delivered the opinion of the court:
Dynalectron Corporation (Pacific Division) seeks review of the decision of the Armed Services Board of Contract Appeals (ASBCA), -69 — 1BCA ¶ 7595, affirming a default termination. This court earlier by order, 199 Ct. Cl. 996 (1972), refused to review the Board denial of excess reprocurement costs to the Government. We cited S & E Contractors, Inc. v. United States, 406 U.S. 1 (1972). As frequently happens in this type of case, a recommended decision of Trial Judge Harkins has become the focal point of the parties’ adverse attention. However, under the law it is the Board decision we are reviewing. Trial Judge Harkins has aided us greatly by his able analysis but we reject 'his affirmance of the Board, *353which we find incompatible with his criticisms of the Board’s conclusions and a misconstruction of the contract.
The subject matter of this case is a fixed price, supply contract for 258 Very High Frequency (YHF) jamming antennas for use on B-52 Air Force bombers as part of the AN/ALA-15 Electronic Countermeasures (ECM) system. Plaintiff’s contract was awarded December 31, 1964, pursuant to a bid submitted on December 1, 1964, in response to an Invitation for Bids (IFB) issued on November 13, 1964. Plaintiff was the sole bidder from the 58 firms solicited. The contract called for “First Article” tests and test reports as well as an initial increment of 258 antennas (with options for an additional 449), each approximately $368.80 (minor variations in this price were negotiated during the life of the contract). The original contract involved $107,284.63, and each of the two supplemental orders added another $99,107.40, to the overall value of the procurement. The contract was terminated for default on June 16, 1966. Plaintiff had delivered only the 25 antennas whose delivery was allowed prior to completion of the First Article test program.
The original contract was twice formally amended: once on February 11, 1965, and a second time on December 20, 1965. It was the failure to have successful test data completed and available for Government review on January 3, 1966, as required by the December 20 agreement, that was the assigned cause for the default termination. The January 3 test data showed that plaintiff had not solved several critical engineering design problems, whose impossibility forms the basis for this appeal.
The standard for review is that of the Wunderlich Act, 68 Stat. 81,41 U.S.C. §§ 321-22 (1970). Upon careful review of the Board’s and trial judge’s decisions and of the entire record, we agree with the trial judge that the Board’s key factual finding of a waiver by contractor of all claims of contract impossibility is not supported by substantial evidence. Waiver requires knowledge, and the record is unconflicting that plaintiff (and defendant for that matter) did not realize the full impossibility of the defective specifications until May 1966, or about 18 months after the Board held plaintiff had full knowledge. We further observe the various defects in *354the specifications to be independent causes of impossibility with, tiie result that knowledge of some defects is not knowledge of all.
Therefore since we agree with the trial judge that at least one cause of impossibility was not known or realized until May 1966, and that it was then promptly communicated to the Air Force, said impossibility was not waived and serves as an absolute bar to a def ault termination. On the other hand, plaintiff did waive, by not communicating, some of the other defects in the specifications, which defendant might, if notified, have corrected to its own as well as plaintiff’s advantage.
By waiver in this case we mean that plaintiff, discovering some defects in the specifications, instead of refusing further effort at performance, as was its right (as even the Board admits), elected to treat what had been intended as a fixed-price production contract as if it were a ^Research and Development (It & D) project, experimenting with new techniques at its own sole expense, unless it were to succeed. If this had occurred as to all the defective specifications, plaintiff could not recover. But it did not.
I
A chronology of this case follows:
1960-1964: Tamar Electronics, Inc., developed the original “Tamar Antenna” to meet the pure performance specifications of Exhibit WCLG-3B (an Air Force document prepared by a Wright-Patterson AFB engineer to govern all AN/ALA-15 system component procurement contracts, including the two types of antennas required on the ALA-15 system.) Tamar Electronics subsequently turned over its “Tamar Antenna” to its wholly owned subsidiary, Stoddart Aircraft Radio Company, Inc. And in 1964 Stoddart sold its rights in the “Tamar Antenna” to Dynalectron. During the lifetime of the three procurements from Tamar/Stoddart, the Air Force attached the Tamar/Stoddart antenna drawings to Exhibit WCLG-3B as design specifications (whereas previously WCLG-3B had only limited antenna size, weight, and shape to “good engineering practices.”) The Air Force failed to keep the drawings attached to the exhibit current with the latest production model of the antenna so that by 1964 the drawings, e.g., showed a “no-eyelet” anten*355na while an “eyelet” antenna was being produced and delivered.
