OPINION AND ORDER RE DISCOVERY
WAYNE D. BRAZIL, United States Magistrate.The Court addresses here complex issues of first impression. Plaintiff seeks to discover from defendant documents of two different kinds: (1) drafts of a declaration that defendant’s expert, Maurice F. Holmes, submitted to the Patent Office as part of defendant’s effort to survive a reexamination of the patent whose validity plaintiff is challenging in this litigation, and (2) a draft of a “Reply Under C.F.R. sec. 1.111” that outside counsel for defendant prepared at the request of defendant’s in-house lawyer (after making revisions in the draft, defendant’s in-house counsel submitted the Reply to the patent examiner for consideration in the reexamination proceedings). For reasons that will become clear, the court will consider the two different kinds of documents in separate sections of this opinion. At the outset, however, it is important to describe briefly the procedural setting in which these issues arise.
This is a struggle of potentially significant economic proportions between two large corporations. Formally it began some three years ago when plaintiff filed its Petition for Reexamination in the Patent and Trademark Office (referred to hereaf*535ter as PTO). Through that Petition plaintiff challenged the validity of the patent here in suit, a patent currently owned by defendant. In particular, plaintiffs petition sought to persuade the PTO that certain prior art rendered defendant’s patent invalid. One month after it initiated the reexamination proceeding in the PTO plaintiff commenced this litigation. Plaintiff’s objective in this lawsuit is identical to its objective in the reexamination proceeding: namely, to have defendant’s patent declared invalid. There also is an important functional interplay between the two proceedings. The reexamination proceeding in the PTO has been completed. It resulted in reaffirmation (the route was not linear) of the validity of defendant’s patent. At trial of this lawsuit defendant plans to rely in part on that reaffirmation, while plaintiff plans to attack it. It is quite clear that both parties were well aware from the outset that the reexamination proceeding and this litigation were closely intertwined (at one point the district court stayed certain aspects of the development of this suit pending conclusion of the proceedings in the PTO) and that the result of the reexamination could play a significant role in the trial of this case.
It also is important to emphasize that this is a struggle that is in every meaningful sense of the phrase fully “adversarial”. Both parties have committed great resources to this combat. Both are well-represented by sophisticated and energetic counsel. Counsels’ work before this court demonstrates beyond doubt that there has never been a dearth of incentive in either party to pursue its interests vigorously in the PTO and to litigate this matter with great intensity.
DRAFTS OF THE HOLMES DECLARATION
Maurice F. Holmes is not a lawyer. He is Manager of Reprographic Development for Xerox Corporation and an expert in paper and paper handling technology. Defendant retained Mr. Holmes to act as its expert consultant in both the reexamination proceeding in the PTO and in this litigation. Mr. Holmes is expected to testify at trial about matters identical to those covered in the declaration he submitted on defendant’s behalf to the PTO during the reexamination. This is a technologically dense case in the outcome of which expert testimony will play a pivotal role.
The subjects about which Mr. Holmes has formed expert opinions, and which he addressed in his declaration to the PTO, are at the center of this litigation: in both settings he attempts to distinguish the arguably relevant prior art and to show why it does not render the claims in defendant’s patent “obvious”. Thus the drafts of the declaration that are in issue here cover material that is pivotal to the resolution of this suit and about which Mr. Holmes is expected to testify at trial.
Moreover, the ostensible subject both of Mr. Holmes’ declaration and of his contemplated trial testimony is not legal but engineering. He describes physical differences between prior art and the matter covered by the patent in suit and draws inferences about mechanical processes from those differences. In other words, his declaration and his testimony focus on the real world facts which should hold the key to the outcome of this case.
The Relevance of the Drafts of the Holmes’ Declaration.
Plaintiff’s counsel seek the drafts of Mr. Holmes declaration for a number of purposes. One is to enrich their ability to cross-examine him at trial in the pivotal subject area about which he will be testifying. Given the central role experts will play in determining the outcome of this case, that purpose is important.
There is a second context in which these drafts might be relevant at trial. Defendant concedes that it will attempt to use at trial the favorable outcome of the reexamination. Because plaintiff will attempt to get whatever mileage it can out of that result, plaintiff has an interest in attacking its significance. In particular, plaintiff hopes to show that the information and argument presented to the PTO by defend*536ant (through Mr. Holmes) was incomplete and misleading, thus compromising the reliability of the examiner’s finding. Plaintiff hopes that drafts of the Holmes’ declaration will show that Mr. Holmes and defendant decided not to disclose certain matters to the examiner and that they packaged the material they did present in a manner that did not do justice to the arguments in favor of the plaintiff’s view of the prior art. I hasten to add that I do not know whether the drafts of the Holmes’ declaration will even indirectly support plaintiff’s theories. But this is the discovery stage, and the court’s role is not to speculate about what more sophisticated eyes might find in such documents. The question at this juncture is not whether defendant and Mr. Holmes made fair and forthright presentations to the patent office; rather, the question is whether plaintiff should be given access to documents that would help it, and, ultimately, the trier of fact in this lawsuit, assess the quality of those presentations.
Are the Drafts of the Holmes’ Declaration “Work Product” Within the Meaning of Federal Rule of Civil Procedure 26(b)(3) or Should Their Discoverability Be Governed by Rule 26(b)(4)?
