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 135 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 135 of 140)
Font size
QR-code for this ebook


with any kind of diligence in the protection of its interests.

15.40 I find that if the agreement violates the antitrust laws that plaintiff
has proved injury.

15.40.1 The creation of the mutually shared pool of technological know-how
on the part of SR and IBM, who at the time of the Agreement had about 95%
of the world EDP market, placed the other industry members not party to the
Agreement, including Honeywell, at a severe competitive disadvantage.

15.40.2 In agreeing to exchange know-how and information on developments,
SR assured itself that other EDP manufacturers would confront a decisive tech-
nological disadvantage or excessively high R&D costs, either of which would
reduce technological competition in the EDP market.

15.40.3 An analysis of the ratio of "imputed rental value" of annual EDP
system shipments to R & D expenses for the period 1957-1967 shows that Honey-
w-ell's R&D outlays relative to value of shipments were consistently much
higher than those of SR.

15.40.4 "Imputed rental value" as herein used means the value arrived at by
converting all EDP sale transactions in each year to lease transactions by a
formula which spreads the revenue derived from these sale transactions over

\



5880

a 3% year period commencing witli tlie first year the revenue was earned ; this
"imputed" value is then added to the actual lease revenue to obtain the "im-
puted rental value" (hereinafter "IRV") ; this is a valid base for comparison
purposes.

15.40.5 From 1957 to 1960, the four years immediately foUovping the SR-IBM
Agreement, Honeywell's R&D outlays exceeded total EDP revenues.

15.40.6 During the same period, SR's R&D outlays ranged from 10% to 22%
of its EDP revenues.

15.40.7 Although the productivity of a particular R&D investment may not
be immediately measurable, it is well accepted statistically that given the fact
that a company is investing R&D dollars every year, the ratio of annual
revenue or value of equipment shipped to annual R&D expenditures is a valid
basis for considering how productive R&D efforts are over a period of time.

15.40.8 It was not until 1965 that Honeywell attained a ratio (4.787) of IRV
of EDP systems shipped to R & D costs as favorable as the 4.619 ratio which SR
had enjoyed in 1957, the year immediately following the technological merger.

15.40.9 Utilization of the ratio analysis technique makes possible the ascertain-
ment of the competitive disadvantage suffered by Honeywell as a result of the
1956 Agreement in either of two ways: (1) the lower revenues of Honeywell
compared to its R&D outlays as compared to SR, or (2) the excess R&D
outlays Honeywell was forced to make in order to achieve any given level of
revenues as compared to SR.

15.40.10 The competitive disadvantage suffered by Honeywell as a result of
the 1956 Agreement (compared with SR's experience) measured in terms ol
lost IRV in the years 1958 to 1967, ranges from $516,877,000 (calculated with
no time lag between R&D spent and IRV earned) to $360,541,000 (one year
time lag) ; generally speaking, this means that had Honeywell's R&D expendi-
tures been statistically as productive as SR's for those years, Honeywell would
have generated between $360,000,000 and $517,000,000 more revenue than it
actually did.

15.40.11 The competitive disadvantage suffered by Honeywell as a result of the
1956 Agreement (compared to SR's experience) measured in terms of excess
R&D costs incurred in the years 1958 to 1967, ranges from $55,314,000 (calculated
with no time lag between R&D spent and IRV earned) to $36,871,000 (one year
time lag) ; generally speaking, this means that Honeywell had to spend between
$36,000,000 and $55,000,000 more for R & D than SR did in order to generate
the same dollar revenue.

15.40.12 The SR-IBM technological merger had the effect of causing a tech-
nological lag of several years on Honeywell, and it required unusually high
R&D expenditures in an attempt to overcome that lag.

15.40.13 To determine the extent of the injury to Honeywell caused by the
1956 Agreement and the technological sharing thereunder, the nature of that
agreement must be evaluated in light of what Honeywell stood to gain from
having access to it, or alternatively, what it stood to lose from not having
access to it, in terms of its competitive position.

15.40.14 What Honeywell's activities and subsequent competitive position
would have been must be compared to what its activities and subsequent com-
petitive position actually were in order to measure the total impact of the tech-
nological merger on it; essentially, the extent of the competitive advantage
Honeywell would have received must be added to the extent of the competitive
disadvantage it actually sustained.