Summer 1964: Stoddart Aircraft Eadio Co. and the Air Force became embroiled in a default termination dispute because the Tamar antenna would not meet all of the WCLG-3B specifications; the Air Force backed down on the dispute and production of the Tamar antenna with eyelets continued.
August 1964: Dynalectron and Stoddart completed negotiations for sale of Tamar antenna line to Dynalectron.
November 1964: After the Air Force-Stoddart-Dyna-lectron negotiations to license further Tamar antenna procurement from Stoddart fell through, the Air Force issued its IFB. The IFB contained the defective specifications.
December 1, 1964: Dynalectron submitted only bid received by the Air Force. Bid letter noted intention to supply Tamar antenna with eyelets and to submit copies of prior test procedures for use with new contract. Bid letter also questioned what Air Force meant by “qualification tests” in IFB.
December 10, 1964: Dynalectron TWX accepted Air Force definition of “qualification tests” as synonymous with “engineering acceptance tests” and “pre-production teste.”
December 21, 1964: Dynalectron received classified test data on Tamar antennas from Stoddart Aircraft Eadio Co.
December 31, 1964: Dynalectron contract with Air Force was legally effective on this date.
Mid-January, 1965: Dynalectron completed review of classified test data received in December and became aware of some defects in the specifications.
February 11, 1965: First formal amendment to contract signed. Dynalectron agreed to non-eyelet antenna (new production process and design) and to new test procedures.
February 26,1965: Other Air Force procurement officials issued unilateral incremental order for additional Dynalectron-Tamar antennas (delivery to be concurrent with delivery under original order).
March 2, 1965: Dynalectron protested unilateral incremental order as unreasonable in light of unexpected lengthy delay in Air Force approval of test procedures. Dynalectron was allowed and required to deliver only 25 antennas prior to completion of entire test procedure and testing program.
*356April 21,1965: Dynalectron. sent Air Force first of several notices about defective Government-furnished equipment (GFE), signal generator. Generator was a classified piece of test equipment available only from the Air Force.
April 29, 1965: Dynalectron sent Air Force first of several requests for Air Force to obtain extension of FCC frequency-authorization permit needed by Dynalectron to legally test Tamar antenna.
May 12, 1965: Dynalectron gave Air Force notice that introduction of crash safety test requirements into WCLG-3B test requirements was a change. In regard to the same test requirements, Dynalectron ashed the Air Force to clarify which of two “last” tests was to be last — since both tests tended to destroy the test antenna and preclude further testing.
June 18, 1965: Dynalectron notified the Air Force that Air Force changes in test procedures would destroy Air Force signal generator — and that the risk of destruction was the Air Force’s responsibility.
July 13,1965: Even though the Air Force had not completed its review and approval of submitted test procedures, the Air Force threatened plaintiff with default termination for delays.
July 20, 1965: Dynalectron’s response to Air Force threats of default termination blamed the delays on Air Force slowness in reviewing and approving test procedures and to substantial changes that the Air Force made in test procedure requirements. This default termination dispute was dropped by the Air Force after Dynalectron underwent a change in management.
August 18, 1965: Final approval of all test procedures was given to Dynalectron.
September 1, 1965: Dynalectron frequency-authorization permit expired without being renewed as requested by Dynalectron. Certain required electrical tests were illegal until a new permit issued.
September 8, 1965: Air Force was given full technical details of the initial-test failures of the Dynalec-tron-Tamar Antenna (which primarily was a failure to successfully pass the vibration test requirements). Dynalectron sought waivers of certain vibration test requirements.
September 1965: Air Force commenced procurement of twice-as-heavy and sturdier Adams-Russell an*357tenna. Adams-Russell antenna for acceptability required waivers of all the design specifications in WCLG-3B (weight, shape, size).
October 28,1965: Dynalectron sought change in test procedures to eliminate requirement that thermocouples be mounted on antenna during high-power radiation tests. The thermocouples burned up during the power tests, destroying the surrounding surface of the antenna (burning metal on a fiberglass-and-plastic surface is a bad combination).
November 24, 1965: The Air Force issued more default termination threats.