In briefing this matter the parties have assumed that the issue before the court is whether the drafts of the Holmes’ declaration meet the criteria which define “work product” within the meaning of Rule 26(b)(3). In particular, the parties have struggled over whether it is fair to conclude that these drafts were “prepared in anticipation of litigation or for trial” as that Rule requires. In thinking through this problem, however, the court has realized that it must first address a different and potentially dispositive question. That question is: should the discoverability of the drafts of the Holmes’ Declaration be governed by the provision in Rule 26 [sub-paragraph (b)(4)(A)] that covers discovery from a person “whom the other party expects to call as an expert witness at trial”? And even if the answer to that question is not clearly “yes”, should the court, in resolving this dispute, attend to the policies that inspire Rule (b)(4)(A)?
It is quite clear that Rule 26 does not include within the definition of “work product” documents generated or consulted by experts retained in connection with litigation. The discoverability of such material is not governed by Rule 26(b)(3). The Advisory Committee’s Note accompanying the 1970 amendments to Rule 26 states explicitly that the “new provisions of subdivision (b)(4)” “reject as ill-considered the decisions which have sought to bring expert information within the work-product doctrine.” Advisory Committee’s Note re 1970 amendment to Rule 26, reported at 48 F.R.D. 487, 504-505. Many courts have recognized this important distinction and have refused to treat opinions generated by retained experts, or the documents they generate or consult in forming their opinions, as “work product”. See, e.g., Beverage Marketing Corp. v. Ogilvy & Mather, 563 F.Supp. 1013, 1014 (S.D.N.Y.1983); Quadrini v. Sikorsky Aircraft, 74 F.R.D. 594 (D.Conn.1977). Instead, courts determine the discoverability of materials generated or consulted by retained experts under principles that emanate from the separate provisions of Rule 26(b)(4).
That subparagraph of Rule 26 governs the discovery of “facts known and opinions held by experts .. acquired or developed in anticipation of litigation.” Fed.R.Civ.P. 26(b)(4). Many courts have construed this subparagraph to authorize orders compelling experts to disclose not only the opinions they hold, but also all the documents the expert generated or examined in the process of forming those opinions. See, e.g., Quadrini v. Sikorsky Aircraft, supra, and Eliasen v. Hamilton, 111 F.R.D. 396, 400, note 5 (N.D.Ill.1986). In particular, this subparagraph has been used as the source of authority to order disclosure of drafts of reports or memoranda experts have generated as they develop the opinions they will present at trial. Quadrini v. Sikorsky Aircraft, supra.
It is significant that the principal rationale for ordering disclosure of such material is to better equip opposing counsel to *537cross-examine the expert. See Heitmann v. Concrete Pipe Machinery, 98 F.R.D. 740, 742-743 (E.D.Mo.1983); In re IBM Peripheral EDP Devices Antitrust Litigation, 77 F.R.D. 39, 41 (N.D.Cal.1977); contra, Breedlove v. Beech Aircraft Corp., 57 F.R.D. 202, 205 (N.D.Miss.1972). The Advisory Committee Note to subdivision (b)(4)(A) points out that this part of the Rule “is responsive to problems” encountered in cases that
“present intricate and difficult issues as to which expert testimony is likely to be determinative. Prominent among them are food and drug, patent, and condemnation cases, (citations omitted). In cases of this character, a prohibition against discovery of information held by expert witnesses produces in acute form the very evils that discovery has been created to prevent. Effective cross-examination of an expert witness requires advance preparation. The lawyer even with the help of his own experts frequently cannot anticipate the particular approach his adversary’s expert will take or the data on which he will base his judgment on the stand.” 48 F.R.D. 503.
Heeding the spirit of this commentary on the Rule, federal courts have been especially sensitive to the need to order disclosure of materials generated or consulted by experts in cases like the one at bar, where the issues are complex and technical and where testimony by experts is likely to play a key role in fixing outcome. For example, in Quadrini v. Sikorsky Aircraft, supra, Judge Newman supported his disclosure order with the following observation:
“Expert testimony will undoubtedly be crucial to the resolution of the complex and technical factual disputes in this case, and effective cross-examination will be essential. Discovery of the reports of experts, including reports embodying preliminary conclusions, can guard against the possibility of a sanitized presentation at trial, purged of less favorable opinions expressed at an earlier date.” 74 F.R.D. at 595.
Thus, the drafts of the Holmes’ declaration would not be deemed “work product” and clearly would be discoverable if the declaration had been prepared directly for use in this litigation (as opposed to being prepared for submission in the first instance to the PTO), or if they were simply drafts of memoranda he had prepared en route to formulating the opinions he will present at trial. That being the case, I believe it would exalt form over substance to rule that the drafts are not discoverable simply because the proceeding for which they were most immediately prepared was the reexamination rather than this lawsuit.
My conviction about this is reinforced by a number of considerations. One is that the reexamination proceeding and the litigation commenced at very nearly the same time and were viewed by all concerned as closely related parts of the same larger struggle. Moreover, Mr. Holmes was not asked to formulate the opinions reflected in his declaration until months after the lawsuit was filed, when it was clear that the central issue in both forums was identical and was the issue on which Mr. Holmes was asked to render an opinion.
Another consideration that supports the conclusion reached here is that it appears that defendant will in some measure rely on the final version of the Holmes declaration during the trial of this matter. If so, the declaration will become part of the presentation of Mr. Holmes’ expert opinion. If it becomes part of the presentation of his opinion the drafts clearly fall within the line of cases that compel discovery of documents generated by experts in the process of forming their opinions.
