United States. Congress. Senate. Committee on the.

The Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) online

. (page 9 of 140)
Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 9 of 140)
Font size
QR-code for this ebook


to forecast that this will approach the $40-billion level in 5 more years.

That is even an acceleration and it is pretty dramatic, I think you will
agree.

And, by the way. it doesn't even take into account any abnormal in-
flation such as that the country seems to be facing these days.

Studying the economy and things such as inflation is not my field of
analysis, but I can read the newspapers and examine my own bank
account. So I am somewhat comforted by the role the computer indus-
try plays in our economy. Salary levels and user spending per em-
ployee are almost certain to increase. But computers, as you have been
hearing already, will continue to provide more "bang for the buck," as
people in the industry call it, as new technology is applied. The in-
creases in productivity per employee will probably keep inflationary
forces minimal insofar as computing is concerned. IMore and more, in
fact, computer use will be the only answer to these and other pressures.

But let's return to the user spending chart, now. I put in the
])ackage of handouts some of you have a circle that cuts the 1973
figures into the maior wedges.

Wo can see that just over one-third of user spending last year was
on salaries. About one-third was for EDP systems. That is the actual
computers themselves. And the final 30 percent was split among such
items as services used by computer users as well as by people who don't
need a computer all the time, supplies such as paper and magnetic
tape, things that feed data in and out of computers, software, the
instructions that tell the computer what to do. and finally, support,
such as data entry equipment — like keypunches, communication lines,
et cetera.

♦See p. 4954.



4937

I am not going to say very much about salaries today, gentlemen.
That is the one thing that is not IDC's specialty. And I might note,
however, that, as we spelled out in our background submission, the
data about employment is extremely sketchy.

According to AFIPS — the American Federation of Information
Processing Societies — which is about to publish a study, there are
about 114 million jobs in the Ignited States, jobs for people who manu-
facture computers, people who work with computer service com-
panies, and finally, the people who operate computers within their
own companies.

Today we are going to focus, as Ave have in our detailed background
report, on the $18 billion in this 1973 pie chart here that is spent for
things other than personnel in the United States.

We are going to add to that in our focus the $6 or $7 billion in
comparable spending abroad that comes into U.S. companies.

So we have, again, another $20 billion pie that we are talking about,
a revenue pie, but its pieces, or at least some of them, are somewhat
different from the $20 billion on this chart.

For a start, we are going to take a look at the population of com-
puters. It is a count, if you will, of how many boxes there are out
there and what they are worth.

This second chart shows the historical buildup of computers and
what we foresee in 5 years. It corresponds to page 30* of the back-
ground material.

You will notice that the blue, or the shaded part, represents the
value of computers installed, and the light yellow or blank bars in
the book represent the number of computers.

As you can see, there is a tremendous growth in the number of
computers expected over the next 5 years : from 133,000 in this coun-
try last year to almost half a million.

The value of these is not going up as much because we are going to
see an influx of small computers — lots of the so-called minicomputers
you've probably heard about, as well as many more small business-
oriented computers.

So you can see that the buildup — from either a numbers or dollars
standpoint — has been dramatic in the past. But what lies ahead, as we
measure it, is even more dramatic than anything we have seen in the
past.

Therefore, I think this is an appropriate point to take a look at the
various ways you can measure the size of the computer industry.

Senator Hart. We have to recess. I apologize. I hope this rollcall
will not run too long.

[Whereupon, a recess was taken.]

Senator Hart. I do apologize, gentlemen. The debate was over
copyright reform. It went a lot longer than we expected. Please go
ahead.

Mr. Peacock. Just to review where we had been. Senator, we had
focused on the user spending charts and the tremendous growth on the
first chart there.

We said that from now on we are going to concentrate on the part
of the user spending other than what he spends for personnel.

The next chart shows the computer population and growth, and you
can see there is a tremendous growth in the number of computers

♦See p. 4972.



