(dissenting).
The District Judge has held the defendant liable for seven million dollars in damages because, he said, defendant breached an implied promise to perfect and market plaintiff’s invention. Because I am convinced that both the decision and the manner in which it was reached were clearly erroneous, I dissent.
Although there has been ample judicial discussion of the difference between obligations implied in fact and those implied in law,1 this distinction remains blurred in many decisions. Regardless of terminology, however, one thing is clear. An obligation, implied either in fact or in law, is fashioned from the facts and circumstances surrounding the dealings of the parties; it does not, like Minerva, spring full blown from the head of the trial judge. Implied promises are to be cautiously and not hastily raised, Genet v. Delaware & Hudson Canal Co., 136 N.Y. 593, 608, 32 N.E. 1078 (1893). They are strictly construed, 14 Castner, Business Organizations — Patent Licensing Transactions § 2.09 (1975), and the plaintiff has a heavy burden in convincing the court of their existence. 19 Cornell L.Q. 603, 606 (1934).
Where the obligation asserted is not expressly stated in the contract, “[i]f it is supportable at all, it must be from some implication arising out of the situation and *117acts of the parties, and the circumstances surrounding the property, at the time of the [contract], which is by the obligation of some principle of justice and equity, required to be made.” Dermott v. State, 99 N.Y. 101, 109, 1 N.E. 242, 245 (1885); 3 Corbin on Contracts §§ 562-564 (1960). “[W]here intent, though obscure, is nevertheless discernible, it must be followed. . ” Parev Products Co. v. I. Rokeach & Sons, Inc., 124 F.2d 147, 149 (2d Cir. 1941). In short, before any implied obligation is found to exist, the facts surrounding the making of the written contract should be fully explored. Reback v. Story Productions, Inc., 9 App.Div.2d 880, 193 N.Y.S.2d 520 (1st Dep’t 1959).
In the instant case, the District Judge decided that the defendant had an implied obligation to perfect plaintiff’s non-skid device before he heard one word of testimony. On September 28, 1973, pursuant to Fed.R. Civ.P. 16, he signed an order specifying the issues to be tried. The fundamental issue on liability, as he framed it, was:
Before it abandoned the contract dated December 21, 1964, did defendant use its best efforts for a reasonable period of time in collaboration with plaintiff to perfect the product so as to make it marketable?
With the entry of this order, the battle was half over, although not a single shot had yet been fired. Insofar as the District Judge was concerned, there was no question as to the existence and scope of an implied obligation, only whether the obligation as he had already found it to exist had been performed.
A pretrial order supersedes the pleadings and becomes the governing pattern of the law suit. Case v. Abrams, 352 F.2d 193,195 (10th Cir. 1965). For this reason, a pretrial conference should not serve as a substitute for the regular trial of the case; and pretrial orders should not be used to resolve factual issues properly determinable on trial. Lynn v. Smith, 281 F.2d 501, 506 (3d Cir. 1960). The procedure followed in this case was, I am convinced, clearly and prejudiciously erroneous.
The District Judge stated in his opinion that, while he had defined the issues at the start of the trial, he had permitted “extraordinary latitude” to the defense to prove all that it could. I find this statement to be without support in the record. The District Judge stated repeatedly that he had signed a pretrial order setting forth the issues; that “these are the issues that we are trying here”; that “I am trying to keep this case within the issues as defined and I intend to do just that”; that “I signed an order trying to delineate the issues and these are the issues we are going to try in this case”; that “we are going to try the issues that were set forth in that pretrial order.” He also told defense counsel that if counsel would like to argue that his signing the pretrial order was improper, he would “have to go to our friends on the seventeenth floor to get that decision”.2
The District Judge sustained objections to question after question by means of which defense counsel sought to establish the parties’ mutual belief and understanding that plaintiff’s non-skid device was perfected, marketablé and fail-safe at the time the contract was signed. In rebuffing the attempts of defendant to establish its belief in and reliance upon plaintiff’s representations to this effect, the District Judge’s expressed reaction was, “It couldn’t matter less.” I believe that it should have mattered and that the District Court’s exclusion of this testimony was prejudicially erroneous. Where a party’s intent or belief is material to the issues in a lawsuit, he should be permitted to testify concerning it. Noonan v. Luther, 206 N.Y. 105, 108, 99 N.E. 178 (1912); Bayliss v. Cockcroft, 81 N.Y. 363, 371 (1880); McKown v. Hunter, 30 N.Y. 625 (1864); Sowalsky v. E. F. MacDonald Stamp Co., 31 App.Div.2d 582, 294 N.Y.S.2d 1016 (3d Dep’t 1968); Epstein v. Cuba, 25 App.Div.2d 680, 268 N.Y.S.2d 947 (2d Dep’t 1966); Richardson on Evidence § 384(1) (9th Ed. 1964); 2 Wigmore, Evidence § 581 (3d Ed. 1940). “To shut out the *118light furnished by the parties themselves— to read their words not as they meant them but as they appear when denuded of that meaning — is to decide an unreal, fictitious, hypothetical case.” United States v. Lennox Metal Mfg. Co., 225 F.2d 302, 313 (2d Cir. 1955).
Having decided the issue in advance of trial and thereafter having hamstrung the efforts of the defense, it was a meaningless gesture for the District Judge to “find” at the conclusion of the trial that defendant had an implied obligation to perfect plaintiff’s invention. As he stated in his opinion, he found for plaintiff “on the claim that was tried”.