15.40.15 In 1956-57 and thereafter, Honeywell and others of the EDP in-
dustry had workable central processing units ["main frames"] but were ex-
periencing substantial diflSculties in developing and producing peripheral and
input-output devices and were therefore forced to turn to costlier and disadvan-
tageous sources.

15.40.16 Because it recognized its deficiencies in the peripheral areas, Honey-
well representatives talked with representatives of SR in 1956 and requested that
SR sell Honeywell its high-speed printers and card readers ; SR declined to do
so on any basis ; this equipment was included in technology shared between
SR and IBM.

15.40.17 Therefore, Honeywell was obligated to and did ask IBM to supply
card readers and punches and went elsewhere to obtain printers.

15.40.18 IBM did agree to lease card punches and readers to Honeywell but
only on a full retail price basis, thereby eliminating that part of the EDP sys-
tem as a potential for profit for Honeywell.



5881

15.40.19 Frm 1957 to 1964, Honeywell leased for re-lease card readers and
piiiRlies from IBM.

ir).40.19.1 Beginning with the Datamatic D-1000, Honeywell's first computer
(liv.st shipped at the end of 1957), Honeywell obtained input/output devices
from IBM for use with Honeywell central processors. Among the IBM items
purchased or leased at retail prices by Honeywell were the IBM 407 printer and
the IBM 519 card punch.

15.40.19.2 The IBM 407 printer and the IBM 519 card punch were explicitly
listed on Appendix A to the IBM consent decree as items on which IBM waa
obligated to furnish technical information to all tab patent licensees. Honeywell
jiever sought this 407 and 519 technical information, or any other technical in-
formation, from IBM.

15.40.19.3 Biiiger, Honeywell's chief executive oflicer, did not know why
Honeywell did not invoke the IBM consent decree to get IBM 80 column punched
card equipment.

15.40.19.4 At various times over the years since 1956, Honeywell's Capp Smith
and Walter Finke complained to IBM about IBM's policy concerning prices on
peripherals it was selling or leasing to Honeywell. On such occasions, Birken-
stock testified that he countered that Honeywell was free to apply for IBM
kuuw-how and build the peripherals itself if it did not like IBM's policy. Birken-
stock claimed to have explained to Smith and Finke that IBM would make
available to Honeywell all of the technical information, both tab and EDP, that
it iiad made available to SR.

15.40.19.5 In August of 1059, Birkenstock met with Paul Wi.shart, then Honey-
well's President, to discuss the possibility of Honeywell's taking a license under
IBM EDP patents.

15.40.19.6 Following his August 1959 discussions with Birkenstock, Wishart
advised Finke that IBM's royalty rates were "a great deal more liberal than I
had thought they would be". Wishart also stated to Finke:

-I think what he (Birkenstock) was trying to say in a nice way was that he
knew we were infringing, and that perhaps the time had come to talk about
it * * *. I do not thing there is any occasion to immediately contract him, but
when we start delivering (our first computers), we may well be forced to do
this."

15.40.19.7 Henry Hanson wrote to AVill Freeman, outside patent counsel for
Honeywell, about the August 1959 Birkenstock visit Hanson stated :

"As we have discussed previously, we are not desirous of opening any patent
discussions at the present time. However, we may not be able to stall indefinitely
on this matter. I have been hoping that it would be possible to build up a better
picture in the anti-trust area tlian 1 have been able to do at the present time * * *
I am no planning to put any concentrated effort on this at the moment but to
continue on as we have been trying to locate our potential problems may be
and building up our own patent portfolio."

15.40.19.8 By March ol 1963, IB3I and Honeywell had agreed in principle on a
patent cross-license that was royalty free but with IBM to receive licenses also
under Honeywell's valuable industrial instrumentation patents as a "quid pro
■quo" for IBIM's greater EDP patent position.

15.40.19.9 The final IBM-Honeywell agreement was worked out in an April
1963 session between Will Freeman (late Honeywell patent counsel) and Birken-
stock. During this session, Birkenstock claimed, he had a detailed discussion
with Freeman concerning IBM's tab and EDP technical information exchange
Avith SR in 1950 and the fact that similar IBM information — as of October 1,
195(j — was available to Honeywell.

