Summary
holding that disclosure of aspects of a software program to potential customers before their signing confidentiality provisions did not negate the trade secret status of the software because the portions "represent[ed] a small fraction of the whole, . . . [which was] too limited to permit an understanding of the program's design and architecture."
Summary of this case from Syncsort Inc. v. Innovative Routines International, Inc.Opinion
Civ. A. No. 82-356-Z.
February 17, 1984.
Richard L. Dahlen, Herbert L. Gatewood, Dahlen Gatewood, Boston, Mass., for plaintiff.
Michael P. DeFanti, Hinckley Allen, Providence, R.I., for Tiverton.
Wm. J. Cheeseman, Foley, Hoag Eliot, Boston, Mass., for remaining defendants.
MEMORANDUM OF DECISION
Plaintiff brought this action to redress the alleged misappropriation by defendants of its trade secrets and proprietary rights in computer software. By agreement of the parties, trial of the case was bifurcated and the only issue tried initially was the liability of all defendants except Tiverton Bottled Gas Company. At the close of plaintiff's evidence, the motion of defendant, Rehab Computer, Incorporated, was allowed for failure of plaintiff to prove that defendant's complicity in any wrongdoing. The defendants remaining are Brian J. Vaill ("Vaill") and Dealers Management Services, Inc. ("DMS"), a company formed by Vaill and of which he serves as an officer and director.
The parties agreed before trial that defendants' liability depended upon proof by plaintiff of three elements: (1) that its computer program is a trade secret; (2) that it made reasonable efforts to keep the program secret; and (3) that Vaill and an associate, Jay DeYoung, copied plaintiff's program and included substantial portions thereof in the product being sold by DMS. They stipulated that plaintiff's program has a value and that DMS sells a program in competition with plaintiff. The following shall constitute my findings of fact and conclusions of law.
Defendants also assert that plaintiff delayed so long in bringing the action as to be barred by laches. Given that the delay was occasioned by extensive settlement discussions and that the complaint was filed shortly after it was clear that settlement was not possible, I do not deem the delay improper.
In 1976, Standard Oil Company of Indiana, in an effort to diversify, decided to develop a computer program to provide accounting and management information and support to independent petroleum distributors or jobbers, many of whom had, under an earlier distribution system, been employees of the company. A subsidiary, Amoco Computer Services Company, was assigned the development task under the direction of Terry R. Beal, an Amoco Product Manager and experienced computer specialist. He started by purchasing a software package from one John Grady, a programmer. Grady's program was designed for retail fuel oil dealers and was inadequate for the needs of the jobbers or bulk dealers. Beal set about to rebuild and improve the original package and named the new program "Jobber Management System" ("JMS"). He obtained the assistance of Grady, Howard Dickerman, plaintiff's principal and at that time a 21-year-old college drop-out and computer whizz, and two others. They developed the program in direct response to the articulated needs of prospective users. Indeed, a substantial portion of the time spent by the team was expended in determining the requirements of and designing the specifications for the system. I credit Beal's testimony that the development effort consumed at least twenty-two man months until the first sale and an additional twelve months to the first installation. It cost approximately $400,000, including the price paid Grady.
Although JMS is a good and useful tool, it was not at all successful from Amoco's point of view. Sales of the system lagged far behind projections and in 1981 Amoco sold it to plaintiff for the sum of $125,000. Plaintiff, which was founded by Howard Dickerman, is engaged in the business of licensing, selling and servicing JMS.