Even if it be assumed that the procedure followed by the District Judge was not prejudicially improper, I am convinced, nonetheless, that his “finding” was clearly erroneous. A patentee has the exclusive right to make, use and sell his invention and may transfer all or part of this right to another. Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 135, 89 S.Ct. 1562, 23 L.Ed.2d 129 (1969). Where the right has been transferred, courts have in some instances implied in the transferee an obligation to exercise it. However, the obligation thus implied has been to “make, use and sell”, nothing more. It has been defined variously as an implied promise to “work the patent”, Driver-Harris Co. v. Industrial Furnace Corp., 12 F.Supp. 918, 919 (W.D.N.Y.1935); “work the patent in good faith to make it produce royalty income”, Mechanical Ice Tray Corp. v. General Motors Corp., 144 F.2d 720, 725 (2d Cir. 1944), cert. denied, 324 U.S. 844, 65 S.Ct. 679, 89 L.Ed. 1406 (1945); “manufacture and sell”, Eastern Electric, Inc. v. Seeburg Corp., 310 F.Supp. 1126, 1151 (S.D.N.Y.1969), aff’d, 427 F.2d 23 (2d Cir. 1970); “manufacture and market”, Perma Research and Development Co. v. Singer Co., 410 F.2d 572, 574 (2d Cir. 1969); and “exploit”, Bellows v. E. R. Squibb & Sons, Inc., 359 F.Supp. 204 (N.D. Ill.1973).
While the phrase “best efforts” is often used to describe the extent of the implied undertaking, this has properly been termed an “extravagant” phrase, Becker, Licensing in the Chemical Industry, 55 J.Pat.Off.Soc’y 759, 774 (1973); and it should not be literally interpreted. A more accurate description of the obligation owed would be the exercise of “due diligence”, Vacuum Concrete Corp. of America v. American Machine & Foundry Co., 321 F.Supp. 771, 775 (S.D.N.Y. 1971) , or “good faith”. Mechanical Ice Tray Corp. v. General Motors Corp., supra, 144 F.2d at 725. Thus, courts have not required the transferee to enter into a hopeless contest, Eclipse Bicycle Co. v. Farrow, 199 U.S. 581, 26 S.Ct. 150, 50 L.Ed. 317 (1905), to resort to “any means whatever” to produce royalty for the transferor, Tibbetts Contracting Corp. v. O & E Contracting Co., 15 N.Y.2d 324, 337, 258 N.Y.S.2d 400, 206 N.E.2d 340 (1965), or to manufacture and sell “come what may”. Parev Products Co. v. I. Rokeach & Sons, Inc., supra, 124 F.2d at 150; United States v. Combustion Engineering, Inc., 364 F.Supp. 181, 186 (D.Conn. 1972) . He has not been compelled to manufacture and sell without profit, Tesra Co. v. Holland Furnace Co., 73 F.2d 553, 555 (6th Cir. 1934); HML Corp. v. General Foods Corp., 365 F.2d 77 (3d Cir. 1966), or to attempt to sell an unmarketable product. Bull v. Logetronics, Inc., 323 F.Supp. 115, 132 (E.D.Va.1971). I know of no case, with facts similar to those of the instant one, in which a licensee or assignee was found to have impliedly promised to perfect a defective product in order to make it marketable.3 For the reasons which follow, I do not believe that such implication should have been made in this case.
INTENT OF THE PARTIES
In determining whether defendant had an implied obligation to perfect plaintiff’s invention, a controlling factor should be the intention of the parties. Parev Products *119Co. v. I. Rokeach & Sons, Inc., supra. A contract should not be implied against the intention and understanding of the parties. Adams & Co. Real Estate v. E & B Super Markets, Inc., 26 App.Div.2d 365, 274 N.Y. S.2d 776 (1st Dept.1966).4 I think it irrefutable that if plaintiff was convinced that its non-skid device was marketable and fail-safe, it could not have intended that defendant was going to perfect it so as to make it marketable and fail-safe. The proof is clear and uncontradicted that from the first meeting between representatives of plaintiff and defendant until the last day of trial, plaintiff considered its anti-skid apparatus to be both marketable and fail-safe.
In 1963, before it had any contact with defendant, plaintiff arranged to have its device manufactured by Worcester Stamp Metal Company; and Worcester expended over one million dollars for tooling and inventory in preparation for performance. However, after plaintiff had marketed approximately 3,500 units manufactured by Worcester, it became dissatisfied with Worcester’s quality control; and, when production was interrupted by a strike, plaintiff was in the market for a new manufacturer. It was at this time that defendant entered the picture. During negotiations which followed, plaintiff repeatedly assured defendant that its device was both perfected and fail-safe. . These assurances were made orally, in writing and in a promotional film. Plaintiff has never deviated from this position.
When defendant quite properly determined that plaintiff’s device was neither perfected nor fail-safe and ceased performance,5 plaintiff sued to rescind the contract. In its complaint, it alleged that defendant was not bound to perform any covenants or agreements and retained complete discretion as to marketing and manufacturing. In its reply to defendant’s counterclaim for fraud and misrepresentation, plaintiff alleged that defendant was obligated to “expend time and money in the manufacture, distribution and sale, testing, recalling and retrieving” of its product. Nowhere was there any mention of “perfecting”.