4938

expected over the next 5 years, from the 133,000 level in this country
now up to almost half a million 5 years from now.

The value of these will not be »-oino- up as much, however, because
we are goino- to see an influx of small computers — lots of minicom-
puters, lots of small business computers.

This buildup, both in numbers and dollars, has been dramatic in
the past, but what lies ahead as we measure it is even more dramatic
than anything we've seen in the past.

So, I think this is an appropriate point to take a look at how one can
measure the size of the computer industry.

The chart we are looking at right now measures the value of equip-
ment installed. What we at IDC do essentially is to count up the com-
puters, figure out what the original purchase price of each one was.,
and multiply that out and add it up.

Unfortunately, this is not an indication of actual revenues. Like-
wise, any figures we can develop on what is normally measured in a
hard goods industry — the level of shipments — are not exactly a true
measure of the situation either.

"VVhy? A large number of computers — a decreasing percentage
today, but still a large number — are rented from the original
manufacturer.

In other words, the equipment is built and the people who make
it are paid, and the products are shipped out to the customers, who
then start paying the manufacturer on a monthly basis at rates of
approximately y^g the value of the computer for each month they
keep it.

So theoretically the computer is paid for after a customer keeps
it operating for 4 or 5 years. And I said 5 years because the rental
rate includes maintenance of computer equipment, and quite often
that is a pretty substantial job — one that involves one or more full-
time people to keep an eye on these big pieces of electronic gear.

This upkeep would be paid for separately if the computer equipment
had been purchased by the usei'. Last year, as a matter of fact, users
spent about 8 percent of the total dollars they paid out to computer
manufacturers for the maintenance of equipment that had already
been purchased.

This concept of the rental of capital equipment — the expensing
of capital goods items — is one of the things that makes the computer
industry, in my opinion, very different.

It is extremely important that you understand and appreciate the
dynamics of a rental-based business. For one thing, it alloAvs users
great flexibility in their choice of what computer they are going to
use and Avhen thev are going to upgrade or change models, and so
forth.

This concept of independence — real or merely believed by the user —
is proliably one of the important reasons that computers caught on
so quickly after they were introduced for the masses, as it were, in 1959.

On the other hand, the supplying of rental equipment means that a
computer manufacturer must liave funding, equity, capital with w^hich
to build the equipment, pay for the parts, pay the people who make
it, and sustain his business operation until enough e(piipment has
been out on rent long enough so that he starts to make a profit
on his original investment.



4939

In the long run, the snpplyinr; of rental equipment can be more
profitable than the outrifjht sale, because there is a <rood chance people
will keep the equipment for a longer period of time than necessary
to pay otf the manufacturing costs. But it does require a large amount
of capital, especially if a company wants to grow, penetrate the in-
dustry, and gain market share in whatever part of the industry he is
competing.

I recall quite vividly, about 10 years ago, having long discussions
with the head of what has become one of the major main-frame manu-
facturers. The ({uestion, simply : Can a company entering the computer
l)usiness grow so fast that it will never make it — never get into the
black — because it has to keep spending to increase market share in a
rental environment?

He finally admitted this could happen, and as a matter of fact he
actually had to slow his company's growth and obtain additional
financing until rental revenues caught up with operational and manu-
facturing costs.

Up to this point we have implied there are two ways for people
to acquire computers : buy them, or rent them from the manufacturer.

There is a third technique, one that came into vogue in the computer
industry during the mid-1960's. It is the concept of the third-party
lease.

This is not an unusual business arrangement, although there seems
to be some question as to whether it is the most viable in a high-tech-
nology industry.

You know how it works. A company with excess capital, or equity,
goes out and purchases computers that a specific user wants or that
the company thinks a user will want.

It then supplies the computer to the user at a monthly rate that is
less than the rental charged by the manufacturer. The economics of
this technique, according to the prospectuses that many of the com-
panies used to raise the funds with which to operate, is the fact that
a computer has a much longer life than 5 years.

