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 112 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 112 of 140)
Font size
QR-code for this ebook


tliat IB^I would he successful in that suit. Kevill and Hancock informed Wade
tliat Telex wanted to build the Merlin control unit in such a way that they would
nut be sued. Kevill and Hancock told Wade that he wonld have to disguise the
control unit in such a way tliat it could not be easily proven that it was an exact
coiiy of the IB^M Merlin control unit.

Kevill and Hancock indicated to Wade that if he were to be sued for copying
tlie TB^r Merlin control unit Telex would provide legal support. In form, former
IBM employees were told by Telex oflScials that they were not expected to
utilize IBM trade secrets and on occasion they were expressly forbidden to do
so : and generally forme IBM employees were hired on a specified ''condition''
that they would not utilize nor employ trade sercets or confidential information
of IBM.

It is quite apparent from the entire record and the very circumstances of hiring
that these latter conditions and directives were largely mere formalizations and
protective devices. There is no doubt that a number of the f(M-mer IBM employees
liired by Telex conscientiously endeavored to separate in their minds the
detailed confidential and trade secret information of wiiich they had knowledge
and which they had retained, from their general competence, and thoir judgment,
expertise and experience of which Telex was entitled to the benefit.



5752

But Telex's necessity for such confidential and trade seciet information was
so manifest, tlie pressure so great and Telex's program and policy to obtain that
information so prevasive during this particular period as to have rendered it
difficult if not impossible for engineering personnel to fully protect IBM's trade
secrets.

FI06. Accordingly, the court finds that Telex deliberately set out to misappro-
priate the Merlin trade secrets, hired Clemens and other IBM engineers for this
purpose, and succeeded in misappropriating a substantial number of those
secrets and incorporating them into the Telex 6830, built at the Telex TDAS
facility.

F157. The ^lerlin control unit and disk drive announced by IBM in June, 1970,
as the IBM 3S30 control unit and 3330 disk drive were the result of a five year
development program at IBM, which cost in the neighborhood of $30 million.
At the time of announcement six or seven of the features were disclosed but na
design information for implementation was included.

The design features listed below were largely new in the context of this
device, were individually valuable for the purpose and provided in combination
a control unit having improved characteristics over any other which had been
introduced at that time.

A substantial number of these features were confidential and constituted trade
secrets in the context of the 3830 and the combination of them in the 3830 was
new, innovative and constituted valuable confidential information and trade
secrets. These were misappropriated by Telex and incorporated in the Telex
6830 prior to the first customer shipment of the Merlin which was made in
August. 1971 :

Semiconductor read/write control store

4K (4096) words of control store

Microinstruction word of 32 instruction bits and 4 parity bits

Variable format microinstruction word

Use of 4 format definition bits

Use of 2 field suppress bits

Use of a 4 bit CK field

Eight specific ALU operations

Four specific direct branch conditions

Use of a specific bit to gate the CK field

A technique of adding a field to the data address register

Command retry

Rotational position sensing

Record reorientation

Upper-A microinstruction word format

Lower-A ^Microinstruction word fonnat

Upper-D microinstruction word format

Lo\ver-D microinstruction word format

Error detection and correction capability.
The use of each of these features in the Merlin, and their details, were known
to one or more of the following IBM employees hired by Telex from IBM : J. K.
Clemens, Robert J. Hancock, Sterling Hon, W. Edward Ice, B. O. Glover, and
J. Kevill. The Telex Meriln 6830 disk file control unit produced in 1971 by the
Telex TDAS group at Santa Clara, California, prior to delivery to customers
by IBM of the first IBM Merlin disk file control unit embodied all of these design
features.

F158. It Is true that some of these features were in general context neither
new, novel, secret nor innovative ; but they all had considerable value in addition
to their value as separate elements since they represented a design composite
which allowed IBM to achieve its goals of the 3830 in terms of cost and per-
formance, and since their use was calculated to put Telex in a position to
compete upon a comparable basis of cost/performance.

No skilled microprogi-ammer with no exposure to the IBM 3830 control unit
in its specific implementation could have arrived at the same degree of con-
formity or similarity to the IBM product as did the TDAS 6830 control unit. The
control unit designd adoi'ted by TDAS for llie Telex 6830 disk file control unit
was only one of enumerable design choices which could have been selected to
achieve this same function. It would not have been possible for Telex to
incorporate the precise technical designs previously incorporated in the 3S30
without coyping those designs.