But even if the Holmes’ declaration is not formally introduced into evidence during the trial it is clear that the substance of the opinions reflected in that declaration will be at least very similar (if not identical) to the substance of the opinions he articulates from the witness stand on the central issues in the case. This follows because the central issue before the PTO was identical to the central issue to be resolved at trial. And defendant retained Mr. *538Holmes to formulate and express opinions on this one key subject not only in the reexamination proceeding, but also in this lawsuit. In this setting disclosure of the drafts of the declaration serves exactly the same purpose (better equipping plaintiffs counsel to conduct cross-examination, and thereby improving the reliability of the truth-finding process) as would disclosure of other documents Mr. Holmes generated as he developed his expert opinions for defendant. Thus it would violate the policies that inform Rule 26(b)(4) not to order disclosure of these drafts.
Ironically, defendant has gone to great lengths in the papers it submitted in connection with this dispute to persuade the court that the Holmes declaration was prepared “with an eye toward litigation” and not solely for purposes of the reexamination. That fact, which the court accepts, supports the conclusion herein reached.
Defendant has argued that compelling disclosure of the drafts of the Holmes’ declaration would do great violence to the values that the work product doctrine was developed to protect. Even though the conclusion reached in the preceding paragraphs makes it unnecessary to do so, I will address these arguments in case a higher court is not sure that the discoverability of these drafts should be governed by Rule 26(b)(4).
The work product doctrine reflects an attempt to resolve tensions between important competing values. As originally articulated in Hickman v. Taylor, 329 U.S. 495, 67 S.Ct. 385, 91 L.Ed. 451 (1947), the purpose of the work product doctrine was to preserve “the historical and the necessary way in which lawyers act within the framework of our system of jurisprudence to promote justice and to protect their client’s interests.” Id., at 511, 67 S.Ct. at 393. In developing the work product doctrine the Hickman court focused exclusively on the privacy interests of lawyers, on shielding “the privacy of an attorney’s course of preparation”. Id., at 512, 67 S.Ct. at 394. The Court sought to protect this privacy interest not out of some abstract commitment to natural rights, but because the Court assumed that protecting the case preparation privacy of counsel was necessary to preserve the incentive system on which it assumed the adversary system depends. That incentive system has two closely related dimensions. One consists of controlling fear, the other of attacking temptation. The fear that the work product doctrine is designed to cabin is the lawyer’s fear that his opponent will gain access to his ideas, his perceptions, his assessments, his plans. The temptation the doctrine is designed to attack is the temptation individual lawyers might feel to be lazy, to let opposing counsel do the investigatory homework, then force him to disgorge it.
The Hickman Court appreciated that our system of adjudication is at its heart dialectical. If one side relies (for ideas or information or legal research) on the other, half of the dialectic is lost. Half a dialectic is no dialectic at all. Thus the principal purpose of the work product doctrine is to preserve the vigor of the dialectic by creating an environment in which each lawyer feels not only free, but pressured, to do his own work and to make both his investigation and his evaluation of the case as aggressive, as thorough, and as internally frank as possible.
When the Advisory Committee drafted the codification of the work product doctrine that became Rule 26(b)(3) in 1970 it extended the doctrine so that it also reached documents generated in anticipation of litigation by persons other than lawyers (i.e., clients or their other agents). But two considerations support the conclusion that the most important values protected by the work product doctrine relate to work by lawyers. First, courts have continued (well after the adoption of the 1970 amendments to Rule 26) to view the overriding purpose of the work product doctrine to be “to protect the integrity of the adversary system” Hercules Inc. v. Exxon Corp., 434 F.Supp. 136, 151 (D.Del.1977). Lawyers are the motor force of the adversary system. The primary concern of *539a doctrine whose purpose is to protect our adversary system must be with lawyers and their work, not with other agents of clients.
Second, it is important to emphasize again that when they codified the work product doctrine into Rule 26(b) the rule-makers decided to remove entirely from its ambit work done by the category of agents of clients that is undoubtedly the second most important to the vitality of the adversary system: namely, retained experts. The work that retained experts do in anticipation of litigation is not work product and its discoverability is governed by a separate rule [26(b)(4) ]. Thus the principal concern of the work product doctrine remains work by lawyers and the lawyering process.
Compelling disclosure of the drafts of the Holmes declaration would not threaten values at the center of the work product doctrine in part because Mr. Holmes is not a lawyer and his declaration consists of descriptions of physical facts and of engineering reasoning, not of legal reasoning or litigation strategy. In other words, the Holmes declaration is primarily evidence, not law or lawyering. The work product doctrine was designed to protect the lawyering process, not to prevent access to the raw evidence or facts on which the lawyering process is supposed to operate. There are countless cases where courts have held that the work product doctrine may not be used to block access to otherwise unattainable facts or evidence. See Hickman v. Taylor, supra, 329 U.S. at 511, 67 S.Ct. at 394; and cases cited in 8 Wright, Miller and Elliot, Federal Practice and Procedure, section 2025 (most recently supplemented in 1986).
Defense counsel might counter by arguing that this characterization of the Holmes declaration is naive. Counsel might argue that lawyers draft experts’ declarations, especially in sophisticated patent litigation like this, and therefore that the drafts of the Holmes’ declaration are quintessential lawyer work product. Or defense counsel might argue that declarations like the one in issue here are the product of intense and very private dialogue between retained expert and lawyer, a dialogue in which the expert contributes the raw data and the lawyer packages it into a form that is likely to go the farthest toward satisfying the relevant legal standard. Counsel might contend that in patent litigation, especially where outcome turns on interpretation of prior art, the line between physics and law is so thoroughly blurred that to characterize a document like the Holmes declaration as consisting of engineering reasoning rather than legal reasoning is to misperceive the nature of this litigation and to abuse reality.
