Newsgroups: sci.space.history
From: henry@spsystems.net (Henry Spencer)
Subject: Re: If you could "do it right" how would you do it?
Date: Wed, 25 Oct 2000 14:35:46 GMT

In article <atqcvsgcoamqn4lnor4mgbov7f12u8rimj@4ax.com>,
OM  <om@RE_MOVE_THIS.ccsi.com> wrote:
>>This depends heavily on *exactly* how you define "first computer".
>
>...Especially since we should probably start with Charles Babbage's
>machines first, then.

Not if we count only the ones he *built*, or half-built.

The Difference Engine, which is the only one that got (partially) built,
was a complex specialized calculator, not a computer.  It computed
successive Y-values, for regularly-spaced X-values, of a high-order
polynomial, and that was *all* it could do -- it was just some memory and
some adders.  (The idea was that a mathematician would figure out how to
approximate something like a transcendental function, over a limited X
range, with a polynomial -- then, as now, a standard technique -- and then
the Difference Engine would crank out the actual table of values, greatly
reducing calculating time and avoiding arithmetic errors.)

The Analytical Engine, which would have been more or less a true computer
(with some reservations; notably, Babbage never did hit on the idea of
address arithmetic, which badly hampered his attempts to explain how his
design could do things like matrix operations), never got off paper, and
indeed never settled down to a single specific design.
--
Microsoft shouldn't be broken up.       |  Henry Spencer   henry@spsystems.net
It should be shut down.  -- Phil Agre   |      (aka henry@zoo.toronto.edu)