5753

F159. As a result of its access to and use of IBM's trade secrets with respect
to Merlin, Telex was able to develop its (J830 control unit in substantially less time
-iind at substantially lower expense than would have been the case had it been
developed independently. Telex thereby was unjustly enriched. By May 15,
liJTl, Telex TDAS had completed approximately GU% of the 3830 type disk con-
troller internal processor design.

The IBM development effort required between two and three years to reach the
equivalent point in the design of the 3830 control unit. In IBM a competent
group of engineers with disk drive experience but no prior Merlin experience
required six years to develop the Merlin. Telex scheduled the development of au
equivalent Merlin system with a competent group of engineers, including
selected key people with IBM Merlin experience, in only 18 months through the
use of IBM's knowledge of Merlin.

The completion of the project by Telex, however, was not accomplished
within this period, since Telex's Santa Clara, California, facility, established
for the purpose in early 1971, was abandoned in April, 1972. The evidence does
not disclose precisely what percentage of the project was then finished, but it
seems fair to infer from the evidence that it was more than half completetl. In
IBM the approximately (> year development effort ou Merlin cost IBM over
.J?3U million.

A group of people experienced in Merlin and developing Merlin in 18 months,
or even in twice that length of time, would have a substantial amount in
development cost; but tlie saving that could be deemed to flow from the utiliza-
tion of protectable contidential information and trade secrets by Telex would
not be proportional to the time involved, since Telex was entitled to utilize the
general expertise and competence of its employees in view of generally develop-
ing technology, and this in any event should have substantially reduced the
time required for the redevelopment of a Merlin design.

Considering the entire record, and by fair approximation, the court finds that
as a proximate result of the improper utilization of IBM's trade secrets, Telex
saved for itself at least ij^lO million in its Merlin development costs to the extent
tliat development was comi)leted by April of 1972, and was unjustly enriched
to this amount ; by rhis amount, on a standard of comparison basis, IBM was
damaged beyond the sales price Telex received for this and other designs as
hereinafter mentioned, there being no better gauge by which to measure IBM's
entitlement.

F160. As above indicated the Clemens' Santa Clara groiip proceeded with its
assigned work in design and development of a disk drive controller until April
of 1972, at which time it had not been fully completed. At this date Telex
altandoned the Santa Clara activity and .shortly thereafter the Santa Clara
facility was closed. Telex determining to purchase disk drive controllers from
Itel Corporation and to market them under the Telex brand. Telex never pro-
duced or marketed a disk drive controller manufactured to the design which
was in progress at Santa Clara, nor has it since done so.

On May 22, 1972, it agreed to sell, and Control Data Corporation agreed to
buy, the manufacturing rights to the Telex 6830-type disk control unit, pur-
suant to a non-exclusive and non-assignable license agreement, together with
IBM's FRIEND program source deck hereinafter mentioned for the sum of
^500,000.

In addition Control Data agreed to pay Telex an additional $36,000 for proto-
type number 1 and number 2 of the control unit, residual supplies, and shipping,
packaging and handling charges. Telex has sought to derive further benefit from
the sale of others of IBM's Merlin trade secrets, and there is a reasonable
probability that Telex will do so to its further unjust enrichment and to the
great future damage of defendant unless enjoined by the court.

F161. "FRIEND" (Versi(m 2) is an acronym for Fast Running Interpreter
Enabling Natural Diagnostics. The source code for FRIEND competitively is
extremely valuable to IBM and would be of great value to IBM's comijetitors.
The FRIP]ND program is a diagnostic program used by field engineering and
development engineering personnel to assist in the diagnosis, checkout and de-
hugging of various devices in a computing system.

It permits more rapid diagnosis of malfunctions and assists in identifying
errors in design. It has lieen u<ed in the development of various IBM products.
The program was developed at the IBM San .Jose laboratory prior to 1970. one
Findlay having worked nearly two years on the program and having received an
outstanding contribution award from IBM for his efforts.



