(dissenting). Had the defendant accountants at Hawkins, Ash, Baptie & Company (HABCO) been lawyers, they would have been disbarred and prosecuted. Stealing from clients is outrageous behavior, and deserves to be substantially penalized. That is what the jury decided when it awarded $1,750,000 in punitive damages to Management Computer Services, Inc. (MCS). But a majority of this court has reduced MCS's compensatory damages from $2,585,750 to $65,000 and has awarded punitive damages of only $650,000 because that figure is exactly ten times the amount of the reduced compensatory award.
The majority fails to understand MCS's theory of recovery and it also adopts a new rule for determining whether a contract is indefinite which is wholly at variance with past precedent. I cannot accept the *625majority's analyses or conclusions, and, therefore, I dissent.
BREACH OF CONTRACT
The basic problem with the majority's analysis of the contract between HABCO and MCS is its conclusion that because nothing in the wording of the contract expressly required HABCO to purchase more than one computer from MCS, HABCO did not breach the contract by buying additional computers from other vendors. But MCS did not and does not contend that the words of the contract required HABCO to purchase computers from MCS. Robert A. Sierp, president of MCS, testified: "[HABCO wasn't] required to buy computers from MCS." But that is not the end of the analysis because the contract between HABCO and MCS dramatically limited the use that HABCO could make of the software MCS sold to HABCO. Only if HABCO purchased additional computers from MCS could HABCO use MCS's software on those computers.
The contract reads:
This software is proprietary to MCS and shall be furnished upon completion under license to HABCO for use on the HABCO computer installation described in this Agreement. MCS shall provide the non-applications software described above at no extra charge for each additional computer system purchased by HABCO through MCS.
Sierp testified:
[T]he objective was if [HABCO was] going to use that software, that contract software, that [it] would buy a computer from us. If [it] wanted to go off and do some tax reporting, [it] could buy any *626computer [it] want[ed] and we weren't — we would not have been involved in that transaction.
Thus, it is apparent from the parties' contract and from Sierp's testimony that although the contract between MCS and HABCO did not require HABCO to buy MCS computers, if HABCO bought computers from another supplier, it could not use MCS software on those computers. What the majority fails to recognize is that without software, the computers bought from another vendor would be useless to develop turnkey computer systems for public housing authorities. The software furnished by MCS was highly specialized software developed at great cost and it was very valuable to a company in the business of licensing computer systems to public housing authorities.
The jury was asked whether HABCO breached its contract with MCS in three respects. It answered "yes" to all three special verdict questions. The majority dislikes the questions asked, and proposes an alternative. But that is not the test appellate courts use when faced with an argument that a jury question was misleading. In Topp v. Continental Ins. Co., 83 Wis. 2d 780, 785, 266 N.W.2d 397, 401 (1978), the court said:
This court has frequently stated that the form of the special verdict rests in the discretion of the trial court, and the court's chosen form will not be rejected unless the inquiry, taken with the applicable instruction, does not fairly present the material issues of fact to the jury for determination.
Instead of using this deferential review of the verdict used by the trial court, the majority, without considering the court's instructions, reviews the verdict de novo, and decides that it would have used alternative language. I cannot join in this sub silentio overruling of *627Topp. It will only cause confusion for future cases where the proper standard of review for special verdict questions is at issue.
HABCO, MCS, the trial court and the jury all understood MCS's theory of its case. I recognize that the jury questions could have been better worded, but when the jury was first asked whether HABCO breached the parties' contract by failing to purchase computers from MCS, the jury answered "yes" knowing that the contract was not worded: "HABCO agrees to purchase all computers from MCS." The jury responded affirmatively because it knew that the only practical way HABCO could have legally used MCS software for public housing authorities was to buy MCS computers. The jury also knew that HABCO avoided the increased cost of purchasing MCS computers by purchasing computers from another vendor and using MCS software on them in breach of the parties' contract. Because purchasing computers from MCS was the only practical way HABCO could avoid breaching the contract, question one of the verdict settled the real issue over which the parties contended. In any event, I cannot conclude that asking question one was an erroneous exercise of discretion, nor that the question failed to fairly present what everyone knew were the material issues of fact in the case.
The same is true of the second breach question. The jury was asked whether HABCO breached the parties' contract by failing to pay twenty-five percent of the program value to MCS for the use of the contract software. The majority's conclusion, that because HABCO did not purchase additional computer systems from MCS, HABCO did not breach the contract, is an overly simplistic answer and does not address MCS's theory of the case. The breach is HABCO's use of MCS *628software on non-MCS computers, not HABCO's failure to buy MCS computers. Had HABCO done what it agreed to do, it would have had to purchase MCS computers. It then would have used jointly owned software on those computers and paid twenty-five percent of the program value to MCS. MCS's damages for HABCO's breach of its agreement not to use MCS software on non-MCS computers equalled the twenty-five percent it would have received had HABCO not breached the contract. I conclude that the verdict form for the second question was not an erroneous exercise of discretion and that it fairly presented the material issues of fact in the case.
The final breach question is the only one the majority directly addresses, but it does so in a way which ignores our standard of review of a jury verdict. The jury was asked whether HABCO breached the parties' contract by failing to compensate MCS for the use of the proprietary software. Had HABCO not breached the contract by using MCS software on non-MCS computers, HABCO would have been required by the contract to pay twenty-five percent of the program value because it would have then used the software purchased through MCS. That was exactly what the jury was asked. Though the jury verdict might have been better drafted, or drafted in a manner which the majority would prefer, that is not what this court reviews. Our review is deferential, not de novo. See Topp, 83 Wis. 2d at 785, 266 N.W.2d at 401. I conclude that the trial court did not erroneously exercise its discretion in wording the third question of the breach of contract verdict as it did. And, given the focus of the trial, I have no doubt but that the form of the verdict fairly presented the material issues of fact to the jury.