December 20, 1965: Air Force entered final signature on second and final formal amendment to contract. Contractor agreed to have test data available for Air Force review on January 3, 1966 (which was done). Contractor also agreed to complete testing by December 18, but this did not occur because of severe weather delays in testing during December.
January 3, 1966: Test data showed substantial failure of Volt-standing-wave-ratio (VSWR) requirements, vibration requirements and radiation pattern requirements. (Other minor discrepancies were also present.)
January 11, 1966: Air Force sent Dynalectron notice of a default-termination investigation.
March 14,1966: Joint meeting between parties occurred. Record indicates that parties came away from meeting with different views of what had been accomplished. Air Force officials testified that plaintiff had accepted a 90-day time limit to complete its antenna or be defaulted. Plaintiff disputes this and says the Air Force authorized plaintiff to procure or make their own version of Adams-Russell single-blade antenna (with implicit waiver of WCLG-3B design specifications — just as with Adams-Russell). Dynalectron did in fact commerce its effort to build its own version of Adams-Russell antenna at this time, while continuing to try to solve design problems with Dynalectron-Tamar antenna. Air Force was timely notified of this new effort to comply with Air Force demands for antennas.
May 10-11, 1966: Another joint meeting occurred.
May 11, 1966: Air Force notified plaintiff in writing that all antenna work (on single-blade Adams-Russell type antenna) and all use of unapproved test procedures was at plaintiff’s “own risk.”
*358May 24, 1966: The Air Force trip report summarized situation:
(a) plaintiff bad just failed power-bandling test with latest Tamar-design antenna and was giving up Dynalectron-Tamar antenna effort as hopeless and impossible;
(b) plaintiff was proceeding on its version of a single-blade antenna, and asking an extension of time to July 25;
(c) the Air Force did not know whether this new effort would succeed; and
(d) the contracting officer recommended a default termination to his superiors in light of these facts.
June 16, 1966: Plaintiff was formally terminated for default.
July 1966: Plaintiff completed work successfully on a prototype single-blade antenna (similar to Adams-Bussell).
II
The following major problems afflicted this contract:
1. Bad specifications:
(a) The IFB in November 1964 included the entire WCLG-3B exhibit with out-of-date drawings attached— the drawings should have been those of the Tamar Antenna with eyelets then in production. Instead, the drawings were of earlier Tamar Antennas no longer in production or acceptable for delivery to the Air Force.
(b) The IFB did not show any of the many waivers and deviations to WCLG-3B granted to current and prior Tamar Antennas in order to be acceptable — no Tamar Antenna ever met all WCLG-3B specifications. Yet plaintiff initially assumed it would supply the Tamar Antenna which the Air Force was then accepting, unchanged.
(c) The IFB did not contain any actual test procedures, which (in addition to formal waivers) was another means the Air Force used to grant de facto waivers of WCLG-3B specifications.
(d) The contract was a collection of documents not readily found together (since several classified documents were included by reference). As a result it is doubtful that the entire collection ever received the type of engineering and legal review usually accorded Government contracts by the Air Force.
*359(e) Technical deficiencies occur in the specifications: WCLG-3B refers to “unit control drawings” to determine applicable tests for components obtained separately from the entire system. The unit control drawings for this particular antenna specify the (a) temperature-altitude, (b) vibration, (c) shock, (d) humidity, (e) immersion, (f) salt spray, and (g) sand and dust tests, as well as any other tests necessary to insure operation of the entire system. These test requirements of WCLG-3B were inconsistent with the production model of the Stoddart-Tamar Antenna with eyelets.
(i) The eyelets were incompatible with the full immersion test which required an electrical performance test immediately after removing the immersed antenna from water. With a wet interior the antenna did not work properly — and the eyelets insured a wet interior.
(ii) The specified physical length of the antenna was too short by about half an inch to comply with electrical (YSWR) frequency and power requirements of WCLG-3B.
(iii) The shape and size of the antenna were incompatible with the classified radiation pattern specifications.
(iv) The stress characteristics of the light-weight materials used did not meet the temperature, altitude (pressure), vibration, and power-handling specifications of WCLG-3B without waivers.