These leasing companies think that the computei- has an 8- to 10-year
life. So they w'rite contracts, set their rates, with the prospect of get-
ting the purchase price of the computer, plus a profit, either from the
original pei-son who leases the computer or from a subsecpient lessee.

Some leases are written as full payouts. Others are shorter term,
sometimes only 1 to 2 years. Thefc shorter term leases ai'e called risk
leases or opei'ating leases, and some of the people who wrote these
back in the mid-to-late 1960's have had second thoughts once 5 years
or so passed and the original mamifactui'er. operating as it Avere on a
planned, o-year cycle, has introduced a new family of computers,
computers with — here again — more bang for the buck, which we see
in technology' today.

Some of these third-party leasing companies have been forced by
their accountants to take write-downs on the book value of the com-
puters they originally acquired undei- such a scheme. And one of the
most pertinent comments on the concept is that there is, with some ex-
ceptions of course, a different group of companies leasing computers
today than the original set. They are pj'imarily writing medium- to
long-term leases this time around.



4940

The leasing concept blossomed shortly after the introduction of
IBM's System/360. This family of machines quickly became the norm,
as it were, and it looked to the lessors as if the system's popularity
would guarantee a long marketplace acceptance.

This has been true to a great extent, but the marketplace accept-
ance for 360's today is at a price considerably lower than their origi-
nal value.

So it may take the lessors longer to recoup their money than they
originally planned. Some of them, of course, have managed to do all
right.

Let us now take a look at the computer systems marketplace, that 35
percent of the user spending pie from about 12 to 4 on the clock.

We will consider this to be the whole pie in the next chart that we are
going to look at.*

This pie chart actually shows several things. The big piece, repre-
senting slightly over three-fifths of the system's pie, we have identified
as IBM systems. This is that group of equipment that was either built
by IBM or was designed to operate with equipment built by IBM, and
it can be further separated into four pieces.

The first is the equipment currently owned by IB]VI and rented to
users.

Next is the equipment previously manufactured and sold by IBM to
users and theoretically available for use in any way the current owner
chooses.

Next is the equipment originally sold by IBM to a third party who
has ultimate control over the system even though the user — as in the
case of a user who has bought liis equipment directly from IBM^ — the
case just above — will probably keep up with the operating procedures
specified by IBM for its own equipment.

Finally, in that segment, there is equipment originally built by
PCM's — plug-compatible manufacturers — to operate directly with
IBM main frames with as much or higher performance than the IB]M
counterpart.

We can also look at this same big piece in a different way that I have
not shown on the charts. Taking the cut of the IBM system's base, we
find that 5 percent of the equipment, a very small amount, is old —
built back before system/360 started to go in about 10 years ago.

Thirteen percent is the small equipment like IBM's System/3, or the
control-type computers — IBM System/7 — similar to the minicom-
puters we will discuss in a moment.

So that essentially takes almost 20 percent out of this pie. The re-
mainder, we have 32 percent — just below one-third of IBM's installed
base — is System/360; 7 percent is that — as we discussed above — built
by the plug-compatible manufacturers; and the remaining 43 percent
is System/370, the family IBM introduced in 1970 and first delivered
in 1971, just 3 years ago.

That shows you how fast users tend to move up with their eriuipment.

Getting back to the big pie on the chart, the remaining 37 percent
is split among seven main-frame manufacturers as shown and the more
than 50 minicomputer and dedicated application computer manufac-
turers who have built more than half the computers installed in the
United States by number, but that represent less than 5 percent of the
installed base by value.

*See p. 4973.



4941

You can see there are two different ways of measuring what is going
on. So that is the way the systems pie looks in the United States.

On the next chart we are going to relate this $30 billion worth of
U.S. computers with the remaining $20 billion that are installed
throughout the world.

This chart, I hope, is not too confusing. It is designed to show a
couple of things.

