like 0.25 do have axact
binary representations.
----- Original Message -----
From: "Sherrill WATKINS" <SEWATKINS@dgs.state.va.us>
To: <tentec@contesting.com>
Sent: May 17, 2001 18:41
Subject: RE: [TenTec] Re: TenTec Amateur Radio Digest V4 #145
>
> Rich: What you are saying is extremely interesting. As a follow on
question and as an example; irrational transcendental numbers, like pi and
e, cannot be described 100% accurately in our standard base 10 number system
because they go on forever and have no end. So the computer system must
first convert this type of a number to a binary number, processes it and
then convert it back to base 10 for our use. Somewhere along the process,
the computer is going to have to decide just where it will round off or
truncate the digits within the number and this is where the error is
introduced? Is this the type of error to which you allude? - 73's-
Corn -k4own.
>
> >>> "Richard B Drake" <rich@w3zj.com> 05/17/01 01:05PM >>>
>
> Even the worlds largest super computers can not to floating point
> arithmetic 100% accurately. This is because it is physically
> impossible to represent fractional values 100% accurately in
> binary. The more bits that are used in the process, the more
> accurate it becomes but complex calculations have to be planned
> out very carefully to minimize the error and in the end there is
> always some rounding required. When you are talking about needing
> accuracy in area of fractional parts per million (10 Hz out of
> 30MHz) a 32 or even 64 bit floating point processor is in over its
> head.
>
> ----
> 73, Rich - W3ZJ
>
> > -----Original Message-----
> > From: owner-tentec@contesting.com
> > [mailto:owner-tentec@contesting.com]On
> > Behalf Of Carl Hyde
> > Sent: Thursday, May 17, 2001 10:59 AM
> > To: W B Reese; tentec@contesting.com
> > Subject: Re: [TenTec] Re: TenTec Amateur Radio Digest V4 #145
> >
> >
> >
> > The drifting Mark is seeing in Digipan is due more to
> > the software and the
> > computers Real Time Clock that to the Omni V's drift.
> > Remember PC Processors
> > don't do floating point math very well.
> >
> > TT sets the offset to 600 Hz so just tune 600 away from
> > the point where the
> > other guys cw tone dissapears.
> >
> > Hell we've been sending and receiving CW with drifty
> > old tube radios with no
> > crystals since ham radio began. How accurate does it
> > have to be? Not as
> > accurate as it is.
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Yahoo! Auctions - buy the things you want at great prices
> > http://auctions.yahoo.com/
> >
> > --
> > FAQ on WWW: http://www.contesting.com/FAQ/tentec
> > Submissions: tentec@contesting.com
> > Administrative requests: tentec-REQUEST@contesting.com
> > Problems: owner-tentec@contesting.com
> >
> >
>
>
> --
> FAQ on WWW: http://www.contesting.com/FAQ/tentec
> Submissions: tentec@contesting.com
> Administrative requests: tentec-REQUEST@contesting.com
> Problems: owner-tentec@contesting.com
>
>
>
> --
> FAQ on WWW: http://www.contesting.com/FAQ/tentec
> Submissions: tentec@contesting.com
> Administrative requests: tentec-REQUEST@contesting.com
> Problems: owner-tentec@contesting.com
>
>
--
FAQ on WWW: http://www.contesting.com/FAQ/tentec
Submissions: tentec@contesting.com
Administrative requests: tentec-REQUEST@contesting.com
Problems: owner-tentec@contesting.com
|