5754

F162. Findlay developed FRIEND (Version 2) by writing instructions in
programming language intelligible to a human reader. In this form the program
is called a "source code". The source code was subsequently translated to an
"object code" which is the form of the program that can he utilized by comput-
ers. While the source code of a program such as FRIEND can be translated to
an ob.iect code, the object code cannot be translated to the source code.

A printout of the object code of the FRIEND 2 program is nothing more than
an unintelligible mass of letters and numbers which only a computer can utilize.
The source code for FRIEND, unlike the object code, can be used by development
engineers to design and test new products and for other pui'poses. The source
code for FRIEND (Version 2) has always been treated as proprietary and con-
lidential by IBM.

Since its completion, it has Ijeen stored on a reel of magnetic tape and kept
in Findlay's possession. IBM has never released it. and only the object code for
this program 1ms been made available to IBM field engineers and those outside
IBM's development lal)oratories.

Fl()3. Prior to leaving IBM to work for Telex, an IBM employee secured a
copy of the FRIEND (Version 2) program and brought it with him to Telex.
While the circumstantial evidence establishes this fact beyond question, the
identity of the employee has not been determined. In the spring of 1971, Neil
Glover, then an IBM employee, left IBM to begin work at Telex.

Glover and Findlay worked together in the same diagnostic programming
group at the IBM San Jose laboratory. Glover in his work had access to the
source code for FRIEND (Version 2). Brigitte deSaint Phalle, then an IB^I
employee, was hired by Telex in July of 1971. When she arrived at the TDA
facility she was given a source listing and comments of FRIEND.

The source listing and comments belonged to Glover. The source listing was
in the form of a computer printout. She transferred this printout to a deck
of pxuiched cards. When the TDAS facility was closed by Telex, much of the
matrial from TDAS was sent to Tunisa and some of the material in punched
card form was placed on magnetic tape. This work was done by John Hasty, a
Telex employee.

During pre-trial discoveiy Telex produced to IBM a number of reels of mag-
netic tape. One of these reels identified as "Hasty 003" contained material from
the TDAS facility and the first portion of that tape consists of a source listing
of the FRIEND (Version 2) program.

The source listing of the FRIEND (Version 2) program in Telex's possession
was copied from a soui'ce listing created by Findlay, as shown by irrefutable
circumstantial evidence: Both listings have the same numl)er of pages — "t-t ;
each of these pages has the same format and the same number of lines; idio-
synchratic notations used by Findlay and not generally used by other pro-
grammers appear in both listings ; spelling errors made by Findhxy appear in both
listings ; the only difference in the listings consists of 22 keypunch errors that
appear in the Telex version and normally could have been made in copying.

Telex used the source code of FRIEND (Version 2) to help in the design of
the controller under development at Telex's TDAS facility, even tliough the pro-
prietary status of that program was later sold to Control Data Corporation by
Telex as part of the May 22, 1972, agreement above-mantioned. The court finds
that Telex deliberately misappropriated, iised and sold to Control Data this
FRIEND (Version 2) source code listing and comments.

F164. Prior to November, 1970, IBM initiated several secret and confidential
advanced disk development programs at the IBM San Jose lahoi-atory which
were given the code names Winchester, Icel)erg, Apollo and ^lidas. Except fin-
certain characteristics of the Winchester program which were dis-closed as iiart
of the announcement of the disk product as the 3340 in March, 1973. the Win-
chester, Iceberg, Apollo and Midas development programs have been treated
as .secret and confidential IBM development programs by IBM, and all depart-
ing employees were reminded of this confidentiality.

F16f). Beginning in early 1971. Telex began developing products intended to lie
equivalent to Midas. W^inchester and Iceberg, based in part on IBM trade secrets
and confidential information solicited and obtained by Telex from its employees
who had previously been employed by IBM,

Shortly after Telex embarked upon its extensive hiring program, it revised
its product calendar to include the 7330 (characterized as the "next generation
6330") which was Telex's planned copy of Iceberg and the 7312 (characterized
as a "head in. pack; 30 MB/cart") which was Telex's planned copy of Win-



Chester. In early 1972 Telex tenninated its 3330 niannfacturing activity at
Santa Clara.

