concurring in part, dissenting in part.
I write to express two areas of concern.
I
I agree with the court’s decision that Biomet’s confidential sales information is not subject to discovery in the circumstances of this case. However, I observe that this court reached a different conclusion on similar facts in Truswal Systems Corp. v. Hydro-Air Engineering, Inc., 813 F.2d 1207, 2 U.S.P.Q.2d 1034 (Fed.Cir.1987). I had hoped that the majority might seize this opportunity to shed light on the boundary between these decisions. Although the majority has attempted to point out some differences, I discern scant guidance on which one might rely.
In Truswal this court reversed a district court’s denial of discovery of confidential sales information of a non-party competitor, stating that the discovery could be had under a protective order although an acceptable order had not yet been fashioned. We now uphold a district court’s denial of discovery of confidential sales information of a non-party competitor, despite the existence of a protective order.
In both cases the discovery was sought for the same stated reasons: to bolster the patentee’s case of validity in different litigation by showing commercial success of the patented invention in the hands of the non-party. In both cases the non-party had been accused of infringement but not yet sued. In both cases discovery of detailed sales figures was resisted by the non-party on the grounds that it was a competitor, that the detailed sales figures were highly confidential and sensitive, and that the detailed sales figures had not been shown to be necessary to the patentee’s case. In both cases the district court granted or the non-party voluntarily provided extensive other discovery.
The majority herein, consisting of the same judges as the majority in Truswal, reports that it reversed in Truswal because that district court denied discovery “on the sole ground that discovery was sought from a nonparty” (emphasis by majority). However, the Truswal district court referred to three aspects of its denial of discovery: (1) that the cited case authority did not support production of (2) confidential sales information by (3) non-party witnesses. Truswal, 813 F.2d at 1208, 2 U.S.P.Q.2d at 1035.
In both Truswal and the case at bar the district courts exercised their discretion to reach the identical result. Yet in Truswal the majority held that the district court abused its discretion, while in this case the panel unanimously upholds the district court.
In matters of discovery the district court has broad discretionary authority, and discovery rulings should not easily be overturned. My colleagues on this panel surely have in mind some controlling distinction between the two cases. I should like to partake of that knowledge, at least to learn whether the opposing results were reached because of differences in facts, or in procedure, or perhaps in quality of advocacy.
II
I am greatly concerned by the majority’s treatment of attorney-client privilege.
As the majority reports, Biomet obtained an opinion of counsel as to the validity of the ’123 patent. Biomet’s counsel gave a copy of the opinion in confidence to De-Puy’s counsel. DePuy in turn produced it to American Standard in response to a discovery request, along with two validity opinions that DePuy had obtained from its counsel. DePuy asserted no attorney-client privilege as to any of these opinions.
Thus American Standard had a copy of Biomet’s counsel’s opinion that the ’123 patent is invalid, and sought to depose and have discovery of Biomet’s counsel and of a Biomet employee on matters related to the ’123 patent.
The district court held that although “Biomet sought from outside counsel, an analysis of the validity of the ’123 patent”, “there is no privilege in the document” because it “relies on non-confidential information gleaned from public records or doc*748uments”. The court therefore held that the disclosure did not waive any attorney-client privilege because there was none to waive, and on this basis refused to compel the requested discovery.
Biomet did not appeal this ruling, which achieved Biomet’s purpose of avoiding the requested discovery. Rather, American Standard is in the curious position of arguing on appeal that Biomet’s legal opinion was indeed privileged, that the privilege was waived, and thus that the requested discovery should be compelled.
Of course, only the holder of the privilege can assert it. However, the issue was not decided on that basis. Instead, the majority herein agrees with the district court that the patent validity opinion was never privileged. This ruling negates decades of hard won precedent, and is a giant step backward into uncertainty, confusion, and prejudice.
A
The Biomet validity opinion, like most patent validity opinions, describes the attorney’s review of “prior art” published patents and scientific articles. All patent validity opinions applying the laws of anticipation and obviousness of 35 U.S.C. §§ 102 and 103 evaluate validity in view of “prior art” published knowledge. Published knowledge is, by definition, not confidential. It is of public record. Yet the majority holds that the Biomet opinion is not privileged because it relies on prior art that is found in public records and is not confidential.
The majority explains:
The district court said the opinion letter was not a privileged communication because it relied on nonconfidential information and stated the source of that information. Contrary to American Standard’s assertion, it did not say the opinion letter was not privileged merely because it relied on publicly available information. It clearly said the letter relied on nonconfidential information gleaned from public records. American Standard simply ignores the finding of nonconfidentiality and focuses alone on its source. Indeed, a mere reading of the above-described opinion “letter” demonstrates its nonconfidential nature, [emphasis in original]
Although the majority protests that it is not deciding for all fact situations, in the typical fact situation before us the majority negates the privilege of patent validity opinions based on prior art.
