dissenting.
I would grant the request of the United States to recall the court’s mandate, and I dissent from the court’s denial of that request.
I would also grant the request of the United States to have this significant case heard by the full court and I would hold that an immediate appeal lies from the district court’s order of deposition. I believe that such is required under both the Classified Information Procedures Act and the collateral order doctrine. I also believe that such is necessary in the interests of national security, as has been represented to us on behalf of the President of the United States.
In my judgment, first the district court, then the panel, and now, regrettably, this court en banc, have all failed to appreciate the fragility of the intelligence gathering process and the susceptibility of that process to influences that might be thought remote and inconsequential in other contexts. Because of this failure, I believe my colleagues have gravely underestimated the effect that their respective orders and decisions have already had, and now will continue to have, on the Nation’s intelligence gathering during this critical period of our history, as we wage war against terrorism and its sponsors around the globe.
I.
The issue that has been brought before the court has profound implications for the Nation’s security. That this is so, however, does not mean that the government is relieved of its obligations under the Constitution and laws of the United States. Quite the opposite. The government’s efforts to ensure security to its citizens, like all other governmental activities, are subject to law (if they are not especially subject to law in this context). However, in this instance, neither the Constitution nor statute requires the risk to our Nation’s security that the court, by its decisions, confidently but no less improperly accepts on the public’s behalf today.
The panel opinion, whose mandate it is asked that we recall and of which en banc reconsideration is sought, is ambiguous. This fact, given the signal importance of the litigation and the corresponding imperative for clarity, itself provides sufficient reason for the full court to rehear this case. But under any of the three arguable readings of the opinion, the court also plainly erred in its conclusion that the district court’s order of deposition is not immediately appealable under the Classified Information Procedures Act, and therefore that the national security must yield, at least for the time being, to other interests.
A.
The panel unfortunately does not articulate either its reasoning or its conclusion sufficiently clearly that the parties, the public, or even the members of this court can understand its decision. The analytical ambiguity that inheres in the panel’s analysis of CIPA in particular is evident in the single passage of analysis offered for the court’s decision. In relevant part, that brief passage states as follows:
Here, the Government contends that the order of the district court directing the deposition of the enemy combatant witness is “a decision or order ... authorizing the disclosure of classified in*287formation,” from which it may take an immediate appeal. We disagree. CIPA § 6, to which the Government points, is concerned with the disclosure of classified information by the defendant to the public at a trial or pretrial proceeding, not the pretrial disclosure of classified information to the defendant or his attorneys. It is true, of course, that the district court issued the testimonial writ based in part on its assessment that the enemy combatant witness’ testimony would likely be helpful to Moussaoui’s defense. But, neither this conclusion, nor the fact that the purpose of the deposition is to preserve the enemy combatant witness’ testimony for potential use at trial, is sufficient to establish the applicability of CIPA. At its core, the order of the district court concerned only the question of whether Moussaoui and standby counsel would be granted, access to the enemy combatant witness (and if so, what form of access), not whether any particular statement of this witness would be admitted at trial. -The district court was thus correct to conclude that CIPA applies here only by analogy. Because CIPA is not directly applicable, § 7 does not authorize an interlocutory appeal.
Slip op. 7-8 (internal citations omitted). This text can be read, and fairly so, in any of three ways. It can be read to hold that the district court order did not “authorize the disclosure of classified information.” Or it can be read to hold that, even though the district court order did authorize disclosure of classified information, section 7 of CIPA does not permit an interlocutory appeal from this order because it was not, by its terms, issued under CIPA or under authority of that Act, and therefore CIPA’s appellate review provisions do not apply to this court’s consideration of the instant appeal. Finally, it can be read to hold that the district court order authorized the disclosure of classified information, but not the particular type of disclosure with which CIPA is concerned.
On any of these understandings of the panel’s opinion, the court erred in its essential conclusion that an immediate appeal of the district court’s order does not lie under CIPA.
B.