However, Martin ordered Kevill to continue work on the "3330 spindle as
a stepping stone for the Midas, Apollo, Iceberg and Winchester type product."

It is clear from numerous recruitment activities and Telex's internal com-
munications and documents that Telex's advanced development disk jirogram
included future disk products identified as Iceberg, Midas, Winchester and
Apollo, and that an important effort of Telex was to obtain not only skill per
se but information concerning IBM's confidential plans and designs and that,
indeed. Telex did obtain as a result of its efforts from Kevill and others confi-
dential IB:M engineering design information, confidential IBM planning infor-
mation, and confidential IBM pricing information was used among other things
to prepare a Midas-Iceberg cost estimate.

During August, 1972, Telex's President James was negotiating a contract with
Hitachi, a Japanese manufacturer. One of the inducements Telex proposed to
Hitachi was access to information relating to IBM's unannounced disk programs
known to Telex employees. Telex also offered to provide Hitachi with informa-
tion that would enable Hitaclii to design an equivalent to the unannounced IBM
Apollo.

In these and other ways. Telex has ■deliberately adopted a continuing policy
for penetration of IBM trade secrets and confidential information on unan-
nounced products.

F1G6. Telex's memory program began in 3970, on the basis of confidential
IBM business information, and since then Telex has attempted to obtain addi-
tional information and trade secrets of IBM. Telex launched its memory pro-
gram partly on the basis of information obtained by James from IBM confi-
dential documents.

James' recommendations were based to a substantial extent on his knowledge
of the unannounced prices and confidential forecast data relating to IBM's
NS memories which he disclosed in the June 11 memorandum to Martin.

Among the material James took with him when he left IBM was information
from registered IBM confidential documents, the SCAN forecast assumptions
and the Plan 2.5 Forecast. The px'ices disclosed in James' June 11 memorandum
for the "XS-2" CPU and its associated memory were identical with the prices
contained in the IBM SCAN forecast assumptions.

The forecast of total installed NS .s.vstems on which James' projection of
memory potential was based, was identical with the forecasts of NS installations
contained in the IBM Plan 2o Forecast. IBM employees experienced and knowl-
edgeable in IB^VI's plans and designs were sought by Telex through the offer
of exceedingly liberal salaries and bonuses, in some cases as much as a quarter
of a million dollars, if equivalent memories' could be produced.

Telex's recruitment program in the memory field were not signally successful.
While it is clear that substantial confidential information was obtained by
Telex concerning the IB^M memory program, the evidence is insufficient to per-
mit the court to quantify such information. A similar situation exists with
respec't to Telex's attempts at misappropriating IBM's trade secrets and pro-
prietary information regarding communications controllers.

F167. There is evidence in the record showing that Telex attempted to recruit,
and in several instances did recruit, IBM employees with the intent to mis-
appropriate IBM trade secrets and confidential information concerning IBM
CPU's, and that Telex has attempted, and is now attempting, in negotiations
and arran'iements with Hitachi to capitalize on its ability to obtain IBM con-
fidential information.

On November Ifi. 1972. Telex's representative Demmer, a former IB]\I em-
ployee, represented to Hitachi among other things : '"The way to optimize on
tlie benefits of liaving the IBM knowledge that exists in Tnlsa. and also the
technical resources that exist in Kanagawa, is to do the basic design of the CPU
in Tulsa until such time that the specs are frozen and the first prototype has
had some testing to verify this basic design." Joint development efforts are now
proceeding as between Telex and Hitachi.

F168. In all of the areas above indicated, where IB]M confidential and trade
secret information has been sought or utilized by Telex, the line of demarcation
between such use and legitimate utilization of the skills, knowledge, judgment
and expertise of former IBM employees is often difficult, and on occasion im-
possible, to delineate with accuracy or assurance.



5756

However, in all of the areas discussed above, it is quite clear, and the great
preponderance of the evidence shows, that Telex has had the intent to l)enefit
not only from these appropriate elements of utilization but from confidential
information and trade secrets which IBM, within the awareness of Telex, has
had the right to preserve : and. in the areas where the court has found that the
evidence is sufficiently definite to authorize awards to IBM the court believes
and finds that the preponderance of the evidence indicates that the utilization
of such confidential information and trade secrets knowingly and willingly by
Telex, without reference to the other areas of legitimate utilization of persf)nnel
formerly employed by IBM, proximately caused and authorized the damages
found herein.