There are several levels at which the court can respond to arguments like these. The first is to emphasize whose signature is on the Holmes declaration. It is not counsel’s. To be of any real utility to the truth finding process, an expert’s opinion should be honest and his own. The court is not interested in furthering the corruption of the truth finding process by announcing doctrine that has the effect of approving and reinforcing the practice of lawyers formulating and writing opinions that are presented to the outside world as the independent opinions of “technical experts”. Counsel who believe they are technical experts and who want to present their expert opinions should submit declarations over their own signatures or take the stand and testify. It is worth noting here that one of the purposes of the work product doctrine as articulated in Hickman v. Taylor was to protect lawyers from being converted into witnesses. Justice Murphy pointed out that if a lawyer’s notes from witness interviews were routinely discoverable and thus available for use “for impeachment or corroborative purposes” the effect would be to “make the attorney much less an officer of the court and much more an ordinary witness. The standards of the profession would thereby suffer.” Hickman v. Taylor, 329 U.S. 495, at 513, 67 S.Ct. 385, at 394, 91 L.Ed. 451 (1947). Thus one purpose of the work product doctrine was to preserve a clean separation between the role *540of lawyer and the role of witness. It would hardly be consistent with that purpose to permit lawyers to use the work product doctrine to hide the fact that they are secretly functioning as witnesses by writing declarations that purport to be from experts.
Moreover, the PTO and juries have a right to know the identity of the source of the opinions presented to them. Where the outcome of a case is likely to turn on the testimony of competing experts, the trier of fact has a right to know everything that the experts considered in the process of formulating their opinions. Since the job of the trier of fact is to decide which expert’s opinion is more reliable and compelling, it is difficult to imagine there being more important documents for the trier of fact to see than drafts of the expert’s opinion that were written or edited by someone else, e.g., by the lawyer who retained him.
Defense counsel argues that compelling disclosure of drafts of expert's declarations will chill or constrain the give and take between lawyer and retained expert that precedes the crafting of expert declarations. That may be. It is not clear to me, however, that the system will be worse off if one effect of compelling such disclosure is to make the process by which such declarations are prepared less of a dialogue between expert and counsel. In fact, I suspect that the truth finding process would be just as well served if lawyers played a lesser role in the formulation of other people’s opinions.
Defense counsel also has argued (with something less than full consistency) that the only effect in the real world of compelling disclosure of these kinds of documents will be to increase dramatically their destruction (Transcript of hearing on June 4, 1987, at 24, where counsel said: “All that his policy decision will do is make sure that people destroy their drafts.”). Accepting the full implications of this line of argument would compel repeal of the entire system of civil discovery. Documents are not immune from discovery merely because they are damaging to the party from whom they are sought. Yet, remarkably, lawyers and clients routinely honor their obligation not to destroy, but instead to produce in response to discovery, damaging documents. It also should be emphasized that this argument by defense counsel implicitly has been rejected by the substantial number of courts that have compelled disclosure of all the materials (including drafts of reports and memoranda) that experts have consulted while developing their opinions. See, e.g., Quadrini v. Sikorsky Aircraft, supra, at 594; and Carter-Wallace, Inc., v. Hartz Mountain Industries, 553 F.Supp. 45, 52 (S.D.N.Y.1982). Moreover, not all drafts of experts’ reports are destroyed despite the fact that it is well known that such drafts are discoverable. And if the choice is between adopting a rule that would block discovery of such drafts, on the one hand, and, on the other, a rule that would permit discovery of such drafts when they exist, the court favors the alternative that will at least occasionally result in enrichment of the truth finding process.
Finally, I emphasize again that it is very clear that all counsel in this litigation are energized by powerful incentives to lawyer this case as aggressively and as thoroughly as possible. There is literally no risk that the ruling made here will undermine those incentives or diminish the vigor with which every aspect of this case will be litigated. Nor is the possibility that drafts of documents of this kind will be discoverable likely to have any meaningful adverse effect on incentives parties feel to develop full presentations for the PTO during reexamination proceedings. The risk that the party will lose his patent rights by an adverse decision from the PTO usually will constitute, as it surely did here, considerable incentive. Moreover, as Judge Wright’s thoughtful opinion in Hercules Inc., v. Exxon Corp., supra, at 152, points out, parties proceeding before the PTO have an independent duty of full disclosure that can serve as at least a partial substitute for the incentives that the work product doctrine was designed to foster.
*541In sum, the court is not persuaded that the result reached here poses a serious risk of harm to the principal values the work product doctrine was designed to protect. As important, the rule here adopted could advance significant competing interests. The court hopes that permitting searching discovery into the process by which experts develop the opinions they present to the PTO will help discipline the reexamination process. Knowing that the process leading to the formulation of their opinions would remain largely insulated from searching review might make technical experts sloppier in their reasoning or more vulnerable to the temptation to be less than fully frank with the PTO in order to increase the odds of a favorable decision for the client who retained them. By contrast, an expert who knows that the process by which he developed his opinions (the data he considered and failed to consider, the interpretations he adopted and those he rejected, the logic he employed, etc.) will be subject to thorough scrutiny by a highly motivated peer is more likely to proceed carefully and less likely to feel that he can get away with some conceptual slight of hand. Thus the rule adopted here is designed to intensify the pressures and incentives experts and parties will feel to disclose all relevant information to the PTO during ex parte or quasi ex parte proceedings.