The prevailing view, until today, was that patent validity opinions based on prior art were legal opinions, subject to the attorney-client privilege as any legal opinion. That privilege could be claimed or waived by the client, as for any legal opinion. But if not waived, voluntarily or involuntarily, the privilege could be asserted.
The majority denies any intent to revive that aspect of United States v. United Shoe Machinery Corp., 89 F.Supp. 357, 359, 85 U.S.P.Q. 5, 6 (D.Mass.1950), which held that “there is no privilege for so much of a lawyer’s letter, report or opinion as relates to facts as gleaned from ... a public document such as a patent”. That and various other aspects of United Shoe relating to patents have long been distinguished or overruled. See, e.g., Sperry v. Florida, 373 U.S. 379, 383, 83 S.Ct. 1322, 1325, 10 L.Ed.2d 428 (1963) (advising clients as to the patentability of inventions under statutory criteria constitutes the practice of law); Ledex, Inc. v. United States, 172 U.S.P.Q. 538, 539-40 (Ct.Cl.1972) (noting that trend away from United Shoe began with Sperry, and holding that documents relating to attorney-client communications while application was pending before the Patent Office were privileged); Knogo Corp. v. United States, 213 U.S.P.Q. 936, 942 (Ct.Cl.Tr.Div.1980) (“There is no longer any question” that opinions of the attorney on patentability and scope of patent claims “are legal opinions.”); Hercules Inc. v. Exxon Corp., 434 F.Supp. 136, 143, 196 U.S.P.Q. 401, 405 (D.Del.1977) (refusing to “resuscitate the question of the applicability of attorney-client and work product protection to patent matters”). The majority’s affirmation of the district court’s holding that there is no privilege in a lawyer’s opinion that “relies on non-confidential in*749formation gleaned from public ... documents”, wording reminiscent of United Shoe, is blatantly contrary to this precedent.
The weight of modern authority is reflected in In re Ampicillin Antitrust Litigation, 81 F.R.D. 377, 390, 202 U.S.P.Q. 134, 143 (D.D.C.1978), where the court wrote:
If an attorney-client communication could be discovered if it contained information known to others, then it would be the rare communication that would be protected and, in turn, it would be the rare client who would freely communicate to an attorney.
In Nestle Co. v. A. Cherney and Sons, Inc., 207 U.S.P.Q. 930, 933 (D.Md.1980), the court considered the privilege of “information obtained through searches of public records, such as materials on file at the Patent and Trademark Office” stating:
[The best approach of prior cases] imply that a document does not fall outside the attorney-client privilege merely because it contains technical or publicly-obtained information. If the party invoking the privilege can show that the document has some legal significance, then the document may be immune from discovery, [citation omitted] More specifically, the communication of the publicly obtained information “should be privileged to the extent that the communication was treated as confidential by the client and would tend to reveal a confidential communication of the client.” In re Ampicillin Antitrust Litigation, 81 F.R.D. at 389, 202 USPQ at 142. [emphasis in original]
In EZ Loader Boat Trailers, Inc. v. Shoreline Trailer Sales, Inc., 207 U.S.P.Q. 1002, 1003 (N.D.Tex.1979), the court stated:
The opinion of a patent attorney deduced from a search of trademark registrations and patents is indistinguishable from legal advice sought in any other confidential relationship between an attorney and his client. This opinion is a communication to the client made in confidence and relating to the matter on which legal advice is sought.
The majority, attempting to explain its aberration from the prevailing view, states that “a mere reading of the opinion ‘letter’ ... demonstrates its nonconfidential nature”. The challenged “letter” is headed on the first page “OPINION”. It first identifies the prior art on which the attorney bases his conclusion of invalidity, as follows:
The prior art to be considered is: (1) Davila, “The Development of Artificial Heart Valves”, Special Technical Publication No. 386, American Society for Testing and Materials (1965); (2) Wright U.S. Patent 448,745 (1891); and (3) Smith, “Ceramic-Plastic Material as a Bone Substitute”, Archives of Surgery, 87, 653-661 (1963) and U.S. Patent No. 3,314,420.
Under the heading “Conclusion” the attorney states:
Hahn U.S. Patent 3,605,123 is invalid under 35 U.S.C. § 103 as obvious from Davila, alone or in combination with Wright and Smith.
In the “Analysis” section the attorney discusses various published documents, states his view as to what was known at the time the invention in the Hahn (’123) patent was made, and explains the basis for his conclusion of invalidity. An extensive extract illustrates the content:
It was well known prior to Hahn that making an implant of a porous material, or adding a porous surface to the implant, would foster the growth of bone and other tissues into the pores.