15.40.19.10 At their April 1963 meting. Birkenstock said. Freeman advised
that since the IBM know-how (teehSical information) was limited to October 1,
1956 he believed it had little value to Honeywell at the time. Freeman also
■expressed a reluctance, according to Birkenstock, to request IBM know-how
because he did not want to expose Honeywell to the possibility of IBIM's asking
for Honeywell know-how in the instrumentation field.

15.40.19.11 At his April 1963 meeting with IBM's Birkenstock, Freeman sug-
gested, according to Birkenstock, that the IBM-Honeywell agreement specifically
>;tate that no know-how was to be provided by either party ; such a provision
was incorporated into the agreement.

15.40.19.12 Final drafting of the IBM-Honeywell agreement was concluded,
.and it was executed in April of 1964.

40-927—75 67



5882

15.40.19.13 The IBM-Honeywell agreement called for nonexclusive patent cross-
licenses on existing "information handling" patents and on patents issuing on
applications filed prior to May 1, 1968.

15.40.20 In December of 1964, IBM notified Honeywell that it would no longer
lease for re-lease card readers and punches to Honeywell.

15.40.21 That meant that Honeywell was forced to purchase card readers and
punches at full list price from IBM in the current year and then wait several
years for a return through rental payments of money which was currently paid
to IBM for its readers and punches.

15.40.22 Although tliis placed Honeywell at a significant competitive disad-
vantage in the EDP market, Honeywell nevertheless continued to buy readers
and punches from IBM since SR would not deal on any terms and there was
no other alternative pending final completion of design and development of
Honeywell readers and punches.

15.40.23. In May of 1966, representatives of Honeywell visited representatives
of IBM to ask that IBM reconsider its policy that it would not lease for re-lease
readers and punches to Honeywell.

15.40.24 IBM refused to reconsider both policies, SR knowing of IBM policy
and vice versa.

15.40.25 From at least as early as 1956-57, it has been very important for an
EDP manufacturer to develop its own complete system because customers dis-
liked "split-system responsibility" with one manufacturer maintaining the central
processor and one or more others maintaining the peripheral devices.

15.40.26 The purchase or lease of peripheral devices by an EDP system producer
from another manufacturer removed that area of the system as a potential profit-
maker.

15.40.27 Peripherals have contributed an increasingly greater share of total
system price during the period 1955 to 1970.

15.40.28 Peripheral and input-output devices have been and are now critically
important to EDP system customers and hence to sales since they determine
the "through-put" of the system, the speed and flexibility with which the system
can meet the user's needs, and often, the quality of the system user's perform-
ance for its customer.

15.40.29 A range of types of peripheral devices (as well as a range of speed
and capacity within each type) offered with a central processor permits the
selection of the best combination of capability and price for the customer.

15.40.30 Tlie EDP manufacturer with a limited set of peripherals was and is
therefore severely limited in its ability to compete.

15.40.31 Since the late 1950's, emphasis has been moving increasingly away
from the central processor and toward peripheral capabilities in the EDP custo-
mer's comparison of performance of data processing business applications by
competing EDP systems.

15.40.32 It is common today, and has been increasingly so since the late 1950's,
to refer to a system with apparent limitations in eflaciency as "printer-bound"
or "input-bound" or "tape-bound" meaning that the speed of a particular periph-
eral unit is the measure of EDP system performance.

15.40.33 Honeywell experienced substantial deprivation and competitive dis-
advantage because of nonaccess to the exclusively held technology which IBM
and SR shared under the 1956 Agreement.

15.40.34 Although the Honeywell EDP organization had a strong management
position and a growing sales capability in 1956/1957, it had little if any special-
ized know-how with respect to peripheral or terminal equipment design and
evaluation or the production of electronic or mechanical units except on a modol-
shop basis.

15.40.35 Had Honeywell participated in the information sharing between IBM
and SR, it would have obtained valuable, needed know-how to supplement that
which it already possessed.

15.40.36 With access to the SR-IBM shared technology, the overall course of
Honeywell would have changed noticeably ; access to the exchanged information
would have significantly changed the climate in which business decisions were
being made at Honeywell EDP in the areas of management, finance, engineering
and marketing during the entire period from 1956 until 1970.