First, this big pie represents the $50 billion worth of computers
installed in the free world, the non-Communist world, at yearend
1973.

The most striking thing to me about it is that only about 10 per-
cent of the equipment was built by a non-U.S. -based manufacturer.
The European computer makers and the Japanese computer manufac-
turers have supplied about 5 percentage points each of the world
supply of computers.

Second, you can see that U.S. computers only account for about
55 percent of the total. I am not going to try to get into balance-of-
trade questions liei-e, but I hope this chart will give you a feeling
for what I suppose you will be hearing later on this week.

In a nutshell, the United States uses just slightly over half the
world's computers; but U.S. companies have made 90 percent of them.

Third, as you can see on the chart, of the worldwide total, 61 per-
cent — just over three-fifths — are IBM systems, some of which are
owned by IB]M and on rent, some of which are owned by users them-
selves, some of which are leased by third parties — and that includes
banks abroad and other financial institutions — and some of which were
manufactui'ed by a plug-compatible manufacturer to attach into the
central box made by IBM, just as in the U.S. chart I explained briefly
a few minutes ago.

Before I attempt to get out any crystal ball and describe where
things might be going in the computer industry, I'd like to ask you
for just a minute to think back to the user-spending chart we started
with. - I

If you will recall, we haven't really yet mentioned that 30-percent
wedge of the pie, the wedge almost as big as this systems pie we have
been looking at here just recently, just describing.

This other sector, as the user-spending pies describe, represents
various types of support and alternative spending done by users and
nonusers of computers — nonuser being a person who doesn't have a
computer in his house but uses computer power from time to time.

On a percentage basis it is a bigger piece of the much bigger pie
in 1973 than it was of the 1968 pie. Incidentally, that figure in 1968 —
24 percent — is lower than it was the two prior time periods mentioned.

This happens as one support area, such as supplies, keeps decreasing
in relative importance while others grow. Thei-e is a pattern in the
shift, however, and it is a growing percentage for services.

Today — and we think more so tomorroAv, according to the user
studies we do — services represent alternatives to the ownership of
computers, and according to IDC's estimates will continue to grow as
a percentage of user spending.

This sector also includes the user spending for communication lines
and the various adaptors or modems he must use to pipe his com-
puter's messages over the telephone lines or to a satellite and back to
another part of the world.

40-927 O - pt. 7-3



4942

In general, the support portion of the pie will continue to grow at
the expense of systems spending.

I have now just given you my best 20-minute summary — with a
slight pause— of what the computer industry is today, how computsr
users spend their money, how the money spent for computer systems
is split among the various manufacturers in various countries.

I would, however, feel negligent in my duty to this subcommittee
if I didn't take just a feAv moments to share with you my educated
guesses as to where things might be going.

As a framework for doing this, I believe we have to take a look at
computers and computer equipment. After we see what the tech-
nology holds out. perhaps we can define how it might affect the
industry.

My last chart of the submitted background material is called "Gen-
erations of Progress.*"* It is an attempt to put on one piece of paper
highlights of what computers and computer use, and computer
acquisition and the alternatives available at a point in time, are all
about.

It is strictly my invention, if you will, and as such it may not be
precise to some of the technologists who will follow me in this seat.
But I find it useful.

Across the top I have broken the history of the computer industry
into time periods, or generations, of roughly 5 years each, a little
longer at the start, and no cutoff date for the future.

I think the dates are generally accepted ones. I haven't begim to
list specific products on the big chart up there. In general, as the typed
version in your book indicates, we had the named machines at the
beginning of the computer age ; the large, slow, clunky monsters that
ushered in the age of the computer.

In a sense each had a personality, and few of any particular model
were built. It was during this time period that some very important
people were saying that 2 or 3 dozen of these machines would take
care of the world's computational needs.

At any rate, we have gone from these scientific wonders of the mid-
dle lOSd's — machines with names like Eniac and Alwac and Univac —
to the business-oriented second generation computers, epitomized by
IBM's 1401, that brought data processing for the first time to many
businesses.