When defendant moved for summary judgment before Judge MacMahon, plaintiff’s president, the inventor of the unit, swore that in his opinion it was fail-safe. In plaintiff’s statement submitted pursuant to S.D.N.Y. R. 9(g), it asserted that the unit was fail-safe and was in no other way unsuitable for commercial use. It stated:
As plaintiff has said repeatedly, the product at all times was ready for mass production and was marketable.
At the trial, plaintiff’s president testified that the product was fail-safe, was perfected, was acceptable and doing what it was intended to do and was ready to be put into mass production. In light of plaintiff’s undeviating position on this issue, the District Court’s statement that both parties knew the device was not perfected and was unmarketable was clearly erroneous and without support in the proof.
The District Judge categorized plaintiff’s assertions as “puffing”, and concluded that defendant could not have relied upon them. If they were “puffing”, it is clear that plaintiff “puffed” its way right through the lawsuit, because it never changed its position. Moreover, there is ample proof that defendant did rely on plaintiff’s statements. For example, when the contract was entered into in December 1964, defendant assumed plaintiff’s obligations under five distributor contracts which called for delivery of over 50,000 units in 1965; and, within a few weeks thereafter defendant agreed to sell Monitor Enterprises, Inc. a minimum of 50,000 units, delivery also to be made in 1965.
However, reliance by defendant is not the issue with which we should be concerned; rather, it is plaintiff’s own belief in the *120truth of its statements. If plaintiff believed that its device was perfected, fail-safe and marketable, and it is clear that it did, it could not have intended, relied upon or even contemplated an implied promise by defendant that it would make it fail-safe and marketable. For the District Judge to now give plaintiff the benefit of such an implied promise to the tune of seven million dollars, is, in my opinion, a clear miscarriage of justice.
THE WRITTEN CONTRACT
The contract between the parties was not scribbled on the back of an old envelope. It was a “detailed written Agreement reached as the result of skilled legal negotiation.” See Vacuum Concrete Corp. of America v. American Machine & Foundry Co., supra, 321 F.Supp. at 775. Not only was there no reference to “perfecting” the device in the written agreement; there was no discussion of perfecting in the negotiations which preceded it. Moreover, there was no “grant-back” provision, as is customarily found in contracts calling for experimentation and development.6
The contract provided that defendant “in its absolute discretion shall determine the method of manufacturing, exploiting and marketing the Product” but gave plaintiff the right to reacquire its device if defendant failed to spend at least one hundred thousand dollars for “marketing, promoting and advertising” in any year beginning with 1966.
It provided further that, in addition to specified royalties, defendant would pay plaintiff twenty-four thousand dollars in cash, would purchase tools and parts for approximately seventy-nine thousand dollars, would assume payment of certain obligations of the plaintiff approximating two hundred ten thousand dollars and would take over five distributor contracts to which plaintiff had already committed itself and which called for delivery of over 50,000 units in 1965. These payments, of course, were in addition to the one and one-half million dollars that defendant had already expended for purchase of parts and tooling from Worcester Metal. Finally, the written agreement stated that it contained “the entire understanding of the parties” and that “there are no restrictions, promises, warranties, covenants, or undertakings other than those expressly set forth herein.” It is in the light of these provisions that the trial court’s finding of an implied duty to perfect must be weighed.
Where a formal written contract results from elaborate negotiations between the parties and their counsel and contains an explicit statement that it incorporates the complete agreement of the parties, additional undertakings should not be implied except to rectify a manifest injustice. HML Corporation v. General Food Corp., supra; Vacuum Concrete Corp. v. American Machine & Foundry Co., supra; Eastern Electric, Inc. v. Seeburg Corp., supra. With the execution of the contract herein, defendant had approximately two million dollars invested in a device which, despite plaintiff’s representations to the contrary, was unmarketable and dangerous. Implication of an obligation to perfect this device, in direct contravention of the terms of the written agreement, exacerbates, rather than rectifies, an injustice.
The District Judge’s statement that plaintiff’s device could have been perfected “without unreasonable cost or effort” is completely without support in the record. Defendant did not believe that the device could be perfected. The process which plaintiff’s expert opined would lead to perfection required at least a fifteen month program of experimentation and development at a cost in excess of seven hundred fifty thousand dollars. Inequity would be heaped upon inequity if, after defendant had spent several years in attempting to perfect the device, plaintiff then exercised *121its right of reversion because defendant had failed to spend at least one hundred thousand dollars per year for “marketing, promoting and advertising”.
