BKMKNNMB.RVW   20031214

"Makin' Numbers", I. Bernard Cohen/Gregory W. Welch, 1999,
0-262-03263-5, U$40.00/C$67.50
%E   I. Bernard Cohen
%E   Gregory W. Welch
%C   55 Hayward Street, Cambridge, MA   02142-1399
%D   1999
%G   0-262-03263-5
%I   MIT Press
%O   U$40.00/C$67.50 +1-800-356-0343 fax: +1-617-625-6660
%O  http://www.amazon.com/exec/obidos/ASIN/0262032635/robsladesinterne
  http://www.amazon.co.uk/exec/obidos/ASIN/0262032635/robsladesinte-21
%O   http://www.amazon.ca/exec/obidos/ASIN/0262032635/robsladesin03-20
%P   279 p.
%T   "Makin' Numbers: Howard Aiken and the Computer"

In teaching about emerging technologies, I frequently point out to the
classes that those who fail to learn about computer history are going
to buy the same old failed ideas again, repackaged with new buzzwords. 
Nowhere have I found this more amply demonstrated, within the compass
of a limited total of pages, than in "Makin' Numbers."  Time and again
I found intriguing tidbits addressing concepts which we currently
consider highly advanced.

There is, for example, the concept of pipelining, and the speeding up
of execution time within the central processor.  The devotees of this
practice would be astounded to find the lengths to which Mark I
programmers took the idea.  Not content with simply preparing in
advance of an operation, they would actually start extra operations
with unused parts of the machine, such as getting in some extra
additions while the multiplication or division unit was crunching
through a multi-cycle function.

In one piece, Grace Hopper speculates on what Howard Aiken meant by
his continual reference to computing "engines," concluding that he saw
a computer as a kind of number factory, in which were employed a
number of specialized machines with differing functions.  This
corresponds with the prevailing thinking about embedded or pervasive
computing.

As a virus researcher, I am very sensible of Aiken's antipathy towards
von Neumann architecture, with no distinction between instructions and
data, and his pursuit of the forgotten Harvard architecture.  Making a
division between code and information that is processed would
eliminate viruses as a possibility.  It is, however, intriguing that
Aiken championed the idea, given his insistence on the pursuit of
usability in computers, and his prediction that programmers would be
more important than the fabricators of computing machinery: von
Neumann architecture is certainly much easier to use in developing
systems.

Even more than in the companion "Howard Aiken: Portrait of a Computer
Pioneer" by Cohen (cf. BKHAPOCP.RVW), "Makin' Numbers" provides a
wealth of ideas from the history of the field.

copyright Robert M. Slade, 2003   BKMKNNMB.RVW   20031214