X. STATUTE OF LIMITATIONS AS TO THE MISAPPROPRIATION OF CONFIDENTIAL
INFORMATION AND TRADE SECRETS

F169. Resolution of the reserved issue whether the statute of limitations has
barred IBM's counterclaims against Telex primarily depends upon the conclu-
sions of law to be set out hereinafter rather than upon facts beyond those already
found. However, additional relevant facts must be determined particularly
with reference to IBM's claim that Telex fraudulently concealed its utilization
of confidential information and trade secrets and thus tolled whatever statutes
of limitations may l)e held to be applicable.

F170. Howard Gruver was hired from IBM as hereinabove mentioned in
July. 1970. The first IBM engineer of significance hired by Telex after that time
was John Clemens, who was hired in November, 1970. IBM became concerned
that if Telex hired's number of key IBM engineers and put them to work on the
same project they worked on at IBM there would be risk of disclosure of its
trade secrets and confidential information.

Accordingly on December 3, 1970, J. D. Kuehler, Director of IBM's San Jose
laboratory, wrote to Telex President Jatras concerning Telex's hiring of Clemens.
On December 11, 1970, TCP President Martin responded on behalf of Telex to
Kuehler's letter to Jatras, and assured Kuehler that Telex intended to develop
its own technology from Telex's own, and public, sources.

^lartin also represented to Kuehler that Telex had instructed Clemens that
he was not to bring with him or retain possession of any drawings, specifica-
tions or documents belonging to IBM, and that it was not Telex's policy to
focus recruitment on any one employer.

Martin suggested it would be in the interest of both companies if Kuehler and
he were to have a meeting to discuss the matter of recruiting, trade secrets
and confidential information in more detail. During the period of time between
the letter from Kuehler to Jatras and Martin's response to Kuehler, Jatras and
Clemens were in fact actively recruiting IBM employees Kevill and Hancock
and attempting to recruit IBM employee Wilmer.

On January 5, 1971, Kuehler wrote Martin and accepted his suggestion that
a meeting be held between representatives of IBM and Telex to discuss the
potential exposure that might arise from Telex's hiring of IBM engineers to
develop a Telex system functionallj' equivalent to the system they had recently
developed for IBM.

During the period of time between the Kuehler letter setting up a meeting
with Martin and the actual meeting held on February 17. 1971. Telex success-
fully recruited Mr. Ice and Mr. Hon from IBM to work at TDAS in the develop-
ment of a Telex product similar to the 3330.

F171. On February 17, 1971, Martin met with Kuehler and R. H. Mattern, Jr.,
of IBM at the latter's Menlo Park Laboratory. During this meeting Kuehler
informed Martin that IBM was greatly concerned when a key engineer who had
))een in an IBM development program like the Merlin was recruited by a com-
l)etitor and given an assignment of developing a product compatible to that
which he had developed at IBM under the incentive of a performance bonus
for meeting extremely tight time schedules.

Kuehler stated that such action could deter "independent contributions" and
expose individuals to the necessity of using IBM ti-ade secrets and proprietary
and confidential information to realize tlie bonus.

Martin informed Kuehler that Telex had recently changed its policy con-
cerning bonuses with regard to performance based on schedules and that Telex
with one exception was now using moi-e conventional salary and stock option
plans. Martin further told Kuehler that he did not want key former IBM
engineers to use IBM's confidential information at Telex and that he had asked



5757

•engineers when they joined Telex to sign an agreement that they would not
bring Telex any documents or drawings of a confidential nature.

ilartin assured Kuehler among other things that he would make certain none
of the engineers recreated IBM trade secrets or confidential IBM information
from memory once they joined Telex. He also indicated that he believed that
the percentage of IBM employees with Telex was already too high and that he
didn't intend to recruit any more employees from IBM. Kuehler left this meeting
with the impression that Martin would follow through with the things he had
said.

F172. Contrary to Martin's representation, in April, 1971, Telex hired Neil
Glover from IBM, and in September, 1971, Telex recruited and hired Dick



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 112 of 140)