The technical assistance contract, upon which the District Judge relied heavily in finding an obligation to perfect, is a very common adjunct to licensing agreements. Contracts of this nature are intended as an aid to the licensee “in establishing better methods of operation and overcoming day-to-day difficulties”. Vukowich, Implied Warranties in Patent, Know-How and Technical Assistance Licensing Agreements, 50 J.Pat.Off.Soc’y 307, 337 (1968). These are “commonly included with know-how and patent licenses so that the licensed invention or information can be adapted to the licensee’s particular needs and so that the licensee can be assured that it is propitiously exploited”. Id. The six months’ technical assistance contract in the instant case in no way implied an obligation on the part of the defendant to perfect a device which proved to be unmarketable and not fail-safe.7
THE SEVEN MILLION DOLLAR COMPUTER PROGRAM
Plaintiff offered the testimony of two expert witnesses in support of its claim that its anti-skid device could be made workable and fail-safe. The first of these witnesses based his opinion upon experimental data supplied by the second, and the second secured all his data from some simulated formulas which he fed into a computer. The information upon which these formulas were based was hearsay, having been secured from still a third person. It was conceded by the witnesses that their conclusions were based upon the computer output. Indeed, neither of plaintiff’s experts had ever seen plaintiff’s device in operation or had tested it.8
As one of the many who have received computerized bills and dunning letters for accounts long since paid, I am not prepared to accept the product of a computer as the equivalent of Holy Writ. Neither should a District Judge. To be marketable, plaintiff’s non-skid device had to function on cars of all types and size, on rough roads and smooth roads, in rainy weather and in dry weather, in the winter and in the summer, at high altitudes and low altitudes, and with wheels, tires and braking systems of all kinds, in fact, under all conceivable conditions. In addition, under all conditions it had to be fail-safe.9 Proof that plaintiff’s defective and dangerous device could be transformed in some way to accomplish all this required more than the running of some figures through a computer which involved calculation of a theoretical series of stops, following the assumed installation of a hypothetical or simulated anti-skid device on a 1968 Thunderbird.
Authorities in the field of computer research acknowledge that simulation is “essentially an experimental problem-solving technique”. Gordon, System Simulation 18 (1969). “Simulation is ‘make-believe’ — it’s *122a game — but it should have some solid relationship with the real world.” Favret, Introduction to Digital Computer Applications 122 (1965). It is the “construction and manipulation of a model of a real world reference system by utilizing theoretical simplifications and assumptions”, Computer Simulation and Gaming: An Interdisciplinary Survey with a View Toward Legal Applications, 24 Stan.L.Rev. 712 (1972). A computer model is valid “only insofar as it enables us to make valid inferences about the real-world system being simulated”. Gonzales and McMillan, Machine Computations: An Algorithmic Approach 212 (1971).10
Plaintiff’s expert testified that the computer would not simulate everything that one finds in a vehicle.11 The District Judge himself stated that it was very obvious that plaintiff’s experts did not include all things in the simulation and categorized the testimony of these experts as “an area of pure physics and delightful math”. When defense counsel questioned one of the experts about the specific functioning of plaintiff’s device, the District Judge said, “He has already testified that he did no experiments directly on the thing. How could he know? He is in the realm of theory.”
Both plaintiff’s and defendant’s experts agreed that simulations and development programs must go hand in hand; that the generally accepted procedure followed in a computer simulation is to validate or test the simulated device against a prototype or breadboard. Plaintiff’s expert said, “The only way one is going to tie it down and find out exactly what the thing is is to go out and test it.”
While the “reduction to practice” test commonly used in patent infringement cases is not completely analogous, I believe it is one which could properly be applied here.12 These cases hold that tests under service conditions are required whenever the conditions which will be met in actual use have not been duplicated in the laboratory. Farrand Optical Co. v. United States, 325 F.2d 328 (2d Cir. 1963). Among the devices for which tests under actual working conditions have been required are automobile sparkplugs, Payne v. Hurley, 71 F.2d 208 (Cust.Ct. & Pat.App.1934), an automobile bumper, Fageol v. Lyon, 290 F. 336 (D.C.Cir. 1923), a de-icing device for an airplane windshield, Gaiser v. Linder, 253 F.2d 433, 45 CCPA 846 (1958), a radar device for ships and airplanes, Radio Corp. of America v. International Std. Elec. Corp., 232 F.2d 726 (3d Cir. 1956), an airplane carburetor, Chandler v. Mock, 202 F.2d 755, 40 CCPA 846 (1953), and a fuel pump for an airplane, Burns v. Curtis, 172 F.2d 588, 36 CCPA 860 (1949).
Judge Learned Hand said in Sinko Tool & Mfg. Co. v. Automatic Devices Corp., 157 F.2d 974, 977 (2d Cir. 1946), that tests under service conditions are necessary whenever qualified persons would require such a test before they were willing to manufacture and sell the invention as it stands. Plaintiff’s own expert testified that its anti-skid device was presently unmarketable, that “feasibility” studies involving the testing of actual prototypes on automobiles would be required and that “the whole program would terminate if the feasibility studies showed that it wasn’t feasible.” Despite all this, the District Judge “bought” the hypo*123thetical, untested and unproven simulation of plaintiff’s experts — using, of course, defendant’s seven million dollars.
If the District Judge erred in accepting the testimony of plaintiff’s experts that its device could be made workable, he went even further afield in finding that it could be made fail-safe. The computer simulations dealt only with stopping distances and had nothing to do with the fail-safe characteristics of plaintiff’s device. Plaintiff’s expert testified that, after a six month feasibility study of computer simulations, it would be necessary to build a prototype which would then be subjected to a five month “failure analysis” program using new and different computer programming. This analysis of the device’s fail-safe characteristics would, the witness said, be “an extremely critical thing”, requiring decisions as to what components needed to be completely redesigned.13 Sans feasibility study, sans prototype, sans failure analysis program, sans fail-safe computer programming, sans the making of any decisions on redesign, the District Judge accepted the opinion of plaintiff’s expert that plaintiff’s anti-skid device could be made fail-safe. This was error.