None of these problems were new to the Air Force, and the stringent test specifications had been added to WCLG-3B as an Air Force reaction to earlier antenna failures. Yet, even with prior notice from waivers asked for and granted and failed antennas, the Air Force never corrected or otherwise resolved these technical deficiencies in the specifications.
2. Poor contract management:
Both the Board and the trial judge noted poor contract management as a significant factor in this case.
'On Dynalectron’s side, several of the defects were noted in January 1965, but were not reduced to formal written notices to the Air Force until months later. The February 1965, amendment to the contract gives little hint of the technical problems discussed concurrently with extending the delivery schedule. This overly-cooperative attitude of plaintiff’s working-level personnel with Air Force personnel de*360layed substantially formal contract action on the key technical problems in this case.
Similarly, on the Air Force side, 1.) geographical separation (Los Angeles officials were day-to-day administration delegates; Bobins AFB, Georgia, officials were official procurement officers; and Wright-Patterson AFB, Ohio, officials made all the technical decisions including test procedures decisions and specification waivers), 2.) other work, and '3.) lack of technical expertise in certain key technical areas of the procurement (antenna production problems as opposed to system requirements and testing) precluded proper contract coordination and management. Eveiybody else deferred to Wright-Patterson AFB officials, and the Wright-Patterson AFB officials apparently never realized all the technical impossibilities built into their defective specifications. As a result, the specifications were never corrected.
3. Other problems:
The combination of bad specifications and poor contract administration aggravated other problems. The 10-day test-procedure copying-exercise expanded into an eight-month project as the Air Force officials dictated many substantial changes in prior Stoddart-Tamar test procedures. The problem of the frequency-authorization permit (to broadcast test frequencies to test the electrical jamming capabilities of the antenna) went unresolved for months once the FCC and FAA discovered that the Tamar Antenna jammed commercial air-traffic-control frequencies in the Los Angeles area very effectively. The GFE signal generator broke down so many times and functioned so poorly, when operating, that an uncalculated amount of time was lost, yet the Air Force never provided a replacement. And the months of delay postponed a significant amount of testing into December 1965, when unusual snows followed by a sudden thaw accompanied with torrential rains caused some of the worst flooding ever in the West coast area. Plaintiff’s test range sustained substantial damage and further delays to the program resulted.
Ill
In regard to the claims of impossibility we come to the same conclusion that Trial Judge Harkins did — plaintiff waived the immersion problem and the antenna length prob*361lem, as impossibilities excusing performance deficiencies, but that plaintiff did not waive either the radiation patterns problem or the vibration test/weight-strength problem. (The vibration and power-handling test failures resulted from the inability of various combinations of light-weight materials to withstand the strength requirements of the vibration, temperature, pressure and power-handling test requirements.) This, as noted by both the Board and the trial judge, results in a situation of mutual fault.
But before discussing the legal effects of the parties’ actions, it is necessary to review the facts (since the Board— despite its own contrary conclusions — concluded that plaintiff waived all of its impossibility claims in December 1964).
The waiver of the immersion problem was amply proved by Guy Nicholson’s own testimony and the Dynalectron letter of July 20, 1965, in the Buie 4 File of the Board Becord:
In an effort to cooperate to the fullest extent possible, and after considerable negotiation and discussion, the accelerated schedule as set forth in Modification No. 1 [February 11, 1965, agreement] to the contract was concluded at no increase in cost to the Government. This despite our admonitions the antenna to be furnished, although resembling in outward appearance its predecessor, would have to be considerably redesigned [without eyelets] to meet the ungraded requirements of the Air Force. (Emphasis supplied.)
Similarly we find a waiver by failure to give notice or request a change order in regard to the VSWB antenna-length electrical frequency and power problem. At page 138 of the Board transcript, plaintiff’s engineer Guy Nicholson testified that the classified test data (delivered December 21, 1964, and analyzed by mid-January 1965) clearly showed that:
* * * The low power VSWB measured on the system did not meet the specification maximum of two frequencies, at two frequencies.
The Board (at 4-5 of its opinion) noted that:
Dynalectron did not notify the Government concerning its discovery contemporaneously. * * *
Trial Judge Harkins (at 16 of his opinion) was even more pointed:
*362* * * Plaintiff, moreover, did not ask for any deviations from the specifications, either those that applied to the -14 A-1 or others, until May 12, 1966. [This statement overlooked September 1965, waiver requests.]
