From:	"Blaine Hodge" <bhodge AT gpu DOT srv DOT ualberta DOT ca>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: Floating/fixed point
Date:	Fri, 11 Sep 1998 05:18:35 -0600
Organization:	University of Alberta, Edmonton, Canada
Lines:	28
Message-ID:	<6tb13l$bfc$1@pulp.ucs.ualberta.ca>
References:	<Pine DOT GSO DOT 3 DOT 96 DOT 980910234423 DOT 14317A-100000 AT expert DOT cc DOT purdue DOT edu> <35F8EE90 DOT 5200CCB6 AT alcyone DOT com>
NNTP-Posting-Host:	async20-4.remote.ualberta.ca
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp

Think of fixed as an integer fraction.  Ok, that's exactly what it is.  Just
picture the integer value over 65536 (for 16.16).  Where as floating point
numbers generally are based on an exponent (ie. 1.0e10).  Like Erik stated,
each has its strengths and weaknesses.  Fixed can't represent a number like
1,000,000.  Float takes a bit more work to convert to an integer.

Erik Max Francis wrote in message <35F8EE90 DOT 5200CCB6 AT alcyone DOT com>...
>Nathaniel Johnson wrote:
>
>> fixed point is a special floating point that was made for speed... it
>> uses
>> a 32 bit number to make up a float 16.16
>
>That, of course, makes it not floating point.  Fixed point and floating
>point both represent (approximations of) real numbers, but neither are
>subsets of the other.  They both have different strengths, and different
>limitations.
>
>--
>Erik Max Francis / email max AT alcyone DOT com / whois mf303 / icq 16063900
>Alcyone Systems / irc maxxon (efnet) / finger max AT sade DOT alcyone DOT com
>  San Jose, CA / languages En, Eo / web http://www.alcyone.com/max/
>          USA / icbm 37 20 07 N 121 53 38 W / &tSftDotIotE
>             \
>            / Who's not sat tense before his own heart's curtain?
>           / Rainer Maria Rilke

- Raw text -