Experts are not exempt from the rules of evidence, McFarland v. Gregory, 425 F.2d 443, 448 (2d Cir. 1970), and their testimony must be based upon established facts. Perry v. Allegheny Airlines, Inc., 489 F.2d 1349, 1353 (2d Cir. 1974); Cassano v. Hagstrom, 5 N.Y.2d 643, 646, 187 N.Y.S.2d 1, 159 N.E.2d 348 (1959); S. De Lia Construction Corp. v. Green Island Contracting Corp., 46 App. Div.2d 970, 972, 362 N.Y.S.2d 584 (3d Dep’t 1974); Stracher v. Corning Glass Works, 39 App.Div.2d 560, 561, 331 N.Y.S.2d 764 (2d Dep’t 1972). It may not be based upon unsupported assumptions and opinions, Di Filippo v. Gargiulo, 278 App.Div. 172, 104 N.Y.S.2d 149 (1st Dep’t 1951); Franzese v. Mackay Trucking Corp., 10 App.Div.2d 713, 199 N.Y.S.2d 514 (2d Dep’t 1960), guess, speculation or conjecture, Craig v. Champlin Petroleum Co., 435 F.2d 933, 937 (10th Cir. 1971); Kale v. Douthitt, 274 F.2d 476, 482 (4th Cir. 1960). A naked opinion based upon speculation or conjecture “does not rise to the dignity of evidence.” Atlantic Life Ins. Co. v. Vaughan, 71 F.2d 394, 395 (6th Cir.), cert. denied, 293 U.S. 589, 55 S.Ct. 104, 79 L.Ed. 684 (1934).
It goes without saying that the findings of the trial judge cannot rise above the legitimate testimony of the expert witness. “Speculation cannot supply the place of proof.” Moore v. Chesapeake & Ohio R. R. Co., 340 U.S. 573, 578, 71 S.Ct. 428, 430 (1951). See Galloway v. United States, 319 U.S. 372, 387, 395, 63 S.Ct. 1077, 87 L.Ed. 1458 (1943). Testimony that an undescribed, hypothetical anti-skid device can be made fail-safe in some undescribed, hypothetical manner is pure speculation and does not establish a foundation upon which to base a seven million dollar verdict.
The District Judge’s error was compounded by his rejection of defendant’s definition of fail-safe and the substitution of some lesser standard which he never defined. Singer’s position, as described by the District Judge, was that it would not market a product which “could leave the purchaser . less safe than if he hadn’t elected to purchase it.” This definition warranted approbation, not the ridicule accorded it by the District Judge.14
In contrast to defendant’s laudable standard, the criteria for reliability of plaintiff’s expert, which the District Judge presumably accepted, was that the device “had to function without failures within warranty periods or at least, with a good knowledge of statistical probability, that had to be quite low, within a warranty period of a car.” [Emphasis supplied]. This witness’ *124definition of fail-safe, which is a classic illustration of gobbledygook, was as follows:
Well, inasmuch as a main function of skid control is to displace and reapply brake fluid, or in other words to release and reapply brakes, a fail-safe criteria to monitor this function was established, and it was established in a way that if such function exceeded an acceptable or reasonable period of time, then the system is to bypass itself.
The Federal Motor Vehicle Safety Standards promulgated pursuant to the National Traffic and Motor Vehicle Safety Act of 1966, 15 U.S.C. § 1381 et seq., provide that, in the event of a rupture or leakage type of failure in a service brake system, the remaining portions of the system shall continue to operate and stop the vehicle within certain prescribed distances. 49 C.F.R. § 571.105-75, S 5.1.2. They then provide that, in the event of failure in an anti-skid device, “structural or functional”, the vehicle shall be capable of meeting the stopping distance requirements of S 5.1.2. 49 C.F.R. § 571.105-75, S 5.5.
The standard set forth in these regulations is, I submit, exactly the standard prescribed by the defendant, i. e., one that “does not impair the utility of the underlying system to which it is attached”. It is not an impossible standard as described by the District Judge, but one upon which any member of the motoring public should have the right to rely. The District Judge’s acceptance of a less rigorous definition of fail-safe, a definition which plays fast and loose with the lives of prospective purchasers of plaintiff’s supposed safety device, vitiates his entire opinion concerning the feasibility of perfection. Perfection, if defined with sufficient latitude and imprecision, is readily attainable. I think, however, that both the defendant and the motoring public were entitled to more thoughtful consideration of the standards of safety than was given them by the District Judge.
EVIDENTIARY SUPPORT FOR THE COMPUTERIZED SIMULATIONS
Because plaintiff’s case against the defendant rested almost entirely upon the results of computerized simulations, the accuracy and reliability of such computerization was a pivotal issue in the law suit. The District Judge recognized this at the outset when he said that the opinion of plaintiff’s expert would be “thrown out” if not properly connected. The Judge promised that, when plaintiff’s computer expert was produced, “we will find out where [the figures] came from, in the sense of what mathematical formulas, what type of programming was used, and so on and so forth.”
However, when plaintiffs computer expert finally testified, he refused to disclose the manner in which the computer was programmed to create the evidence used on trial, on the ground that it was his “private work product” and was proprietary information.15 Despite the assurance of defense counsel that defendant would recognize the confidentiality of such information except for purposes of the litigation, the District Judge upheld plaintiff’s refusal to permit defense counsel to examine the programs. This ruling, which required defense counsel to attack the credibility of plaintiff’s proof with one hand tied behind his back., was prejudicially erroneous.