15.40.37 Honeywell would have achieved a strong position as early as the
1960-1964 period if the company's two prime weaknesses had been eradicated in
the late 1950's by access to the shared technology :

.1 The lack of production and commercial product experience ; and



5883

.2 The lack of knowledge and customer and corporate confidence in tlie
peripheral areas from both a design and production viewpoint,

15.40.38 The following benefits would have accrued to Honeywell from the
time Honeywell obtained access to the 1956 knowhow technology :

.1 Electronic products would have been produced at lower costs with better
reliability, and EDP systems could have been brought into the market sooner;

.2 Electromechanical products would have been selected with greater dis-
crimination, implemented for better performance, and produced in-house at a
much earlier date with greater customer acceptance ;

.3 A broader and more reliable product line would have enhanced the image
of the company as a full-line vendor of EDP equipment ;

.4 A full-line vendor image would have resulted in more customers earlier ;
and

.5 A broader product line would have attracted more creative design talent
to Honeywell employment in both hardware and software areas.

15.40.39 The primary technical strength of the Honeywell organization rested,
in its central processor design competence rather than in the areas of implemen-
tation of designs in reliable, economical or producible hardware.

15.40.40 Honeywell would have profited by access to the merged technology hv
the areas of memories, central processors, and control units by knowing cer'
tainly what had already been discarded as unreliable, uneconomical or otherwise'
unsuitable.

15.40.41 Brute-force implementation techniques were utilized in the Honeywell
D-1000 and H-800 in an attempt to resolve problems which surely had more
reasonable solution ; however, the pressures of time and lack of design knowl-
edge and experience forced Honeywell engineers to find their own solutions which
were often clumsy, expensive, or diflicult to reproduce, all of this would have
been alleviated from the time of and by access to the merged technology.

15.40.42 Some of the major potential contributions from Honeywell access to
the shared technology are in the areas of improved implementation of the con-
ceptual design (better documentation, better packaging, better powering) and
in the application of the detailed information to efforts to replace or coexist
with equipment of IBM or SR which a potential customer already had ; none of
these could be supplied by examination of the competing equipment or by "re-
verse-engineering."

15.40.43 The Honeywell D-1000 systems were built one-by-one by highly trained
engineers and technicians, as were the first H-SOO systems ; access to the merged
technology would have ameliorated this condition for Honeywell and the public
would have benefited.

15.40.44 Honeywell EDP engineers had little production experience in 1957;
particularly, a lack of knowledge of what could be produced in a factory as op-
posed to a model shop or test laboratory, and a lack of knowledge of what the
factory had to know to produce such equipment ; the technology shared between
SR and IBM covered such matters.

15.40.45 The lack of production experience among Honeywell EDP engineers
resulted in the use of older technology when current technology of Honeywell
(not equal to that shared between SR and IBM) did not produce reliable prod-
ucts, or resulted in expensive model shop crash redesign programs and other
activities tending to increase costs, delay shipment, reduce reliability and damage
Honeywell's performance and image.

15.40.46 SR's UNIVAO II and File computer, made accessible to IBM, both
were constructed using printed circuit packages ; had information on these been
available to Honeywell as it was to IBM, Honeywell would have been able to
improve both design and documentation by engineers and improvement of pro-
cedures and techniques by production personnel ; IBM had access to all of thi.*?
in the technological merger.

1.5.40.47 The SR-IBM information concerning design and production would
have been valuable to Honeywell with respect to connectors, connector tech-
niques, power supply and distribution, cabling and cooling.

1.5.40.48 Honeywell, having personnel with little knowledge of production tech-
niques, used older technology or over-designed for performance reliability, with
the result that the electronic portions of the H-800/1S00, 400/1400 and 200,
were more expensive to build and operate than competitive equipment, which
affected Honeywell's competitive ability in terms of price and customer costs
in floor space and environmental condition.



5884

15.40.49 These deficiencies in Honeywell's design implementation, documen-
tation and production know-how would have been substantially eliminated by
access to the technology shared between SR and IBM.

15.40.50 These deficiencies resulted in a loss of sales and loss of EDP image
becau.se of the deficiencies in technology displayed by Honeywell.

15.40.51 Honeywell's financial losses were reflected not only in lost sales
dollars and increased costs for production and repair of hardware, but also a
loss in market place status as reflected in the difficulty in hiring quality ijersonnel
and gaining access to sophisticated buyers.