Under the first reading of the panel’s opinion, the court would have held that the order from which the government appeals does not “authorize the disclosure of classified information,” within the meaning of section 7(a) of CIPA. Were this the holding of the court, it would be in error because the district court’s order most certainly does authorize the disclosure of classified information. As the government explains, Moussaoui does not have the necessary clearance to receive classified information from the government. Petition for Panel Rehearing or Rehearing En Banc at 3. Therefore, absent the district court’s order, the government officials in possession of the classified information Moussaoui seeks would themselves have no authority to produce such classified information to Moussaoui; indeed, the unauthorized disclosure of classified information to an individual, like Moussaoui, who is not cleared to receive such information, would likely constitute a felony under Title 18, section 798, of the United States Code.
Thus, the district court’s order unquestionably “authorizes the disclosure of classified information” in the government’s possession to Moussaoui, an authorization that brings the district court’s order squarely within the ambit of the plain language of section 7(a).
Under the second possible reading of the panel’s opinion, the court would have *288held that section 7 only permits interlocutory appeal when CIPA is “directly applicable” to the order appealed by the government. See Slip op. 8. The panel does not explain its reference to a precondition of direct applicability. Presumably, however, that reference is to a belief that, in order for section 7 to authorize interlocutory appeal, either the district court’s order must have been explicitly entered pursuant to, or at least have been authorized by, a provision of CIPA other than section 7(a). But even if the panel’s opinion were read to so hold, the court would yet be in error, because such a holding cannot be squared with CIPA’s plain language.
Section 7(a) of CIPA provides simply that appellate jurisdiction will lie over “a decision or order of a district court in a criminal case authorizing disclosure of classified information.” Section 7(a)’s authorization of immediate appeal is not limited to orders that, by their terms, are entered pursuant to CIPA or even to orders that are authorized by CIPA. (As the government notes, even if such a limitation existed, it would be satisfied here, because the district court’s order is a “determination[ ] concerning the use ... of classified information ... [at] ... [a] pretrial proceeding” within the meaning of section 6.). Nor does any other provision of CIPA even arguably impose such a limitation. And had Congress wished to circumscribe section 7(a) so as to provide appellate jurisdiction only over such orders, it could easily have done so. In fact, section 6(c)(1) of CIPA employs precisely this sort of limitation, providing that the United States may move to substitute admissions or a redacted summary in lieu of classified information “[u]pon any determination by the court authorizing the disclosure of specific classified information under the procedures established by this section.” Id. (emphasis added).
Finally, under the last of the three possible readings of the panel’s opinion, the court would have held that section 7(a) of CIPA is unconcerned with the disclosure of classified information by the government to the defendant, and, instead, protects only against the disclosure of such information by the defendant to the “public” (which does not include the defendant and his attorneys) in a pretrial hearing or at trial. That this was the holding most likely intended by the panel is suggested by the first sentence of its CIPA discussion, which asserts that CIPA was enacted “to combat the problem of ‘graymail,’ an attempt by a defendant to derail a criminal trial by threatening to disclose classified information.” Slip op. 7. And, indeed, that this is in fact the court’s holding is all but confirmed by the statements and conclusions that are featured in the panel’s short analysis. Those statements and conclusions are that “CIPA § 6, to which the Government points, is concerned with the disclosure of classified information by the defendant to the public at a trial or pretrial proceeding, not the pretrial disclosure of classified information to the defendant or his attorneys” and that “the order of the district court concerned only the question of whether Moussaoui and standby counsel would be granted access to the enemy combatant witness (and if so, what form of access), not whether any particular statement of this witness would be admitted at trial.” Id.
It is unsurprising, therefore, that the United States understands the panel’s opinion to hold that an immediate appeal of the district court’s order does not lie because it authorizes disclosure only to Moussaoui and his attorneys, and not to the public at large. See Petition for Rehearing or Rehearing En Banc at 4 (stating that the panel opinion “held that the particular disclosure of classified information authorized by the order does not trig*289ger Section 7, because it involves only disclosure to the defendant, not disclosure to the public”). Indeed, this is what I understand the panel to have held as well.