As courts are drawn willy-nilly into the magic world of computerization, it is of utmost importance that appropriate standards be set for the introduction of computerized evidence. Statements like those of the District Judge that a computer is “but calculators [sic] with a giant ‘memory’ and the simulations the computer produces are but the solutions to mathematical equations in a ‘logical’ order” represent an overly-simplified approach to the problem of computerized proof which should not receive this Court’s approval.
*125Although the computer has tremendous potential for improving our system of justice by generating more meaningful evidence than was previously available, it presents a real danger of being the vehicle of introducing erroneous, misleading, or unreliable evidence. The possibility of an undetected error in computer-generated evidence is a function of many factors: the underlying data may be hearsay; errors may be introduced in any one of several stages of processing; the computer might be erroneously programmed, programmed to permit an error to go undetected, or programmed to introduce error into the data; and the computer may inaccurately display the data or display it in a biased manner. Because of the complexities of examining the creation of computer-generated evidence and the deceptively neat package in which the computer can display its work product, courts and practitioners must exercise more care with computer-generated evidence than with evidence generated by more traditional means.
Roberts, A Practitioner’s Primer on Computer-Generated Evidence, 41 U.Chi.L.Rev. 254, 255-56 (1974).
There are those knowledgeable in the field of computerization who believe that new evidentiary rules will be required to channel and control the use of this new medium. Freed, Computer Records and the Law — Retrospect and Prospect, 15 Jurimetrics J. 207, 208 (1975). Indeed, much nonjudicial effort has already been expended in this field, and specific recommendations have been made. See e. g., Federal Judicial Center, Manual for Complex Litigation § 2.717 (1973); ABA Sub-Committee on Data Processing, Principles of Introduction of Machine Prepared Studies (1964). See also Bibliography of Representative Articles and Comments in Freed, supra, 15 Jurimetrics J. at 218.
Judicial decisions to date have largely skirted the edge of the problem because they have been concerned mainly with computerized records made in the regular course of business. See United States v. Russo, 480 F.2d 1228 (6th Cir. 1973), cert. denied, 414 U.S. 1157, 94 S.Ct. 915, 39 L.Ed.2d 109 (1974); United States v. DeGeorgia, 420 F.2d 889 (9th Cir. 1969); Olympic Ins. Co. v. H. D. Harrison, Inc., 418 F.2d 669 (5th Cir. 1969); D & H Auto Parts, Inc. v. Ford Marketing Corp., 57 F.R.D. 548 (E.D.N.Y.1973). Routinely prepared records, admitted pursuant to business records acts such as 28 U.S.C. § 1732(a) are well recognized exceptions to the hearsay rule, because their regular use in the business of the company insures a high degree of accuracy. Proof of day-to-day business reliance upon computerized records should therefore make less onerous the burden of laying a proper foundation for their admission. United States v. Russo, supra, 480 F.2d at 1239-40.
Where, however, a computer is programmed to produce information specifically for purposes of litigation, an entirely different picture is presented. Its product, which is hearsay and conclusory, is not admissible under 28 U.S.C. § 1732(a) or similar state statutes. United States v. DeGeorgia, supra, 420 F.2d at 895; Melinder v. United States, 281 F.Supp. 451 (W.D.Okl.1968). Under such circumstances, a court should not permit a witness to state the results of a computer’s operations without having the program available for the scrutiny of opposing counsel and his use on cross-examination. United States v. Dioguardi, 428 F.2d 1033 (2d Cir.), cert. denied, 400 U.S. 825, 91 S.Ct. 50, 27 L.Ed.2d 54 (1970). Moreover, such availability should be made known sufficiently in advance of trial so that the adverse party will have an opportunity to examine and test the inputs, program and outputs prior to trial. United States v. Russo, supra, 480 F.2d at 1241.
Long before the age of computers, the law was established that an expert witness might refer to records such as elaborate mathematical calculations, if, but only if, such records were made available for inspection by opposing counsel and thorough cross-examination thereon was permitted. Edwards v. Sears, Roebuck and Co., 512 F.2d 276, 294 (5th Cir. 1975); Di Filippo v. *126Gargiulo, supra, 278 App.Div. at 176, 104 N.Y.S.2d 149. Because of the computer’s “ability to package hearsay and erroneous or misleading data in an extremely persuasive format”, Roberts, supra, 41 U.Chi.L. Rev. at 279, this rule should be strictly adhered to whenever expert testimony is predicated upon specially prepared computerized calculations or simulations.
It is a mistake to liken the program of a computer to human calculation, because the program directs the performance of tasks that humans would not attempt, in a manner that they would not elect. Jacobs, Commission’s Report on Computer Programs, 49 J.Pat.Off.Soc’y 372, 374 (1964). An error in programming can be repeated time after time,16 and simulation with an incorrect program is “worse than worthless”. Favret, Introduction to Digital Computer Applications, supra, at 122. For this reason, programming requires great accuracy, more than that needed in other types of engineering. Ershov, Aesthetics and the Human Factor in Programming, 13 Jurimetrics J. 142 (1973).