The JMS program starts with a "Systems Options Menu," the main menu, which includes five major and five minor groupings of functional options. The five major groupings are called "File Maintenance," "Dispatching," "Posting," "Billing" and "Reports." The five minor groupings are for various housekeeping functions: "Start of Day," "End of Day," and routines for protecting information fed into the system. The major groupings determine the organization of the system and were chosen by the designers, Beal, Grady and Dickerman, from an infinite number of alternatives. (Grady's original system had 10-12 menus.) They are assigned the actual operating programs. "File Maintenance" is a constantly changing store of information of customer names, addresses, requirements, account numbers, and account status, as well as other useful details concerning customers. The "Dispatching" submenu is used to schedule deliveries in advance and to respond to customer requests for immediate delivery. For both contingencies, it prints delivery tickets which are used by the driver as his instructions. "Posting" is designed to keep the information in the computer current and to enter the various business transactions as they occur. The fourth grouping, "Billing," is the smallest. It is used primarily in connection with the preparation of monthly bills and is able to add finance charges or rental fees; it computes sales taxes, and it permits easy management of budget accounts. The "Reports" submenu is the essential management tool. It is designed to produce periodically or "on request" various information, such as reports, showing accounts receivable by age, tax reports, and sales analyses. It can produce these by division of a company, by product, alphabetically, by type of account, or numerically.
The manner in which JMS receives information, processes it, and then produces new data is unique. For example, the "Daily Operating Control Report" has ten or twelve different sections: sales, accounts receivable, activities of trucks, etc. All are produced by the operator simply requesting the computer to "Process All Data," which causes the computer to scan all relevant files. The "File Maintenance" submenu has separate programs for adding, changing or deleting customer records and includes separate files for delivery and billing addresses. The "Posting" submenu separates the entering and processing of transactions, a decision made because the early computers worked relatively slowly. It also divided the transactions to be posted into five categories — "Liquid" and/or "Package Deliveries," "Miscellaneous Debits" and/or "Credits," and "Enter Payments" — although a different number and arrangement was equally feasible. The program displays running totals during the posting process, a feature not necessarily required. The "System Reports" submenu not only permits the generation of a wide range of reports, from customer listings to sales analyses and budget projections, but is programmed to produce mailing labels and Rolodex cards.
When a dealer acquires a new customer, information about him is added into the computer by means of a specifically designed series of questions. A query about the customer's "zone" is designed to assist the driver in finding the customer's home, as is routing information requested. The program includes questions about the customer's needs — the primary product to be furnished, the type of equipment, and the number of trucks the customer has, and at what intervals deliveries are to be made. It has room for additional information, such as the location of the filler pipe or warnings about dogs. The system moreover includes a field whereby a customer may be assigned to a particular division.
The customer display screen demonstrates numerous design decisions. While the inclusion of certain information, name, address, billing and payment data is self evident, it is less imperative that this screen should, for example, show year to date delivery and payment data, or that it be arranged in the precise manner of the JMS.
In addition to choosing the particular five major groupings for the purposes outlined, the designers thus made decisions concerning the specific manner in which these submenus are to be used, how they are to be accessed, how many screens to use, and how to arrange the information on each screen. It is the organization of the JMS program, with its five major groupings, combined with the particular features within the system and the procedures to be employed in its use, which plaintiff claims as its trade secret. The system is complex and is unique.
Defendant Vaill had no formal training in designing computer programs or as a programmer, but he had gained considerable experience in systems design while employed at the Providence Gas Company from 1966 to 1981. During his tenure at the Providence Gas Company, a public utility which supplied natural gas, Vaill had been involved in working on routing systems for meter readers, a customer accounting system, and an appliance inventory system. He had also assisted various computer companies in translating his employer's needs into programs. He had started there as a clerical employee, advanced to the position of management trainee, and by the late 70s had worked himself up to general manager of Tiverton Bottled Gas Co., ("Tiverton") a then newly acquired subsidiary.
Tiverton, an LP-gas dealer, had at the time of the acquisition a manual accounting system. Vaill investigated and considered a number of computerized systems and eventually settled on JMS. In February 1980, Tiverton contracted to purchase JMS, and in April of that year it was installed. Although Vaill did not sign the contract for Tiverton, it is undisputed that the sale was made subject to an agreement by Tiverton to "treat as confidential all programs, documents and other information relating to JMS. . . ." Indeed, all sales of JMS were made under similar agreements. I find that Vaill was aware of the confidentiality clause and of plaintiff's concern to keep the design of JMS secret.