Rules that result in more of the relevant information being presented to the PTO should enhance the quality of the PTO’s deliberations and the reliability of its decisions. Since the PTO’s decisions affect valuable property interests, rules that improve the reliability of the PTO’s work contribute to an important end. Moreover, there probably is some correlation between the quality of the PTO’s decisions and both the amount and outcome of litigation over intellectual property rights. One would hope that the volume of intellectual property litigation would decrease as the quality of the PTO’s work, and public confidence in the reliability of PTO procedures, increases. And to the extent that the results of PTO proceedings are introduced into evidence at trial, improving the reliability of the PTO’s determinations should make for more just results in the litigation that is pursued. Thus, to the extent that it fosters an environment in which more information and higher quality reasoning are presented to the PTO, the rule adopted here contributes to important interests that outweigh the marginal harm done to values that the work product doctrine was designed to protect.
By compelling disclosure of the drafts of the expert’s declaration the court also creates an opportunity to directly improve the quality of the truth finding process at trial. Mr. Holmes will be a key expert witness. Access to the drafts of his declaration, a document that addresses the central issues in this case, should equip opposing counsel to do a better job of fully elucidating the process through which Mr. Holmes formulated his views. The more light that is shed on that process the better the odds that justice will be done at trial.
In short, the rule announced here could contribute significantly to interests of substantial magnitude. These competing considerations clearly outweigh whatever marginal harm the rule might do to values that the work product doctrine was designed to serve. Thus, even if the discoverability of these drafts should not be governed by principles emanating from Rule 26(b)(4), the analysis of work product issues offered in the immediately preceding paragraphs serves as an independent basis for compelling disclosure of these documents.
This disposition of the matter makes it unnecessary to reach the issue of waiver raised by plaintiff’s counsel.
OUTSIDE COUNSEL’S DRAFT OF THE “REPLY UNDER 37 C.F.R.
SECTION 1.111”
For the reasons explained below, the court has concluded that the draft of the “Reply” that outside counsel (Mr. Clark) submitted in confidence to Mr. Robbins, defendant’s house counsel, is protected from discovery by the attorney-client privilege.
*542While there is some older, now rejected authority to the contrary, see e.g., Zenith Radio Corp. v. Radio Corporation of America, 121 F.Supp. 792, 794 (D.Del.1954), and American Cynamid Co. v. Hercules Power Co., 211 F.Supp. 85 (D.Del.1962), it is clear today that the attorney-client privilege can attach to documents generated in connection with proceedings before the PTO. Hercules Inc. v. Exxon Corp., supra, at 147, citing Sperry v. State of Florida, 373 U.S. 379, 83 S.Ct. 1322, 10 L.Ed.2d 428 (1963). The document in issue here was written when a client (defendant, speaking through house counsel Robbins) asked its outside counsel, Mr. Clark, to prepare and to submit in confidence, a proposed “Reply Under 37 C.F.R. SECTION 1.111.” The resulting draft clearly is a confidential communication from attorney to client. Whether it is protected by the attorney client privilege turns on the nature of the communication and the principal purpose for which it was made.
A communication that consists largely of technical information that an agent of a client provides to the client’s representative before the PTO so that that representative can pass along the technical information to the PTO generally is not covered by the attorney-client privilege. Hercules Inc. v. Exxon Corp., supra, at 148; Jack Winter, Inc., v. Koratron Co., Inc., 54 F.R.D. 44, 46-47 (N.D.Cal.1971). Such a communication does not constitute legal advice or a request for legal advice, and the principal purpose for making the communication is to not to secure legal advice, but to secure what is essentially a business service (that is, passing along the information to the PTO). In drafting a document of this kind an attorney is, by hypothesis, merely serving as a conduit for factual information. He is not acting primarily as a lawyer. Hercules Inc. v. Exxon Corp., supra, at 147.
If we were to look only at thé content of the document in question here, and ignore the context in which it was prepared, it might be more difficult to conclude that the draft is covered by the attorney-client privilege. But analysis of this kind of issue cannot fairly focus only on the “nature” (or content) of a document. Sound analysis also requires consideration of the purpose for which and the context in which the document was generated.
The court has conducted an in camera examination of the document. It has several components. First, very briefly, it asks for a straightforward, grammatical amendment to one of the claims in the patent. This request, made in response to an observation by the examiner, makes up only a very small part of the document, which is dominated by other matters. The court therefore concludes, contrary to arguments advanced by plaintiff’s counsel, that it is not accurate to characterize this document as nothing more than “an amendment in the patent office.” (Transcript of Hearing of June 4, 1987, at 34). It follows that authorities that suggest that the attorney-client privilege does not cover drafts of requested amendments to patents, submitted as part of ex parte proceedings before the PTO, do not control disposition of the issue here presented. Cf. Choat v. Rome Industries, 462 F.Supp. 728, 732 (N.D.Ga.1978), an opinion this court finds unpersuasive in any event (see below).
The vast majority of the draft consists of characterizations of positions taken by the examiner during the ex parte reexamination interview, characterizations of responses to those positions presented during the interview by defendant and its agents (including counsel and Mr. Holmes), and new or recast responses to the reasoning on which the examiner had based his tentative decision to deny certain claims. So far as the court can tell, none of these paragraphs contain new factual material or raw data. Instead, most of the document consists either of restatements of arguments presented in private to the examiner during the interview or of new or newly phrased arguments developed after the interview.