Because the record in this case contains no information about the programming relied upon by plaintiff’s experts, neither we nor defense counsel have any way of knowing whether it was complete or accurate. It may be that the District Court has “that specialized acquaintance with the subject matter which enables it to dispense with evidence and draw upon undisclosed sources of information not available to us”. Sinko Tool & Mfg. Co. v. Automatic Devices Corp., supra, 157 F.2d at 978. However, if we are to perform our function as an appellate court, we need more to review than the District Judge’s modestly admitted acquaintance with “LaGrange equations”.
Although the admissibility of evidence is ordinarily for the trial court, where, as here, plaintiff’s entire case rests upon the accuracy of its computerized calculations, the refusal of the District Judge to subject such computations to the searching light of full adversary examination constitutes prejudicial error which mandates reversal.17
PROOF OF DAMAGES
There is a long line of New York cases holding it improper for a real estate expert to evaluate property by capitalizing “hypothetical income from hypothetical tenants who would be occupying a hypothetical building.” City of Binghamton v. Rosefsky, 29 App.Div.2d 820, 821, 287 N.Y.S.2d 249, 251 (3d Dep’t 1968); Wer Realty, Inc. v. State, 26 App.Div.2d 732, 271 N.Y.S.2d 714 (3d Dep’t 1966); Levitin v. State, 12 App. Div.2d 6, 207 N.Y.S.2d 798 (3d Dep’t 1960). In the instant case, the District Judge computed damages from the hypothetical income of a hypothetical anti-skid device to be sold at a hypothetical price in a hypothetical market. In my opinion, damages based on proof such as this are too uncertain to support a substantial award. Friedman v. Golden Arrow Films, Inc., 442 F.2d 1099 (2d Cir. 1971); James Wood General Trading Establishment v. Coe, 297 F.2d 651 (2d Cir. 1961); Freund v. Washington Square Press, Inc., 34 N.Y.2d 379, 357 N.Y. S.2d 857, 314 N.E.2d 419 (1974); Hewlett v. Caplin, 275 App.Div. 797, 88 N.Y.S.2d 428 (1st Dep’t 1949), aff’d mem., 301 N.Y. 591, 93 N.E.2d 492 (1950).
Moreover, even if there had been adequate proof, I do not believe that loss of anticipated profits should be the proper measure of damages in a case such as this. *127There are two separate and distinct classes of implied contracts: those implied in fact as a result of the acts of the parties, and those implied in law as the result of the application of equitable principles. See cases cited in note 1 supra. Contracts implied in fact are true contracts; those implied in law are quasi-contracts. Id. A contract should not be implied in fact against the declarations and intentions of the parties or where the facts are inconsistent with its existence. Miller v. Schloss, 218 N.Y. 400, 406, 113 N.E. 337 (1916); 3 Corbin on Contracts §§ 564, 569 (1960). While courts which declare the existence of quasi-contractual liability often do so on the basis of a somewhat fictional finding of intent, we have described this as a “good formula by which to square doctrine with result.” Parev Products Co. v. I. Rokeach & Sons, Inc., supra, 124 F.2d at 149.
There was no true or actual agreement to perfect in the instant case. In the face of plaintiff’s adamant and continued claim that its anti-skid device was at all times marketable and ready for mass production and the specific language of the written contract that there were no undertakings other than as expressly set forth therein, there could have been no actual mutual intent that defendant was obligated to perfect plaintiff’s device and make it marketable. If any such obligation on defendant’s part is to be implied, it must be an obligation implied in law, or a quasi-contractual obligation. If this is so, the quasi-contractual measure of damages should be applied for its breach.
In the early days of the law, when the cause of action for simple and implied contract was in assumpsit, damages were measured by the consideration given rather than the value of defendant’s performance. 11 Williston, Contracts § 1338, at 201 (3d Ed. 1968). Williston says that:
[s]uch a rule was natural when assumpsit was regarded as in the nature of a tort for deceit, and when, therefore, it might well seem that the law should put the plaintiff in as good a position as he was in before the contract was entered into rather than in as good a position as he would have been in had the performance of the contract been fully carried out. Id.
We have long since passed beyond the limitations of this rule in actions for breach of contract. And yet, there are those who have questioned the logic of the change. Fuller and Perdue, Jr., in The Reliance Interest in Contract Damages, 46 Yale L.J. 52, 56 (1936), state that, “In passing from compensation for change of position to compensation for loss of expectancy, we pass . . from the realm of corrective justice to that of distributive justice. The law no longer seeks merely to heal a disturbed status quo, but to bring into being a new situation.” They conclude that it is no easy thing to explain why the normal rule of contract recovery should be that which measures damages by the value of the promised performance.
The theory of the remedy in quasi-contract continues to be corrective justice, i. e., that the parties should be placed substantially in statu quo. Woodward, The Law of Quasi-Contracts § 268. For contracts implied in law, or quasi-contracts, the measure of recovery is generally the amount of the unjust enrichment of the defendant. 1 Corbin on Contracts § 19, at 50 (1963).
Courts which parrot the word “intent” in the process of finding a quasi-contractual obligation have little difficulty in applying contractual rules of damage for its breach. However, if quasi-contracts are not true contracts but rather obligations implied in law to prevent injustice, the remedy for their breach should be no more than the restoration of the status quo, not an award based upon the expectancy of gain from the impliedly promised performance. Hewlett v. Caplin, supra.
It is obvious from the foregoing that the issues raised by appellant on appeal are more numerous and complex than would appear from the majority opinion. While I sympathize with my brothers’ desire to bring this protracted litigation to an expeditious end, I think we should lay it to rest only after the parties have had their case *128fairly heard and decided. For all of the foregoing reasons, I respectfully dissent.
