Learning The Benefit Of Hindsight]

Robert S. Thau (rst@ai.mit.edu)
Sun, 20 Apr 1997 20:10:43 -0400 (EDT)

Ernest N. Prabhakar writes:
> (Watson -had- to want
> to sell more than five computers...).

Well, it's one thing to want to, and another to believe that there's a
reasonable prospect that you can. At the time of the supposed quote,
electronic computers were a new and extremely flaky technology; there
were a few people who firmly believed that they had business
applications, including pioneers Eckert and Mauchly, who had started
Univac to prove it, but there were also plenty of doubters; Eckert
and Mauchly found it *very* difficult to get financing for their

In fact, IBM's own first commercial computer, the 701, was sold to
management less as a product per se than as a patriotic gesture to
help national labs --- who were, in fact, a lot of the customers ---
because IBM's product planning committee found it very hard to believe
that many, or any, businesses would actually buy a computer, at least
given the then-current state of the art. IBM's first business
oriented computer (that meant hardware decimal arithmetic in those
days) was a productized version of an inadequate lab prototype which
got rushed out the door when IBM's customers started buying Univacs.

As a sidelight, even some pioneers of automatic computation were
skeptical about the prospect for large numbers of commercial sales;
Howard Aiken of the then-prominent Harvard Computational Laboratory
was well known for pooh-poohing the notion. (Then again, Aiken also
pooh-poohed other notions which turned out to have wide application,
such as full stored-program operation, which none of his machines
supported; this is the main reason that the Harvard Computational
Laboratory didn't *stay* prominent). I have seen the "at most five"
quote attributed to Aiken as well as Watson; from him, at least, it is
in character and completely believable.

(Aiken and IBM had cooperated in the early and mid 1940s; by the late
'40s there had been a break in the relationship, but some within IBM
might still have been influenced by the guy... though none of them
would have admitted it to Watson).