I agree with the government that the panel held that immediate appeal does not he under section 7(a) of CIPA where at issue is only an authorized disclosure to the defendant. However, on the assumption that this is the court’s holding, it, too, is in error.
To be sure, CIPA is concerned with a defendant’s disclosure to the public of classified information that is already in the defendant’s possession. But this is not the only concern of the statute, and to fail to appreciate that there are also other concerns embodied in the Act is to misunderstand the Act. CIPA is equally concerned with the foreseeable (and foreseen) event that defendants will attempt to delay, if not altogether derail their prosecutions by attempts to obtain access to classified information that they do not already possess. For example, sections 2, 4, and 6 of the Act are all concerned with the production of classified information by the government to the defendant. This additional purpose of the statute has nowhere been more forcefully recognized than in the Ninth Circuit’s opinion in United States v. Clegg, 740 F.2d 16 (9th Cir.1984), where the court found it “clear that CIPA is as concerned with controlling disclosures to the defendant as it is with controlling disclosures to the public.” Id. at 18.
The conclusion that CIPA extends to the circumstance, such as that here, where the defendant seeks access to classified information, is unavoidable based upon the plain language of CIPA. But, in fitting coincidence, even the Senate Report relied upon by the panel for its mistakenly narrower interpretation of the Act confirms as much. That report clearly states, that the problems CIPA was meant to address are not “limited to instances of unscrupulous or questionable conduct by defendants since wholly proper defense attempts to obtain or disclose classified information may present the government with the same ‘disclose or dismiss’ dilemma.” S.Rep. No. 96-823, at 3 (1980), reprinted in, 1980 IJ.S.Code Cong. & Admin.. News 4294, 4296-97 (emphasis added).
Accordingly, I believe that, under section 7(a) of CIPA, an immediate appeal of the district court’s order of deposition lies to this court.
C.
That an immediate appeal of the district court’s order does lie under CIPA, and thus that the panel erred in its contrary conclusion, is only further supported by the fact that our sister circuits have rejected the very analysis the panel adopted — in opinions that are not even cited, much less discussed, by the panel. Refusing to fall into the error committed by our court, the Ninth Circuit held unequivocally in United States v. Clegg, in an opinion joined by then-Judge Kennedy, that section 7(a) of GIPA provides for appellate jurisdiction not merely over court orders authorizing disclosure by the defendant to the public, but also over orders authorizing disclosure by the government to the defendant. See Clegg, 740 F.2d at 18; see also United States v. The LaRouche Campaign, 695 F.Supp. 1282, 1288 (D.Mass.1988) (noting that one “manifest objective of CIPA is that classified information should not be disclosed to anyone needlessly” and that “when classified information is not yet in the hands of defendants and their attorneys and they are making demands for disclosure, the court must consider whether defendant’s rights can be fully protected by an alternative procedure that does not result in the disclosure of classified information.”) (emphasis added); The District *290of Columbia Circuit, in United States v. Yunis, 867 F.2d 617 (D.C.Cir.1989), although it was not required to decide the question, even assumed that it had jurisdiction over such an order under section 7(a) of CIPA.
Thus, in its indefensible reading of CIPA, our panel stands not only alone, but in direct conflict with our sister circuits that have considered the question.
II.