The record is also clear that Stoddart had had the same problem and sought a waiver — asking for longer length as in the original Boeing antenna and instead receiving a waiver of the electrical requirements and being told to keep the shorter length (the shorter length resulted in a smaller total exposed surface area of the antenna — a key factor in the aerodynamic stresses imposed on antenna and aircraft, and a factor which the Air Force determined outweighed the slight loss in electrical performance).
No adequate explanation is shown in the record for this failure to give the Air Force prompt notice of the defective specification as to antenna length incompatible with VSWR specifications.
However, as regards other technical deficiencies, the trial judge differed (at 23-24) with the Board:
* * * Plaintiff, however, did not know that the “totality” of the specification was deficient until after failures in the vibration tests in September 1965, and in the power-handling test on May 16, 1966. Plaintiff acquired actual knowledge that the specification may have been deficient in its “totality” only after it had testing-experience under conditions in which deviations were not permitted.
After carefully reviewing all the evidence, we agree with the trial judge that plaintiff did not recognize the full impossibility of the specifications until it had completed testing every alternative method of performance in May 1966, at which point it so informed the Air Force. We can cite no stronger evidence of this state of knowledge of plaintiff than plaintiff’s own actions which fully accord with the trial judge’s conclusion. Only in May 1966, after exhausting all alternatives to solve vibration/altitude-temperature/power-handling specifications, did plaintiff stop and tell the Air Force its specifications were defective.
We do not choose to discuss at length parties’ semantic arguments over “deficient”, “impossible” and “impractical”. The real test, as we noted in Foster Wheeler Corp. v. United States, 206 Ct. Cl. 533, 545, 513 F. 2d 588, 594 (1975), is *363whether the specifications require: “performance beyond the state of the art * * The Board used concerning them the terms “deficient” and “impractical” to indicate that they thought performance was beyond the state of the art. There is no documentary evidence or expert testimony to indicate that the antenna-length (VSWB) and vibration/altitude-temperature/power-handling test specifications were within the then-defined state of the art. All of the evidence points to the single conclusion that neither plaintiff nor anyone else could perform the Air Force specifications without waivers).
IV
The question that must now be asked is what legal conclusions are to be drawn from the foregoing facts.
The trial judge correctly distinguished certain cases where “impossibility” was held the contractor’s risk: Austin Co. v. United States, 161 Ct. Cl. 76, 314 F. 2d 618, cert. denied, 375 U.S. 830 (1963) ; Bethlehem Corp. v. United States, 199 Ct. Cl. 247, 254, 462 F. 2d 1400, 1404 (1972); United States v. Wegematic Corp., 360 F. 2d 674 (2d Cir. 1966); and J. A. Maurer, Inc. v. United States, 202 Ct. Cl. 813, 485 F. 2d 588 (1973). Austin is distinguished since the contractor did not furnish the specification as was done in Austin. Bethlehem is not applicable because the Government did not rely here on any superior expertise claimed by the contractor. Nor did contractor propose a revolutionary advance in the state of the art, as in Wegematie. Nor was there a contractor’s mistake in technical judgment associated with inferior expertise of the Government as in Maurer.
The trial judge also correctly found that plaintiff had not waived all the deficiencies in the specifications, though it had waived some. In short, the trial judge found fault on both sides and yet upheld the default termination. The trial judge declined to find the fault entirely beyond the control of the contractor.
We have a somewhat novel problem. In our precedent cases deficient specifications have been entirely the risk of one party or the other. Here we have multiple and independent deficiencies, so that even if one were properly cured, the others would still bar the door to success.
*364This leads to the first question: was plaintiff properly-defaulted?
Y
A default termination is the Government’s way of telling its contractor that he has breached the contract. In a default termination, the Government assumes the risk of having the default termination converted to a convenience determination if the Government is wrong and the contract so provides, as here.
The second item to note is that it takes only one valid defense to void a default termination. If the Government is not sure whether plaintiff has defaulted, it can seek further assurances as the Government did in this case several times. But when the Government actually issues the termination notice, it runs the risk of being wrong.
Impossibility is a classic defense to breach of contract.