Superficially considered, some of this material might not appear to consist of argument. But a more careful consideration, in context, persuades the court that it does. *543First, it is important to keep in mind that this “Reply” was written after the examiner had tentatively decided to reject key claims in defendant’s patent. Thus when the document was written attorney Clark’s client was in a defensive position and, in essence, in an adversary relation to the examiner. Clearly the purpose of the document was to try to persuade the examiner to change his mind.
Moreover, these paragraphs represent an effort by the claimant, through counsel, to persuade the examiner to reason from the raw data (published information about certain prior art) toward certain debatable inferences or conclusions. While it is arguable that some of the material in this draft “Reply” consists only of descriptions of facts, somebody decided which facts to describe and which words to use to describe them. That somebody was a lawyer, and in the world of patent litigation, as preoccupied as it is with words, the court can safely assume that a great deal of calculation went into the choices of words here. It is important to emphasize that the criteria that dominate this kind of calculation ultimately are quasi-legal, not scientific.
The lawyer who wrote this material was not engaged in a dispassionate exercise in cataloguing differences everyone would concede existed between patent and prior art. First, he was trying to persuade the examiner to draw inferences, from ambiguous words and from the context in which they appeared, about the physical character of earlier invented processes. In other words, he was trying to persuade the examiner that certain differences existed. In addition, he sought to focus the examiner’s attention on certain other differences that he thought were “significant.”
Most important, in this context the criteria that define “significance” are at least quasi-legal. The target of the persuasive effort was the quasi-legal norm of “obviousness”. What attorney Clark was doing was selecting and verbally packaging certain facts, and crafting arguments about what the facts were, so as to maximize the odds that the examiner would reach a favorable conclusion when he compared the facts as he found them to the norm of “obviousness”. There is a great deal of subtle reasoning in these paragraphs, and it is reasoning with a clear legal purpose. This is the essence of lawyering. Clients pay patent lawyers to show them how to reason in the highly specialized, esoteric manner that examiners are likely to find persuasive. Thus, attorney Clark’s draft “Reply” consisted of his quintessentially professional advice to his client about how best to present its case, given the quasi-legal norm that the examiner would use to decide the outcome.
Thus the court concludes that what defendant asked for from attorney Clark, and what it received in the form of this draft “Reply”, was confidential legal advice. As such, it is protected from discovery by the attorney-client privilege. Cf. Jack Winter, Inc., v. Koratron Co., Inc., supra, 54 F.R.D. at 46.
In reaching this conclusion the court declines to follow the conclusory pronouncements in some opinions that imply, at least if taken literally, that the attorney-client privilege simply is not available to documents prepared in connection with ex parte proceedings before the PTO. See, e.g., Detection Systems, Inc., v. Pittway Corp., 96 F.R.D. 152, 155 (W.D.N.Y.1982); Sneider v. Kimberly-Clark Corp., 91 F.R.D. 1, 7 (N.D.Ill.1980); and Choat v. Rome Industries, Inc., 462 F.Supp. 728, 732 (N.D.Ga.1978) [misreading Sperry Rand Corp. v. International Business Machines Corp., 45 F.R.D. 287, 291 (D.Del.1968)]. Rather, the court endorses the view that in certain circumstances this privilege can be available to documents so generated and that to decide whether to sustain any given assertion of this privilege courts should engage in a situation-specific analysis that is sensitive to the competing interests that are involved. See Hercules Inc. v. Exxon Corp., supra, at 146-149, Sperry Rand Corp. v. International Business Machines Corp., supra, at 291. Moreover, as explained in detail below, a reexamination proceeding is not as com*544pletely ex parte as a typical application for initial issuance of a patent. Thus the arguments that have been advanced against recognizing privileges in original application proceedings may not be as persuasive with respect to reexaminations.
IS THE CLARK DRAFT OF THE “REPLY UNDER 37 C.F.R. SECTION 1.111” PROTECTED BY THE WORK PRODUCT DOCTRINE AS WELL?
Despite the fact that resolution of the claim of attorney client privilege in favor of defendant makes it unnecessary to do so, the court addresses this question because there is no published opinion on point and counsel may need guidance in this area with respect to other documents.
The Clark draft was written by a lawyer on behalf of a client and its confidential character has been preserved. Thus, formally, deciding whether it qualifies for some level of protection under the work product doctrine turns on whether it can fairly be said that the document was “prepared in anticipation of litigation or for trial” within the meaning of Rule 26(b)(3).
There is authority for the view that documents like this are not protected by the work product doctrine when they are drafted in connection with an ex parte application for initial issuance of a patent or in connection with an ex parte re-issue proceeding, at least when it is clear that “the attorney's primary concern” in preparing the document is with “claims which have arisen or will arise during the ex parte prosecution of the application” rather than with issues “more relevant to future litigation than to the ongoing prosecution.” Hercules Inc. v. Exxon Corp., supra, at 152. In fact, some courts have gone so far as to declare that the work product privilege can never attach to documents prepared in connection with ex parte proceedings before the PTO. See, e.g., Choat v. Rome, supra, at 732, citing Interlego A.G. v. F.A.O. Schwarz, Inc., 196 U.S.P.Q. 8, 12 (N.D.Ga.1977).