Spongy implants are disclosed in a number of publications. Bone ingrowth into polyvinyl sponge is described by Struthers et al, Plastic and Reconstructive Surgery, 15, 274 (1955); in a bone substitute, by Gilmer et al, Surgery, Gynecology, and Obstetrics, 113, 143 (1961); and in Teflon sponge, by Friedenberg et al, Surgery, Gynecology, and Obstetrics, May 1963, 588-592.
Jardon U.S. Patent 2,688,139 (September 7, 1954) discloses an artificial eye which is porous throughout to foster tissue growth into the pores of the prosthesis.
*750Smith, “Ceramic-Plastic Material as a Bone Substitute,” Archives of Surgery, 87, 653-661 (1963), and U.S. Patent 3,314,420 (April 18, 1967), filled a porous ceramic prosthesis with an epoxy resin, then washed the prosthesis with methylene chloride to leach away the superficial resin to a depth of 50 to 70 mils, leaving a porous surface into which fibrous tissue and bone were said to grow.
******
Thus, Smith appreciated the problem of mechanical weakness in an all-ceramic implant, a problem of major importance in a load-bearing implant such as a hip prosthesis.
The need for strength led naturally and inevitably to metallic-core prostheses of various types. Hirschhorn and Reynolds, in a paper presented orally in October 1968 at a symposium of the Metallurgical Section of the AIME, substituted powdered Yitallium, a cobalt alloy used by Hahn, for the ceramic of Smith, and acknowledged the weakness problem associated with the use of porous materials. Hulbert et al, in a paper presented at the same symposium, described a solid metal implant coated with a porous ceramic material to facilitate tissue in-growth. While it is deemed unlikely that either of these papers is early enough to serve as a reference against Hahn, they show the direction of the art as suggested by the following two references, which are available.
As early as 1891, Wright disclosed, in U.S. Patent 448,745, an artificial tooth made of solid gold or platinum with a surface coating of porous, unglazed porcelain. Wright explicitly taught that the metallic core was used to give strength to the prosthesis, and the porous surface to allow____
******
Thus, by the time of Hahn, the concept of a metal-core prosthesis with a spray-coated outer layer of the same metal to allow tissue ingrowth was old and well known. The Hahn device was therefore obvious and unpatentable under 35 U.S.C. § 103.
One question remains: whether Wright and Davila represent analogous art, properly applicable to Hahn. Two publications and a symposium contemporaneous with Hahn clearly show that they are:
Klawitter (“A Basic Investigation of Bone Growth into a Porous Ceramic Material,” Ph.D. Thesis, Clemson University, 1970)....
Nevertheless, Klawitter is evidence that it was common practice at the time of Hahn to adopt materials developed for other purposes, and an obvious place to look would be____[emphasis in original]
The opinion so continues, for a total of six pages.
This is a typical, classical, legal opinion of patent invalidity applying the law of 35 U.S.C. § 103. The opinion states the conclusion of law and the reasoning on which it was based. Its substance is attorney analysis and evaluation of published prior art, the standard source of section 103 invalidity. It is a standard communication of legal advice, from attorney to client. The district court’s and the majority’s view that it is not subject to the attorney-client privilege contravenes the heavy weight of authority, and belies the content of the opinion.
B
This legal opinion thus is subject to a claim of attorney-client privilege unless the privilege has been waived, voluntarily or involuntarily.
American Standard argues that the privilege was waived when Biomet gave a copy to DePuy. The question of waiver was not decided by the district court, although the court raised the issue, without resolution, during argument before it. Biomet had also argued before the district court that waiver was avoided by the “common defense” rule 1 .This issue was not decided *751by the district court and is not pressed before us. Nor does Biomet pursue the argument that even if the opinion were subject to a waived attorney-client privilege, its attorney had not waived his work product immunity.2
Thus although the district court’s denial of the requested discovery of Biomet’s counsel may or may not have been the correct result, it can not have been correct on the reasoning upheld by the majority. The issue of waiver requires review, in the first instance by the district court, on the proper standard.
. In Stix Products, Inc. v. United Merchants & Manufacturers, Inc., 47 F.R.D. 334, 338, 162 U.S. *751P.Q. 508, 511 (S.D.N.Y.1969), wherein a non-party shared a validity opinion in confidence with a party to the suit, the Stix court held that because there was a community of interest in demonstrating the invalidity of the patent, the sharing of the opinion did not constitute waiver of the work product immunity of the attorney who prepared the opinion.
. See United States v. American Telephone and Telegraph Co., 642 F.2d 1285, 1299 (D.C.Cir.1980) in which the court states:
[W]hile the mere showing of a voluntary disclosure to a third person will generally suffice to show waiver of the attorney-client privilege, it should not suffice in itself for waiver of the work product privilege.
See also Hickman v. Taylor, 329 U.S. 495, 510-514, 67 S.Ct. 385, 393-395, 91 L.Ed. 451 (1947).