Vaill found JMS an easy system to learn and soon became intimately familiar with it. At some time in 1980, pursuant to an agreement with Howard Dickerman, he even assisted in demonstrating JMS to other potential buyers and received a commission from plaintiff for sales he had facilitated. During demonstrations of the system by Vaill or Dickerman, a prospective customer necessarily learns some aspects of the program. Plaintiff did not restrict Vaill in the manner in which he carried out such of the demonstrations as he conducted, nor did it extract confidentiality agreements from prospective customers. Nevertheless, the portions of JMS disclosed during a demonstration represent a small fraction of the whole, and I credit Dickerman's testimony that they are too limited to permit an understanding of the program's design and architecture. Prospective customers at such demonstrations do not see the entire menu and all of the submenus, nor do they see any user's guide or manual. Whenever anyone evinces a serious interest, plaintiff requires a written undertaking to keep confidential anything disclosed before permitting a more detailed review of the program and a look at the user's guide.
In the course of his work at Providence Gas Company and later at Tiverton, Vaill had become acquainted with Jay DeYoung, who was then a computer programmer employed by Rehab Computer Incorporated, a data processing concern based in Washington, D.C. They had cooperated on one project for Providence Gas, and in the fall of 1979, Vaill requested DeYoung to estimate the cost of designing a program for LP-gas dealers and particularly Tiverton. Although they did not then go forward with the latter project, they recommenced discussions about developing a software package for general distribution in October 1980. Vaill went to Washington, D.C. for a meeting with DeYoung and Messrs. McCoy and Quigg, two officers of Rehab whom Vaill sought to enlist in the enterprise. The presentation to McCoy and Quigg said very little about the design and documentation of the proposed package, but focused almost entirely on marketing. No commitments were made, but after additional discussions between Vaill and DeYoung, the latter came to Providence in February 1981 to work on a system.
DeYoung worked at Tiverton for two weeks in February, one week in March, and one week in April, and utilized the Wang computer which belonged to Tiverton. He denied having used the JMS manual. I do not credit that testimony. By the end of February, Vaill and DeYoung had documented sixteen files, including a "Customer Master File" in nearly final form. By that time they had also developed all forty-nine fields which constitute the final version of their program. DeYoung managed all of that even though he worked on the computer for no more than three to four weeks and admittedly did not utilize the Wang utility system, called "IDEAS," at least for the "Customer Master File."
A "file" is the physical area on a magnetic disc where information is stored. A "record" is data in a file which relates to a particular subject, and a "field" is a specific area within a record for storing particular information concerning the subject. If the file is analogous to a telephone book, the record is all the information pertaining to one name and the field, each piece of such information, such as the telephone number or the address.
Although defendants claimed to have incurred substantial "development costs," the evidence is devoid of specific figures. The testimony suggests, moreover, that "development costs" included not so much amounts expended on developing a program but rather amounts spent on developing a market for FuelPak. In late March, Vaill and DeYoung caused the training materials to be erased from one of the JMS disks and used it for their system which they then exhibited at a trade show in early April. They "borrowed" a second disk, a blank one from Tiverton, to make another copy of their program for use at that and other trade shows. These disks were returned to Tiverton sometime in June only after Tiverton demanded their return. Vaill and DeYoung ceased having access to JMS materials, except for the two disks, in early May when Tiverton fired Vaill for his activities in connection with the development of FuelPak.
DeYoung testified that from October 1980 on he made notes about the design of a program, most of which he later destroyed. He did, however, keep some papers, including an outline, dated February 17, 1981, of a customer master file. That document, together with another outline, dated February 19, 1981, which shows numerous items to be included, is remarkably similar to the JMS customer file. Other notes of Vaill, DeYoung and their assistant, Gary Hannon, show a tracking of the JMS "Transaction Posting" screen and the "Add-A-Customer" program. Some of the titles of particular files were changed but the resemblances disclosed by the notes, and others mentioned below, permit the inference, which I draw, that Vaill and DeYoung copied substantial portions of their system from JMS.
