From: "Perry E. Metzger" <perry@piermont.com>
Date: Fri, 24 May 1996 08:47:01 +0800
To: trei@process.com
Subject: Re: ecash representation
In-Reply-To: <199605231726.NAA17935@linet02.li.net>
Message-ID: <199605231858.OAA26827@jekyll.piermont.com>
MIME-Version: 1.0
Content-Type: text/plain



"Peter Trei" writes:
> Back in the mid-80's, I worked for several years at Irving Trust, 
> a (now-gone) major money center bank. One of the financial 
> messaging systems I worked with stored currency amounts
> as 96-bit vectors of a base unit (eg, a penny), and 
> could have a 'binary point' anywhere in the vector. There were
> the usual math functions available to handle this data type.

Sounds like the usual fixed point hack used for manipulating and
storing money. Most systems I've seen use things like this.

Floats have all sorts of defects, like not conveniently indicating to
you the point at which they start dropping precision. A float doesn't
care that it just started dropping vast amounts of the precision in
some calculation where you are unfortunate enough to have done the
order of operations wrong. At that point, a hack will have kept its
precision or will indicate overflow, but floats blythely keep on
going. This leads to very unpleasant problems down the road.

Perry