The rationale for denying access to the work product doctrine in these environments seems to have two related components. One is that courts should adopt rules that pressure parties to make full disclosure to the PTO in ex parte proceedings, in part because the absence of an adversary in such proceedings intensifies the risk that parties will make sloppy or disingenuous presentations and thus compromise the reliability of the decisions the PTO makes. The second component of the rationale for denying work product protection in this setting seems to run like this: a key purpose of the work product doctrine is to preserve lawyers’ incentives to prepare their cases thoroughly, but in ex parte proceedings before the PTO parties are under a separate affirmative duty to make full disclosure, therefore there is no need for the work product protection. See Hercules Inc. v. Exxon Corp., supra, at 152.
As earlier parts of this opinion demonstrate, this court obviously is sympathetic with arguments that focus on the importance of crafting rules that pressure parties to honor disclosure obligations. But before relying on such arguments to justify carte blanche denial of protections to which lawyers attach considerable signifi-. canee it is important to make a judgment about how much denying that protection is likely to contribute toward that end in the specific setting of a reexamination proceeding.
A reexamination is nowhere nearly as non-adversarial and as ex parte as is a typical application for initial issuance of a patent. In this case, for example, the reexamination was initiated not by the patent holder (defendant), but by its competitor in the marketplace and its adversary in this litigation. And when plaintiff initiated the reexamination it submitted a substantial brief attacking the validity of defendant’s patent, arguing in great detail from specific references to prior art that the principal claims covered by the patent were obvious at the relevant historical time. Thus, at the outset of the proceedings, the examiner had the benefit of a major adversarial submission by a sophisticated commercial enemy of the patent holder. If, as happened *545here, the challenger’s papers persuade the examiner that there is a “substantial new question of patentability” the patent holder has a two month grace period during which it may elect to file a responsive statement that can include, among other things, amendments and new claims. 35 U.S.C. sec. 303(a). If the owner files such a statement, the challenger is given an opportunity to file a reply. If the owner elects not to file such a response, the process moves into an ex parte phase that is like the prosecution of an application for initial issuance of a patent, except that the PTO sends to the challenger copies of Office Actions and of responses submitted by the owner.
Moreover, reexaminations often are conducted within a larger context that is dominated by a lawsuit in which the challenger is seeking simultaneously to achieve the same end (invalidation of the owner’s patent). Thus, unlike the situation typically associated with processing an application for initial issuance of a patent, the whole environment is very adversarial. Moreover, the key issues in the two parallel proceedings are identical.
In this environment there is less risk that an owner will be able to dupe the PTO. The challenger’s initial submission can educate the examiner at the outset about the key things to watch for. The knowledge that the process has been commenced by such a submission from a sophisticated and motivated opponent should reduce the likelihood that the owner will make substantially incomplete or poorly developed submissions. If the owner submits a one sided or misleading responsive statement the challenger will have an opportunity to attack and balance it through a reply. And the copies of the Office Actions and responses could equip the challenger to detect at least clumsy efforts to defraud the PTO (apparently the only place a challenger could press such a fraud claim would be in federal court, not in the PTO itself).
Because there is less risk of nondisclosure in a reexamination than in a completely ex parte prosecution of an original application it makes less sense in the reexamination setting to permit fear of nondisclosure to completely dominate our decisions about whether to extend work product protection. The objective of encouraging disclosure should remain a substantial consideration, but when the structure of the procedure permits some adversarial inputs, and thus creates pressures that reinforce the owner’s legal duty to be forthcoming with the PTO, courts can afford to take additional considerations into account when resolving this kind of problem.
In denying work product protection to documents generated in connection with ex parte proceedings some courts, as indicated above, have accepted the argument that since a principal rationale for the doctrine is to preserve counsel’s incentives to prepare their cases thoroughly there is no need for the work product rule in proceedings where independent legal duties and economic self interest normally will supply a full measure of the desired incentive. As this opinion makes abundantly clear, I have no doubt that in many reexamination proceedings, certainly in the one in question here, parties and lawyers have ample incentives to prosecute their cases vigorously, independent of anything that work product protection might add. But there are additional values that the work product doctrine was designed to promote, some of which could be advanced by making the doctrine available in appropriate circumstances to limited kinds of documents that are generated in reexamination proceedings.
As the Hickman opinion made clear, one purpose of the work product doctrine is to protect the morale of the legal profession. 329 U.S. at 511, 67 S.Ct. at 393-94. That morale could be badly damaged if adversaries could rummage at will through a lawyer’s files, especially those aspects of his files that reflect the work that is closest to the center of his craft, his legal reasoning, his strategizing, his packaging of materials for presentation to the adjudicator, and his subjective assessments of strengths and weaknesses of the parties’ *546competing positions. It was in part out of its desire to protect that morale that the Hickman court insisted that lawyers have a right of privacy in the documents they create that reflect how they plan to develop or present a case.
The Hickman court also wanted to discourage lawyers from attempting to litigate their cases through the words and work of opposing counsel. Leaching is demoralizing. It also represents a major misfocusing of counsel’s energy. Since most cases turn on the facts, courts should pressure lawyers to spend most of their energy trying to promptly acquire the relevant evidence directly from the original and most reliable sources, sources in the real world, not in an opponent’s files. Lawyers in patent cases, in particular, too often exhibit an unhealthy preoccupation with words attributable to opposing counsel. Courts should discourage lawyers from spending substantial amounts of their client’s money looking for litigation leverage in the verbiage generated by other lawyers. Intellectual property litigation already looks too much like a semantic shell game.
Another consideration made relevant by Hickman is the fact that the more a lawyer’s words are discoverable, the greater the likelihood that he will be forced to become a witness. And the more often lawyers are witnesses the greater the disruption of the administration of justice (e.g., every time a lawyer becomes a witness there is considerable pressure on him to withdraw as counsel of record, and every substitution of counsel creates transaction costs and risks delaying final disposition).