While I believe that the district court’s order is immediately appealable under CIPA, I believe the panel would be mistaken in its conclusion that the district court’s order is not immediately appealable even were such an appeal not authorized by that Act, because the district court’s order is independently appealable under 28 U.S.C. § 1291, as a collateral order. See United States v. Pappas, 94 F.3d 795, 798 (2d Cir.1996) (holding that “CIPA does not deny appellate jurisdiction that otherwise exists. It simply extends to the Government, but not to the defendant, the opportunity for an interlocutory appeal of protective orders in circumstances where an appeal would not otherwise exist.”). Although section 1291 provides that appeal may be taken only from a “final order,” the panel correctly recognizes that “[t]he Supreme Court has long given the finality requirement in § 1291 a practical construction rather than a technical one.” Slip op. 8 (citing Cohen v. Beneficial Indus. Loan Corp., 337 U.S. 541, 546, 69 S.Ct. 1221, 93 L.Ed. 1528 (1949)). Yet the panel proceeds to apply the finality requirement technically rather than practically, by insisting that the district court’s order is a “discovery order like any other” and that it must be “treated the same for jurisdictional purposes.” Slip op. 10.
If properly applied to take into account the exceptional nature of the order at issue, I have no question but that the district court’s order is immediately appeal-able under the collateral order doctrine, as well as under section 7(a) of CIPA.
The panel’s- rejection of appellate jurisdiction under the collateral order doctrine rests specifically on its holding that the district court order “fails to satisfy the first prong of the Cohen analysis,” slip op. 9, which requires that “an order conclusively determine the disputed question.” Coopers & Lybrand v. Livesay, 437 U.S. 463, 468, 98 S.Ct. 2454, 57 L.Ed.2d 351 (1978). This holding is obviously incorrect. The district court order conclusively resolves the question of the defendant’s right to depose the enemy combatant witness, as that court’s subsequent actions confirm. The district court affirmed its original order of deposition on March 10, 2003, with a written opinion. It later denied the government’s motion to modify the order in a May 15, 2003, ruling, rejecting an affidavit of the government certifying that the disclosure of the information would cause “identifiable damage to the national security.” See Petition for Panel Rehearing and Rehearing En Banc at 3 n. 1. And, on July 7, 2003, the court ordered that the United States appear today to advise the court whether it intends to comply with the court’s order of deposition. Clearly, the district court’s order was not, as the panel erroneously claims, “tentative, informal or incomplete.” Cohen, 337 U.S. at 546, 69 S.Ct. 1221.
The panel reasoned that the order is “incomplete” until the government openly defies it and incurs sanctions. Slip op. 9. But this cannot be so. An order imposing sanctions would resolve an entirely separate question from the one resolved by the order of deposition. The question of whether Moussaoui has a right to depose the witness is entirely distinct from the question of what sanction will be imposed *291if the government defies the court order recognizing that right.
By effectively holding that a district court discovery order can never satisfy the first prong of Cohen unless sanctions are imposed for disobedience of that order, the panel again disregarded contrary authority from our sister circuits. In In re Ford Motor Co., 110 F.3d 954, 958 (3d Cir.1997), the Third Circuit held that a discovery order requiring the production of disputed documents did satisfy the first prong of Cohen because it left “no room for further consideration by the district court of the claim that the documents are protected.” Likewise, in United States v. Philip Morris, Inc., 314 F.3d 612, 617 (D.C.Cir.2003), the District of Columbia Circuit held that a district court order that a particular memorandum was not protected by the attorney-client privilege satisfied the first prong of Cohen because it conclusively and finally determined a distinctly separate issue from the merits of the underlying dispute. In fact, in holding that the order in question was appealable under the collateral order doctrine, the court specifically rejected the contention that a discovery order must be defied and sanctions must be imposed in order for that order to be appealable.
Although the panel opinion does not even discuss the second and third prongs of Cohen, it is evident that the district court’s order satisfies both of these requirements. As to the second prong, the appeal indisputably presents a matter of great importance — implicating as it does the Nation’s security — which is distinct from the merits of the underlying criminal case. The district court’s order also satisfies Cohen’s third prong, requiring that the order be effectively unreviewable on appeal from final judgment. For, were the government to comply with the district court’s order by producing an enemy combatant witness to be deposed by Mous-saoui and his lawyers, the harm to the national security would be irreparable. And, if the government chooses to defy the district court’s order and accept the likely sanction, the harm inflicted could be significant even if the sanction were ultimately reversed by this court on subsequent appeal. See discussion infra.
