COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	THE IBM ANTI-TRUST CASE AND THE PUBLIC INTEREST
C00017 ENDMK
C⊗;
THE IBM ANTI-TRUST CASE AND THE PUBLIC INTEREST


	The Justice Department started a few years ago a new anti-trust
case against IBM.  In addition, CDC, various software companies, various
companies making peripheral equipment, and various leasing companies
have also filed suit.  The purpose of this essay is to try to determine
what the interests of the the various parties are in the matter and
especially what should be the stand of the computer scientists.

	We may consider that the public has the following interests:

	1. It wants good computers at reasonable prices.

	2. It wants a wide variety of choice in computers, peripheral
equipment and software.

	3. It wants the technology to develop rapidly and for new
ideas to be encouraged and tested.

	IBM has the following interests, and I will not presume to
speculate on their relative importance:

	1. Its stockholders want the stock to go up.

	2. Its executives, and engineers and scientists want to do a good
professional job in providing computers.

	3. Its salesmen want to beat the competition and exceed their quotas.

	4. Its executives want to lead a peaceful life and avoid harassment
and lawsuits.

	5. All its people would like to keep it intact as an organization,
because to do otherwise would make life very uncertain.

	IBM's competitors have the following interests:

	1. Survival as organizations.  For some of them, this is not a serious
question, but others have gone broke for various reasons.  Even if IBM is
not directly involved in a debacle, there is often a possibility of survival
by pressing IBM on monopoly grounds.  In one case, a few years ago, it seemed
to me that a company that didn't deserve to survive succeeded in blackmailing
IBM into giving it money and agreeing not to compete with it in a certain
market.

	2. Profits.

	3. Assured markets.

	4. Doing a good job.

	The Justice Department has the following interests:

	1. Their conception of what the public interest is.  Breaking up a
company with 70 percent of the business in an industry is certainly the
obvious solution, but there are drawbacks.  The first is that when elephants
fight, the mice get stomped.  If IBM and its resources were divided into
three or four parts and competed freely, each part might be so strong that
the non-IBM computer companies would be wiped out.

	2. Individually, the Justice Department lawyers would like to make
their reputations by winning the big case.

	In my opinion, the following additional considerations need to be
taken into account:

	1. The present situation is not intolerable.  IBM's fraction of the
market has been stable since the computer business started in the early
1950's.  Its competitors have prospered when they made better computers
than IBM in some market and have declined when they have made worse.  The
same is true of the peripheral equipment manufacturers, the software
companies and the leasing companies.  On the other hand, perhaps the situation
can be improved by some kind of anti-monopoly action.  We shall come to
that.

	2. The interests of the competitors are not identical with the interests
of the public.  If present trends continue, the result may be the creation
of a cartel, i.e. an agreed division of the various markets, created and
maintained by the U.S. government under the illusion that it is promoting
competition.  A prize example of the creation of a cartel is the recent
settlement between IBM and CDC wherein IBM agreed to give CDC its
service bureau subsidiary and not to compete in that business any more in
exchange for CDC calling off its anti-trust suit.  IBM had much less than
half the service bureau business, so that this agreement probably reduced
competition in that business.  As such, some of the other companies attacked
the agreement, but it may be that each of them will be satisfied if it gets
its cut.  Certainly, the fears of going out of business of many companies
would be alleviated by an agreed market division.

	In my opinion, this would be very bad for the country.  Philco and
General Electric and RCA and a number of smaller losers in the computer
business deserved to go out of the computer business in the sense that they
made no product not bettter made by someone else.  Once a cartel has been
created, any company about to go under will be able to demand a redistribution
of the business.  This will be at the expense of the users of computers.

	Another proposal by one of the competitors was that IBM be broken up
according to product lines, e.g. into a 145 company etc.  (This was their
example).  The effect of this would be to preserve a division of
computers by size that is probably already obsolete.  It would also be a legal
nightmare to determine what is a legitimate successor of a 145 in a different
technology.

	3. The public has suffered from the dominance of IBM in the following
ways only some of which may be remediable by the courts:

	a. IBM has built less than optimal computers and their losing
features have been copied by the competition.  Here are some examples:
For many years, no-one built computers that addressed more than 32K words,
because IBM was building successors to its successful 704.  No company
except D.E.C. with the PDP-10 has really made the full transition to
time-sharing computer systems.  No-one had to, because IBM goofed when
they came out with the 360 as they have recently realized and admitted.

	It is not clear whether an altered structure of the computer industry
would help.  It may merely a plea that computer engineers be smarter and
see further ahead.

	b. IBM's secrecy certainly works to the disadvantage of progress in
the computer field.  Because, almost no-one outside the company knows what
they are doing, there is no criticism if they are hatching a large egg for
the public that will appear in five years.  The lack of time-sharing for
the 360 is a case in point, and I have reason to fear they will
lay another suboptimal egg
in a few years.

	IBM's excuse for this secrecy is the anti-trust consent decree of
1965 which forbids them to hint at new products before they are ready to
quote price, performance and delivery.  Certainly, the Justice Department
has always allowed them to maintain this position.  However, IBM has a
tradition of secrecy dating to long before the consent decree, and the
secrecy has been applied in areas far from new products.  Moreover, even in
the area of new products, the secrecy is mainly to the disadvantage of
competition.  This is a case where lack of technical knowledge by the
Justice Department and the courts led to results which were opposed to what
they were trying to promote.  Let me illustrate these assertions.

	First, IBM's tradition of secrecy.  One retired IBM scientist told
me that when he started at IBM in 1949, each senior engineer at Endicott
locked up his laboratory at night, and the laboratory didn't open until
he unlocked it in the morning.  At that time, IBM's business was in
electro-mechanical devices with a long development time, i.e. inventions
in the classical sense, and this has always been the fortress of industrial
secrecy.  IBM became gradually less secretive until the early 1960's and
then got worse again.  At least, this is my personal impression as an
occasional IBM consultant.

	Second, two examples of the secrecy going far beyond new products.
Both of these are based on personal experience.  In 1962, I published
a method for defining the semantics of computer languages, and an
IBM laboratory took up this method, improved it, and used it to
describe the semantics of IBM's new language PL/I.  Because of personal
contacts and because I was an IBM consultant, I was aware of this, and
when a big report was ready after several years I received a copy of
it as an IBM Confidential document.  After considerable protest on my
part, they "declassified" the reports and the work has been public since.
Everyone concerned agreed that the work was of a general scientific
character and revealing it in no way pre-announced a product.  Such was
the atmosphere of secrecy, however, that it required the protest of an
outsider accidentally aware of it, to get it released.

	A second example is the manual for a version of the LISP programming
language that IBM Research prepared on the basis of a version of LISP
for the IBM 360 developed at Stanford University.  It was also declared
condidential, and, so far as I know, was never released.

	I don't know what other purely scientific treasures IBM may be
hiding, more through inadvertence than policy.

	However, the major disadvantage of IBM's secrecy to the public is
its effect on competition and progress in the computer industry.  Perhaps
we should put it another way.  Namely, here are some major advantages to
making IBM modify its secrecy policy: