Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!usc!aero!djp
From: djp@aerospace.aero.org (Dennis J. Persinger)
Newsgroups: comp.lang.smalltalk
Subject: Re: floating point support
Keywords: double precision, VAX G and D
Message-ID: <57988@aerospace.AERO.ORG>
Date: 18 Sep 89 23:02:54 GMT
References: <260@argosy.UUCP>
Reply-To: djp@aero.UUCP (Dennis J. Persinger)
Organization: The Aerospace Corporation, El Segundo, CA
Lines: 32

In article <260@argosy.UUCP> becher@argosy.UUCP (Jonathan D. Becher) writes:
>I've been playing with the number junta in ParcPlace Smalltalk 2.4
>and need to add double-precision floating point support (it only has
>single-precision).  I'd like to include IEEE, VAX D and G formats to
>be complete.

ParcPlace Smalltalk 2.5 (being released as we speak) includes fileins for
Double Precision, Complex Numbers, and Meta Numbers (infinities, infinitesimals,
and a few other special types).

Doubles arithmetic is supported by primitives, but I haven't tried using
it yet.  Below is the class comments from the type 'Double':

    Instances of the class Double represent floating-point numbers in 
    IEEE 64-bit format.

    These floating-point numbers are good for about 14 or 15 digits of 
    accuracy, and the range is between plus and minus 10^307.
    Here are some valid floating-point examples:
	8.0   13.3   0.3   2.5e6   1.27e-30  1.27e-31 -12.987654e12
    The format consists mainly of no imbedded blanks, little e for tens 
    power, and a digit on both sides of the decimal point.

    Instance Variables: *byte indexed*

    Class Variables: 
	Pi			<Double> 3.14159
	RadiansPerDegree	<Double> 0.0174533

Dennis Persinger
The Aerospace Corporation
djp@aerospace.aero.org