RE: [boost] Fixed-Point Decimal Type

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2001-05-09 16:45:55

----- Original Message -----
From: Bill Seymour <bsey_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Wednesday, May 09, 2001 6:12 PM
Subject: Re: [boost] Fixed-Point Decimal Type

> Fernando Cacciola wrote:
> >
> > ... I recommend at least two default variations, one with an
> > efficient internal representation and another with a flexible
> > representation, allowing arbitrary large numbers. Some
> > applicatiosn will need that range flexibility.
> >
>
> I think I want to write a class that one or more vendors could
> implement as a built-in type under the as-if rule.  That's why
> I picked 18 digits (prior art on IBM mainframes, if I remember
> right).
>
> Something that is to the proposed class as vector is to built-in
> array would certainly be useful; but I see that as a separate class.
>
But as a user I wouldn't consider them so much different.
I would rather see it as I see float and double types.
Both would have the same interface and would be used for the same reason,
only one won't overflow.

At least keep this in mind. I believe that the implementation could be done
in such a way
that the internal representation is actually generalized.
Even if you don't let the user know about that generalization, you could
easly support both variations
with the same base code.

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com