Certain of the features of both programs are clearly dictated by the needs of the businesses for which the programs are designed. Fuel oil dealers must have information about their customers' addresses, the status of their accounts, and their requirements for fuel oil, and other products. They must have the ability to predict customer requirements and to route deliveries efficiently. However, these broad needs do not predetermine the precise manner in which a computer system has to satisfy them, and the striking aspect of this case is the similarity of approach to these not uncommon problems. For example, both systems include five major groupings which performed substantially similar functions. The "Transaction Posting" submenu of both uses the same five batch heads. The fields are listed across the screen in the same order, although JMS has one field which is not included in FuelPak and the total is displayed differently. In both, the entry of data is separate from processing data, although they could have been combined. In neither can batches be closed explicitly, rather they are closed by starting a new batch. The "Add-A-Customer" program of both systems includes a field whereby the customer is assigned to a division of the company. Both have three fields concerning deliveries, "automatic," "will call," and "Julian," and where JMS uses for its delivery code numerical designations 1, 2 and 3, FuelPak uses a similar sequential code, 11, 22 and 33. Although the manner of daily closing is somewhat different on the two systems, both produce the same six reports. I credit the testimony of plaintiff's expert, Professor Robert Kinicki, that these and other similarities are not accidental and that certain cosmetic differences were incorporated in FuelPak to disguise copying. I further credit his testimony and find that the time needed to produce a system like FuelPak far exceeds the time, in fact, expended on development by Vaill, DeYoung and their assistants, even taking into account Vaill's familiarity with the fuel oil business.
Trade secret protection may be afforded to any idea, process, or compilation of information valuable and useful in one's business and which is not generally known. A trade secret is any special knowledge developed through skill and ingenuity and the expenditure of money and effort which, by being secret, gives the owner an advantage over his competitors. Atlantic Wool Combing Co. v. Norfolk Mills, Inc., 357 F.2d 866 (1st Cir. 1966); Rapco Foam, Inc. v. Scientific Applications, 479 F.Supp. 1027 (S.D.N.Y. 1979). In determining whether a particular idea is entitled to protection, the Court should consider the degree of secrecy surrounding it, the efforts expended in developing it and preserving its secrecy, its value to plaintiff and defendant, and the difficulty of duplicating it. Rapco Foam, Inc. v. Scientific Applications, 479 F.Supp. at 1029. Thus, plaintiff need not prove that its system is novel in the patent law sense; Cataphote Corporation v. Hudson, 444 F.2d 1313 (5th Cir. 1971). Plaintiff need only show that the particular architecture of its program is valuable, that it is not a matter of common knowledge or readily duplicated, and that it was developed and has been kept secret through plaintiff's efforts.
Plaintiff has met that burden. JMS reflects a series of design decisions which resulted in a complex but coherent program for a specific market. The program is neither random nor self evident and was developed at substantial expense. While, as noted earlier, some of the screens are mandated by the needs of the market, the particular combination of procedures used in plaintiff's system, and the particular features within the system detailed earlier, are neither obvious nor easily duplicated. They constitute a trade secret. Those features differentiated plaintiff's system from other computer programs and they gave plaintiff such competitive advantage as it had.
Such difficulties as exist in applying the general definitions and principles of the law of trade secrets to computer software are particularly apparent when considering the adequacy of plaintiff's efforts to ensure the secrecy of its program. Both parties demonstrated the relative ease with which one knowledgeable in the science or, as the case may be, art of computers can discern not only the general outline of a program but also its source code. Nevertheless, I conclude that plaintiff took adequate precautions to protect the secrecy of its product.
Based upon the similarities of the two programs detailed earlier, DeYoung's notes, the availability of the JMS manual to Vaill and DeYoung, their use of it, and the speed with which they assembled their system, the conclusion that they copied substantial portions of JMS is inescapable. Vaill does not seriously contest that he was bound by the confidentiality agreement of Tiverton. I find that he did violate that confidential relationship.
Accordingly, plaintiff's motion for the entry of a finding that defendants Vaill and Dealer Management Services, Inc. are liable to plaintiff for misappropriation of its trade secrets is allowed. Plaintiff shall submit a proposed form of injunction.