By issuing defective specifications, some not waived, the Government has put this case somewhat in the category of our recent Foster Wheeler Corp., supra, decision and related cases. There the court, 206 at 557, 513 F. 2d at 601, noted certain features common to such cases:
1. a fixed-price, supply contract,
2. detailed performance specifications drawn up by the Government,
3. a requirement that the contractor submit a Technical Proposal,
4. meetings of the parties when the contractor’s difficulties became 'apparent, and
5. an eventual relaxation of specifications by the Government so that the contractor would be able to complete the contract.
Beviewing this list of factors, numbers 1 and 4 apply and already have been amply described. Number 2 'also applies. Although the specifications were mixed performance and design specifications, the design specifications (which by drawing fixed weight, shape, and dimensions of the antenna) here left the contractor no elbow room to avoid the impossibility of the specifications. Had the Air Force given plaintiff the waivers it sought, to avoid the design specifications, this would be a different case. But by insisting on its design specifications, despite notice of their deficiencies, the Air Force accepted the risk that its design specifications were non-*365performable. Number 8 is not applicable here. Any prejudice to the Air Force from the lack of a technical proposal is eliminated by the Air Force’s prior experience with Tamar Antennas which provided the Air Force ample notice of the inadequacies of its specifications — inadequacies that for some reason were not corrected in a timely fashion by the Air Force.
Number 5 of the list points up the unique item about this case. Normally the Government gives in when the error of its specifications is amply demonstrated. Here the Government did not give in. Instead the Government waived any further right to performance by terminating plaintiff for default, running the risk (as in all default terminations) that the termination was not proper.
It seems apparent in light of what has been said that on the termination date plaintiff was entitled either to a convenience termination or to new specifications with the impossibilities eliminated and an extended completion date provided.
Once the Government terminated plaintiff, it found no other manufacturer who would or could perform its impossible specifications upon reprocurement. Instead, the offending design specifications had to be waived and the Adams-Russell antenna procured. Not only was the Adams-Russell antenna twice as heavy (negating the weight saving that had prompted adoption of the Tamar Antenna in the first place) but it also cost significantly more (the Air Force excess — reprocurement-cost notice claimed it cost $410,994, to obtain from Adams-Russell the 749 'antennas Dynalectron had agreed to supply for $288,470.44 (including tests and test data)). Since the reprocurement contract was not signed until January 1967, it is also questionable if the Air Force saved any time. (Plaintiff’s single-blade antenna may have been available sooner — and cheaper.)
VI
However, because of plaintiff’s partial waiver and its deficient contract administration, we do not think plaintiff is entitled to shift all of its losses to the Government. In view of the share plaintiff’s fault played in bringing the situation *366about, and its repeated failure to point out the Government’s errors, which, if done, might have enabled the Government to minimize its own losses as well as plaintiff’s, we do not tbinlr it would be equitable to shift all the costs to the defendant, as in the normal convenience termination.
As the trial judge noted, characterizing the termination is merely the first step in ascertaining liability for damages.
Each tribunal agreed that the specifications were deficient. However, the trial judge disagreed with the Board on the waiver issue, as do we. The trial judge and we conclude that plaintiff had at least two separate and independent impossibility claims — both blocking the road to success. Some were waived. The others not.
The waiver doctrine, as noted in Foster Wheeler, supra, 206 Ct. Cl. at 558, 513 F. 2d at 602, (citing Ling-Temco-Vought, Inc. v. United States, 201 Ct. Cl. 135, 475 F. 2d 630 (1973)), rests on the premise that delay in asserting rights prejudices the Government by running up recoverable costs. Here it is practically impossible to determine the prejudice to the Government. In the past, when notified of the antenna-length specification problem, the Air Force simply granted a waiver to the WCLG-3B specification. Here, even when plaintiff gave the Government notice of the problem on January 3, 1966 (and repeated it again in later summaries of the January 1966 test data), no significant response was forthcoming from the Government over the problem.
We note that the Board and the trial judge thought they were splitting the costs by denying defendant any reprocurement costs. However, the Board also denied defendant repro-curement costs on the obviously valid ground that there was no reprocurement — in effect letting defendant walk away from a mess largely of its own creation with no liability. If there was no reprocurement the Board decision does not divide the costs.