And then, finally, we stretch across that line to families of machines
that each manufacturer offers today. Looking out to the future, it ap-
]:)ears that the concept of network computers — as I will mention as Ave
wander down the chart — and networks of computers will change to-
day's concept of just what a computer is, anyway.

Coming down to the next line we have electronics, the circuitry used
within computers. This is what led to changes being called generations,
because during the time period involved very different and distinct
generations of electronics were coming along.

In the past 20 years we have seen electron or vacuum tubes replaced
bv transistors, and these discrete transistors replaced by integrated
circuits, and the drive for the future will be to make it smaller and
faster.

In the main memory area, delay lines quickly gave way to rotating
drums. By the late 1950's the magnetic core Avas introduced to hold

*See p. 4950.



4943

the computer's information, and this was the norm, in improved ver-
sions, until just the last few yea:-s when semiconductor memories of
various types have begun to be accepted.

In the labs scientists are talking about combinations of circuits and
memories that operate at temperatures near absolute zero but that put
the power of today's largest computers, or even more, into space the
size of a desk drawer.

Auxiliary memory — the device that is used to hold information that
h needed only on an occasional basis — has progressed from decks of
IDunched cards to the tapes so often used to symbolize the computer to
large disk stacks, nuich like phonograph record collections, and to
mechanical tape libraries.

Clearly on the horizon, as a very large auxiliary storage device, is
the Bell Lab invention called the "bubble memory." In fact, these
should be coming down the road in a few years.

Data entry has gone from the card punch, which will remain still
the most populous input method for another few years, to various key-
capture units to the concept of direct data entry.

Output has progressed from the "flashing light" era of the early
computer to faster and faster printers. In the future, with manage-
ment by exception techniques being applied to the computer, rather
than getting reams of paper we will begin to get only meaningful
documents or displays.

Also, paralleling this development, we will have action output, like
direct control of machines — which we have somewhat today — auto-
matic transfer of funds, and so forth.

The architecture of computer systems, the very way their elements
are put together, will continue to move from the classical combination
of processor, memory, control circuits, and input/output that you have
heard described by the people from Control Data, to a network con-
cept, a network oriented around a vast memory, around which many
special-purpose and general-purpose processors are hung and all of
which is really determined by the operating system — the instructions
that tell the computer how to work with the various elements that
make it up.

Throughout all of this movement there will be a thrust toward what
is called "transparency," the provision to the user of a system that he
essentially can't see, and doesn't have to understand; a system that
provides answers.

We believe users won't really have to worry about bits and bytes
and binary arithmetic that much anymore. Along with this answer-
orientation I would expect that manufacturers will begin to price
systems not by the box or by the pound, the way they do today, but
by the transaction.

We at IDC, in fact, are so convinced that the general movement
toward the automation of business transactions is coming that we have
started a newsletter and are looking at other products — at information
needs in this area.

As the Nation's monetary system and various reservation systems
get connected with credit and retail systems and hospitals and hotels
and insurance programs, as this autotransaction concept takes oil' and
eventually brings the advantages of data base access into the home,
things will change.



4944

There will be the true possibility of alternatives to owning a com-
puter. Certain manufacturers— Control Data, in particular— have
alreadv stated goals along these lines.

The' timesharing companies, the turnkey systems houses answering
specific problems with tailored products, even large companies not
closely identified with the computer industry today, yet strongly con-
sidering vast service operations, all of these people will compete for
the increasing computer dollar. Many of them will begin to offer valid
alternatives to computer users.

For the user who wants to continue "doing his own thing,'' how-
ever — and I am practical enough to admit this will be the majority



Online LibraryUnited States. Congress. Senate. Committee on theThe Industrial reorganization act. Hearings, Ninety-third Congress, first session [-Ninety-fourth Congress, first session], on S. 1167 (Volume pt. 7) → online text (page 9 of 140)