The work product doctrine also promotes lawyer creativity. It fosters an environment where lawyers can feel free to consider, in private, a multitude of different approaches and theories, some that at first blush seem hairbrained, or too risky, or too theoretically far out. But it is only when they feel free to indulge in such private intellectual experiments that lawyers can fully explore their client’s options, and it is only in such settings that lawyers are likely to truly exhaust their creative potential.
Giving limited recognition to the work product doctrine with respect to some kinds of documents that'lawyers generate in connection with reexamination proceedings could promote some of these values. I conclude that the Clark draft of the “Reply” falls in the category that should be entitled to protection, at least absent a quite substantial showing of need by an opponent.
At the outset I emphasize that it seems fair to conclude, formally, that this draft satisfies the requirement of being “prepared in anticipation of litigation or for trial.” At the time attorney Clark prepared this draft his client was defending a lawsuit initiated by the same party that forced the reexamination. Moreover, he knew that the subjects addressed in the draft were pivotal to the litigation. In fact, the draft represents the lawyer’s private thinking about how best to argue the issues on which the litigation is likely to turn. When he prepared this document attorney Clark was in a very real sense preparing for the trial of this case. Given these realities of the situation, it would contribute nothing for the court to attempt to idetermine which proceeding, the reexamination or the trial, most preoccupied attorney Clark when he was preparing this document. The answer to that question simply makes no difference for purposes of deciding whether the values that inform work product doctrine would be sufficiently served to extend protection to this document. The careful reader must by now be wondering what distinguishes, for purposes of work product analysis, the Clark draft of the “Reply” from the drafts of the Holmes declaration. There are several relevant factors. The first is that the Clark draft was clearly and thoroughly the work of a lawyer. This is an important point, because, as noted above, the values that originally inspired the work product doctrine, and that continue to be most important to its rationale, relate to the work of lawyers and to the way lawyers function as *547the motor forces of our adjudicatory system.
A second major factor distinguishing the two kinds of documents focuses on differences in their content and in the kinds of inputs the system expects from the two different kinds of authors. Mr. Holmes presents himself to the PTO and to the court as an intellectually independent technical expert. He purports to present scientifically accurate descriptions of physical reality and to engage in engineering reasoning, that is, reasoning based on what we assume are value neutral principles developed in the physical sciences. We know that he has been retained by a litigant, but we have a right to expect his advocacy to be meaningfully cabined by science. We have this right in large part because to get to the truth, or to apply legal norms reliably in litigation, it is essential that the real world facts on which everything ultimately depends be honestly characterized.
Mr. Clark’s work, however, is not presented as engineering reasoning. While his draft includes commentary on reasoning by Mr. Holmes, it is, in essence, argument toward a legal idea, not scientific description of mechanical facts. Moreover, we recognize that attorney Clark’s role is to try to persuade, to create verbal packages and to shape arguments that present his client’s case (rather than reality itself) in the best possible light. We do not expect Mr. Clark to present a detached and scientifically balanced portrait of physical facts. We expect him to argue. We expect him to slant his presentation in whatever manner (consistent with our perhaps over-generous ethical standards) he calculates to be most likely to achieve the litigation objective of his client. And his role, more clearly than Mr. Holmes’, is to argue what the implications of the physical evidence are for legal norms or social values. His is a role in which creativity is less likely to compromise the basic data, the fundamental predicates, of the truth finding process.
Sophisticated counsel might argue that the line between the nature and purposes of the Holmes declaration and the Clark draft is thin and wavering. Perhaps so. But much of our life in the law consists of problematic line drawing. In fact, one of the judiciary’s principal responsibilities in our society is to draw difficult lines. So the fact that the line is thin is not particularly worrisome as long as there is a basis, articulable in terms of the policies that inspire the relevant legal doctrine, for placing the line were it is. I think there is.
There are additional distinctions between Mr. Holmes and Mr. Clark that are important for present purposes. Mr. Holmes already is a witness. There is no danger that refusing to extend work product protection to his documents will increase the likelihood that he will be called to testify. And no threat is posed to the administration of justice, or to the morale of the legal profession, when engineers are called as witnesses. Moreover, it is altogether fitting and proper for opposing counsel to focus their energies on probing the basis for the opinions Mr. Holmes will offer at trial and that are presented in his declaration. Opinions from experts like Mr. Holmes will determine the outcome of the litigation, and rightly so.
Not so with the arguments and opinions of Mr. Clark, defendant’s attorney. We should discourage opposing counsel from pursuing through discovery what Mr. Clark thinks. What he thinks is not evidence. And we do not want Mr. Clark to be forced to become a witness. He is not the source of the technical expertise or of the raw data about real world events on which the case should turn. Thus it makes considerable policy sense to protect from disclosure his private thinking about how best to package the case for presentation to the PTO and to the court, while simultaneously refusing to extend that protection to the documents Mr. Holmes generated as he developed the expert opinions he will present at trial on the technical subjects that are at the heart of this case.
For all of the reasons above set forth, the Court hereby enters the following ORDERS:
*5481. By July 30,1987, defendant shall produce to counsel for plaintiff all drafts of the Holmes declaration.
2. Attorney Clark’s draft of the “Reply Under C.F.R. sec. 1.111” is protected from disclosure by both the attorney-client privilege and the work product doctrine and thus need not be produced.