The trial judge says plaintiff must be held to the “literal terms” of the provisions of the contract it executed, referring in this coimection to Art. 11(e) by which a default termination is converted into a convenience termination only if “the failure to perform the contract is due to causes beyond the control and without the fault or negligence of the contrac*367tor * * Since the contractor was not without fault or negligence, it cannot have a settlement by the convenience termination formula. So much is clear. But in holding that by Art. 11(e) the contractor wholly forfeits an otherwise valid claim, the trial judge goes beyond the literal language and beyond any implication that could reasonably be imputed to its authors.
In the absence of such a clause a wrongful termination for default is a breach. Dale Construction Co. v. United States, 168 Ct. Cl. 692, 721 (1964). As there held, some fault on the contractor’s side does not bar him from breach recovery, with appropriate cost sharing where fault is mutual (at 706). The Government here could have limited its prior breach liability by a proper convenience termination. Nolan Bros., Inc. v. United States, 186 Ct. Cl. 602, 405 F. 2d 1250 (1969). But it would have had to reimburse the costs of the contractor by the convenience termination formula. The clause surely could not have been intended to enable the Government to avoid all liability for antecedent breaches by trumping up a wrongful default termination, yet that is the power the trial judge gives it.
The contract we think does not provide a measure of damages and it has to be provided by common law or equitable principles. When, as here, the defaulted contractor is not without fault or negligence, but the default was wrongful because of the Government’s previous fault or negligence, Art. 11(e) may impose a ceiling in that the damages perhaps may not exceed those that would be computed by the termination formula.
If Art. 11(e) does not reduce the damages awardable to zero, we must go elsewhere to learn how to measure them. We find instruction in National Presto Industries, Inc. v. United States, 167 Ct. Cl. 749, 338 F. 2d 99 (1964), cert. denied, 380 U.S. 962 (1965). There in a mutual fault situation we effected a reformation of the contract on the theory of a mutual mistake, sharing the losses equally. Since, unlike here, we there rejected the claim for breach liability and had therefore to find a remedy under the contract, this case would seem a fortiori eligible for a similar innovative approach. The equitable powers of this court in actions at law *368are not limited to instances of mutual mistake. They extend also to unilateral mistakes when they are knowingly and unfairly exploited by the other side. Chernick v. United States, 178 Ct. Cl. 498, 372 F. 2d 492 (1967). The plaintiff here thought it was contracting to deliver the previously accepted Tamar/Stoddart design of antenna. By the time it discovered its error, it was in a situation of performing an K&D contract under ostensible fixed-price production contract terms. This situation defendant exploited to put plaintiff more and more in the hole, and finally, by a wrongful default termination, defendant sought to throw the entire costs of the numerous errors of both sides on the plaintiff. In National Presto, in effect we converted the experiments there performed into a joint venture at shared cost. Here, since the convenience termination formula is ruled out, except as a ceiling, we think the formula of shared costs as in a joint venture is likewise applicable. The Government benefited from plaintiff’s flounderings at least to this extent: it tried out a number of false leads and learned what would not work at plaintiff’s cost.
The decision we arrive at is far more equitable than the Board’s. The Government retains its liability for issuing defective specifications. Foster Wheeler, supra; Hol-Gar Mfg. Corp. v. United States, 175 Ct. Cl. 518, 360 F. 2d 634 (1966 ); Natus Corp. v. United States, 178 Ct. Cl. 1, 371 F. 2d 450 (1967); Tecon Corp. v. United States, 188 Ct. Cl. 436, 411 F. 2d 1271 (1969); Tombigbee Constructors v. United States, 190 Ct. Cl. 615, 420 F. 2d 1037 (1970). But the Government also carries only a share of the damage assessment because of plaintiff’s contributory errors. The result in this case rewards no one for careless contracting, whereas the mere denial of reprocurement costs exonerates the Government of all fault — -in a situation where Government errors were the greater cause of the failure of performance (even though aggravated severely by plantiff’s failure to give timely notice of the early-detected defects). Consequently we have differed with the Board and the trial judge and applied the precedents of this court to work “jury verdict” equity among the parties and encourage correct contract procedures on both sides.
*369Accordingly the recommended decision of the trial judge is rejected and the parties are to share equally the allowable and reasonable costs normally recoverable by plaintiff in a convenience termination. Since this is a breach case, the cause is remanded to the trial division for quantum proceedings under Rule 131(c).' Both motions for summary judgment are allowed in part and denied in part, pursuant to this opinion, and judgment is entered for plaintiff to the extent called for in this opinion.