15r40.52 As early as 1959, in developing the H-SOO system, Honeywell was
forced to attempt to achieve compatibility with certain SR and IBM equipment
in order to make sales.

15.40.53 In late 1957, Honeywell was weak in the entire area of high capacity,
medium-access-time storage when contrasted with the IBM 305 RAMAC and
the SR Magnetic Drum File Computer, both covered by the shared technology.

15.40.54 In 1958 and thereafter, the weakness of Honeywell in random access
units was increasingly apparent and the competition, including SR, capitalized
on that weakness.

15.40.55 Honeywell had few i>ersonnel experienced in this branch of. EDP
-technology and any realistic prospect of developing random access equipment
from scratch was hopeless.

15.40.56 Honeywell, in 1957 and for several years thereafter, could not have
become competitive in this rapidly evolving technological area without a body
of technological information on which to build; the shared technology would
have provided this know-how.

15.40.57 The Bryant disc memory unit was selected to be the device offered
by Honeywell with its .systems, although not without reluctance.

" 15.40.58 Four systems containing the Bryant disc unit were shipped to Honey-
well customers, but they did not become acceptable in the general marketplace
primarily because of inability to keep them operative in service.

15.40.59 For the immediate future, Honeywell was thus committed to the
use of the Bryant disc unit but the difficulties with this unit coupled with the
increasing pressure of the market required that investigation of other units be
continued.

15.40.00 Honeywell considered as late as 1964, developing a unit using magnetic
tape loops to provide a random access capability because of growing indications
that IBM 1311 (random access) unit would destroy any possible marker .success
for tlie Honeywell 400 and also the Honeywell 200; SR and IBM had shared
this teclmology in 1950-57.

15.40.61 Honeywell personnel concluded that wide tape woiild not be competi-
tive either with" the IBM 1301 disc unit or with the IBM 1311 disc pack unit
and that Honeywell would need another program for random access mass

15.10.62 Large numbers of important EDP prospects, some of whom already
were customers, made it plain in 1964 that they would not continue to consider
Honeywell as a supplier because Honeywell was not competitive in random
access.

15.40.63 In September 1934, Honeywell was forced to -offer to provide IBM
2311 disc pack units with Honeywell systems or be faced with the alternative
of continuing to lose orders.

15.40.64 In December 1964, The Bryant disc unit was removed from the product
line because of difficulty in operation and the conviction by Honeywell personnel
that it was no longer competitive.

15.40.65 Honeywell's attempted internal development of a ma.gnetic card ran-
dom access storage proved to be a disaster both financially and reputation-wise.

15 40.66 The Honevwell card mass storage project was terminated as a product
development in the "first quarter of 1966 because Honeywell could not perfect
the card storage unit to a point at which it would function reliably : access
to the 1956 technology shared by SR and IBM would have provided a substitute
whenever access was achieved. ^^^ .^ ,•

15.40.67 Honeywell immediately instituted a series of orders to CDC for di.sc
pack drive units and began a crash program for control units, both resulting

in heavy expenses. , .:, , ^ ^ ■ inon

15 40 68 In 1966, Honeywell began its own disc pack development and m 1969
nut into production random access equipment of its own comparable to current
offerings of disc pack drive units of other vendors; the shared technology
anticipated this by over ten years.



5885

15.40.69 Had the IBM 305/355 information been made available to Honeywell
in 1957 as it was to PR, the following would have resulted [each year of defer-
ment bevond 1057 would have postponed the results correspondingly] :

.1 Honeywell would have undertaken in 1957 to build random access storage
systems starting with the H-SOO, increasing the customer base and storing up
a' reservoir of know-how on hardware, software, and systems ;

.2 The net result to Honeywell of this development project would have, at
least, yielded a workable, acceptable randt)m access storage unit, and the leverage
on sales would have been markedly positive ;

.3 Honeywell would have avoided involvement with the Bryant units on the
800 and 400\systems. ^^ , ,.,

.4 Honeywell would have avoided the card mass storage debacle which cost
it not onlv inonev but also time and adverse publicity ;

.5 Honeywell would have saved some of the very large disc drive development
ccsts incurred from 1966 to 1970 ;

.6 Honeywell would have avoided investing a very large amount of money in
the purchase of CDC's disc pack drives ; and



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 135 of 140)