*629There is a problem, however, with this final breach question. Once the jury found the first breach and awarded damages, the facts show that there would be either no breach of the contract by HABCO's failure to compensate MCS for its use of the proprietary software or no damages for the breach. The contract provided that if HABCO bought additional computers from MCS, MCS would provide HABCO with proprietary software at no extra charge. The jury's affirmative response to the first breach of contract question put MCS in the financial position in which it would have been had HABCO bought additional computers from MCS. MCS would then have provided the proprietary software to HABCO at no cost to HABCO. I would change the answer to breach question number three to "no."
INDEFINITE CONTRACT
The majority concludes that the parties' contract is too indefinite to be enforced. Yet, it never quotes the part of the contract which it believes is indefinite. Apparently, the majority has adopted a new rule of contract law to the effect that a contract is indefinite if it does not mean what one of the parties contends that it means. I am unaware of such a rule. We look to the contract itself to determine whether it is indefinite and therefore unenforceable. Arthur Corbin notes:
A court cannot enforce a contract unless it can determine what it is. It is not enough that the parties think that they have made a contract. They must have expressed their intentions in a manner that is capable of being understood. It is not even enough that they have actually agreed, if their expressions, when interpreted in the light of accompanying factors and circumstances, are not such *630that the court can determine what the terms of that agreement are.
1 Arthur Linton Corbin, Corbin on Contracts § 4.1 (1993). This is consistent with the RESTATEMENT (SECOND) of Contracts § 33(2) (1981), which provides: "The terms of a contract are reasonably certain if they provide a basis for determining the existence of a breach and for giving an appropriate remedy." The majority holds: "When the court subsequently ruled that the contract was insufficiently definite to be enforced, as MCS would have it enforced, MCS was left with the quoted question and award based on a theory of the case which no longer applied." Majority op. at 597 (emphasis added). This new holding looks to the pleadings and the positions taken at trial to determine whether a contract is indefinite. The majority cites no authority for this dramatic change in contract law, and I find none. I believe that the proper test for indefiniteness is to look at the language of the contract to determine whether the contract is too indefinite to be enforced.1 There is no question but that the contract is sufficiently definite. Even the majority notes: "Since it is undisputed that HABCO used MCS software with computers it bought from vendors other than MCS, it is arguable that HABCO breached the contract when it used MCS software on non-MCS computers." Majority op. at 596 (emphasis added).
Once we conclude that a contract was formed, and that HABCO breached it, facts that even the majority grudgingly admits, the only question becomes the *631extent of MCS's damages. Because there is evidence to support the damages the jury awarded for HABCO's failure to purchase computer equipment from MCS, and for HABCO'S failure to pay twenty-five percent of the program value for the use of the contract software, I would reinstate that part of the jury's verdict.
PUNITIVE DAMAGES
The only rationale that I can discern for the majority's reduction of the $1,750,000 punitive damage award is that the figure it chooses, $650,000, is ten times the amount of the reduced compensatory damages. But why is ten, aside from being a round number with metric significance, the proper multiplier? Would not eleven or nine be just as appropriate? And if " [t]here is no arbitrary rule that punitive damages cannot equal 15 times the compensatory damages," Malco, Inc. v. Midwest Aluminum Sales, Inc., 14 Wis. 2d 57, 66, 109 N.W.2d 516, 521 (1961), why did not the majority award punitive damages of $975,000?
The use of a multiplier as the sole means to determine punitive damages has been specifically rejected. In Fahrenberg v. Tengel, 96 Wis. 2d 211, 235-36, 291 N.W.2d 516, 527 (1980), the court said:
Although the amount of compensatory damages and criminal penalties have some relevancy to the amount of punitive damages and may be factors in determining the reasonableness of the punitive damages award, we have not been willing in the past, and are not willing in this case, to adopt a mathematical formula for awarding punitive damages. In punitive damages, as in damages for pain and suffering, the law furnishes no mechanical legal rule for their measurement. The amount rests initially in the discretion of the jury. We are reluc*632tant to set aside an award because it is large or we would have awarded less. As we have said in cases involving compensatory damages, " '[A]ll that the court can do is to see that the jury approximates a sane estimate, or, as it is sometimes said, see that the results attained do not shock the judicial conscience.'"
(Emphasis added; quoted source omitted.)
The jury saw the witnesses and heard what HABCO did. I cannot conclude that a $1,750,000 punitive damage award against accountants who have stolen their client's software was an insane estimate by the jury. We give juries discretion in their award of punitive damages. I agree that the award is large and I would have awarded less. But that is not the test. We must look at whether "the jury approximates a sane estimate," id. at 236, 291 N.W.2d at 527, and, using that test, I would affirm the jury's punitive damage award. The message sent by the majority in a world where computers have provided extensive profits to those who can market the technology, is that crime pays, and pays well. The decision to risk $650,000 by stealing software can be an easy one where the profits can reach millions of dollars. Perhaps a $1,750,000 punitive damage award is not much better than a $650,000 award, but it is a start. It sends the message that courts will not reverse large punitive damage awards where the conduct is criminal and egregious. It allows for the possibility that even larger awards will be sustained when the conduct merits it. And that possibility will, perhaps, give potential computer thieves pause when they contemplate obtaining desired software by theft.
*633For these reasons, I respectfully dissent.
Even if we conclude that a contract is ambiguous, we do not necessarily conclude that it is void for indefiniteness. If a contract is ambiguous, we may construe the contract through the use of extrinsic evidence. Pleasure Time, Inc. v. Kuss, 78 Wis. 2d 373, 379, 254 N.W.2d 463, 467 (1977).