. See, e. g., Bradkin v. Leverton, 26 N.Y.2d 192, 309 N.Y.S.2d 192, 257 N.E.2d 643 (1970); Adams & Co. Real Estate v. E. & B. Super Markets, Inc., 26 A.D.2d 365, 274 N.Y.S.2d 776 (1st Dep’t 1966); Beidlar & Bookmyer, Inc. v. Universal Ins. Co., 134 F.2d 828 (2d Cir. 1943); Kulukundis Shipping Co. v. Amtorg Trading Co., 126 F.2d 978, 990 (2d Cir. 1942); Parev Products Co. v. I. Rokeach & Sons, Inc., 124 F.2d 147, 149 (2d Cir. 1941).
. This Court sits on the 17th floor of the United States Courthouse in Foley Square, New York.
. “The better reasoned cases have recognized that a licensed product or process must be marketable or commercially exploitable for the licensor to hold the licensee to his contract.” Vukowich, Implied Warranties in Patent, Know-How and Technical Assistance Licensing Agreements. 50 J.Pat.Off.Soc’y 307, 312 (1968).
. “The ‘best efforts’ clause is obviously no more than an expression of the aim of the parties in entering the agreement.” Scales, Best Efforts and Implied Obligations in Patent Licenses, 41 Marq.L.Rev. 385 (1958).
. The District Court correctly found that plaintiff’s device was neither perfected nor fail-safe.
. “A patent license grant-back gives the licensor full access to all future improvements and know-how that the licensee develops as a result of being licensed under the original patent.” Jacobs, Patents for Softwear Inventions, The Supreme Court’s Decision, 13 Jurimetrics J. 132, 137 (1973).
. In a pretrial memorandum filed by plaintiff, it said:
“The technical service contract was entered into so that the technical expertise of Perma would be available to Singer for improving its manufacturing methods.”
. There is substantial authority in both the New York and federal courts that the opinion of an expert witness must rest upon facts, rather than upon the opinions, inferences or conclusions of others. See, e. g., Kreutzwald v. Walters, 242 App.Div. 479, 480, 275 N.Y.S. 878 (4th Dep’t 1934) (per curiam); Bergman v. Mergenthaler Linotype Co., 274 App.Div. 553, 85 N.Y.S.2d 124 (3d Dep’t 1948) (per curiam); Marx v. Ontario Beach Hotel and Amusement Co., 211 N.Y. 33, 38, 105 N.E. 97 (1914); Stancill v. McKenzie Tank Lines, Inc., 497 F.2d 529, 536 (5th Cir. 1974); 6816.5 Acres of Land v. United States, 411 F.2d 834, 840 (10th Cir. 1969); Taylor v. B. Heller and Co., 364 F.2d 608, 613 (6th Cir. 1966). Where, as in this case, the opinion of another which is relied upon by the expert is based upon hearsay information received from a third person one more step removed, the expert’s testimony should be looked at with a jaundiced eye even under the liberal Federal Rules of Evidence.
. Plaintiffs own expert conceded that the anti-skid unit “had to be designed so that it could function under the total range expected in the automotive environment where the unit is to be installed”.
. “Applied mathematicians . are always simulating something and know what they are talking about (but they may be far from certain whether what they are saying is true).” Gonzalez and McMillan, at 186.
. Among the things which were not simulated in the theoretical device of plaintiffs witness were engine stalls and vacuum loss, the transfer value, the plunger in the hydraulic cylinder, the gears, the freedom of motion in the adapter gear, the tortional elasticity of the flexible shaft, the performance of the drive gear, the backlash in the gears, the oscillation between a retaining pin and the cam gear and the friction in the governor. Indeed, plaintiffs expert ignored the measurements and figures in plaintiffs existing unit.
. In patent infringement cases, where priority of invention is in issue, proof of prior conception standing alone may not be sufficient. It is often required that the “prior” inventor establish that his invention has been tested to make sure that it will operate under service conditions. Payne v. Hurley, 71 F.2d 208 (Cust.Ct. & Pat. App. 1934).
. The witness also testified that after the five-month “failure analysis” program, an additional five months would have to be devoted to the making of pre-production samples which would be subjected to environmental and reliability testing. This would be followed by still another period of fleet build up and testing.
. The District Judge stated that the only mechanical device which could meet this test would be a wedge, “the simplest tool known to man”.
. Although it was not disclosed whether plaintiffs witness had protected his program under the copyright laws, such protection was available if needed. Gottschalk v. Benson, 409 U.S. 63, 72, 93 S.Ct. 253, 34 L.Ed.2d 273 (1972).
. A recent news report stated that federal computers automatically waste millions of dollars because of mistakes programmed into their systems. As an example, the Navy was said to have taken five years to correct a computer that was automatically initiating on its own at least three million dollars in unnecessary costs annually. National Review, June 25, 1976, at 8, col. 2.
. My brothers pay lip service to the rules above expressed but then completely emasculate them by holding that “Singer has not shown that it did not have an adequate basis on which to cross-examine plaintiffs experts”. Where all the information concerning the programming upon which plaintiffs experts relied is kept from the defendant, it is impossible for defendant to point to the inaccuracies and omissions that might have been uncovered by full disclosure and proper cross-examination.