At bottom, the panel’s insistence that the United States first be sanctioned before appealing the district court’s order of deposition is premised on its fundamentally flawed belief that that order is “a discovery order like any other,” which “must be treated the same for jurisdictional purposes.” Slip op. 10. The district court’s order simply is not a discovery order like any other. It is, rather, a unique order, perhaps even unprecedented in our jurisprudence.
When such an extraordinary order with such far-reaching effects is appealed, it is plain error to treat it the same as a run-of-the-mill discovery order. As the Supreme Court instructed in United States v. Nixon, 418 U.S. 683, 691, 94 S.Ct. 3090, 41 L.Ed.2d 1039 (1974)—an opinion which the panel nonchalantly distinguishes with a citation to Bennett v. City of Boston, 54 F.3d 18 (1st Cir.1995), a case involving an interlocutory appeal by a district attorney in a routine criminal prosecution, see slip op. 10 n. 8 — “[t]he requirement of submitting to contempt [ ] is not without exception and in some instances the purposes underlying the finality rule require a different result.”
The instant case comfortably falls within the category of exceptional cases envisioned by the Court in Nixon as warranting interlocutory appeal.
Many of the factors present in the “unique setting” of Nixon are also present in this case. As in Nixon, the requirement that the United States place itself “in the posture of disobeying an order of a court *292merely to trigger the procedural mechanism for review of the ruling would be unseemly and would present an unnecessary occasion for constitutional confrontation between two branches of Government.” Id. at 691-92, 94 S.Ct. 3090. If anything, this consideration weighs even more heavily in this case than it did in Nixon because the district court’s order impedes the President’s exercise of power as Commander-in-Chief by limiting his authority to prevent access by and to enemy combatants. And, as in Nixon, the requirement that the United States first be sanctioned by the district court before appealing its underlying claim to this court only “would further delay both review on the merits of [its] claim ... and the ultimate termination of the underlying criminal action for which [its] evidence is sought.”
In sum, the required balance under the collateral order doctrine manifestly ought be struck in favor of immediate appealability in this case, and I would so hold.
III.
I have no doubt whatever that the several orders and decisions issued by the district court and the panel have already not merely caused alteration of, but actually disrupted, the questioning of the particular enemy combatant witness whose production and deposition have been ordered by the district court. The timetables carefully laid out for, and the techniques designed specifically to be employed in, the questioning of this witness have inevitably had to be adjusted both in anticipation of the predicted rulings that would issue from this court and the district court and in response to the actual rulings that have come forth from these courts. Just as certainly, in detriment to the strategic interests of the Country, the government has also been required to artificially structure its questioning of this enemy combatant witness around the real possibility that any information imparted to him in order to obtain information in his possession may ultimately have to be disclosed to the defendant and his counsel during the course of judicially-ordered deposition, if not to the public at large at any ensuing trial. And common sense should tell that our orders and decisions have, as well, exacted untold hidden costs with respect to other existing and potential informants who have, respectively, either withheld information that they would otherwise have already provided or decided not to come forward at all until this litigation is finally resolved and its implications for their considered disclosures fully understood.
Of course, none of this is even to mention the obvious fact that, in their efforts to protect the Nation, the President and his national security advisors, if only incrementally so, have had to proceed differently than they might otherwise have proceeded were this particular matter not hanging over them like the sword of Damocles, as they have awaited final decision from us.
The panel, and now the full court, have comforted themselves in the fact that we will eventually have jurisdiction to review any order of the district court sanctioning the United States for noncompliance with its extraordinary order of deposition. This is, there should be no doubt, to ignore (or to accept without full appreciation of the possible costs to national security) the distinct possibility that, because of the courts’ actions, appellate review of any sanction imposed will never be sought, with attendant consequences for the President’s diplomatic and military conduct of the war against terrorism left to be borne by an unsuspecting public.
In any event, in taking comfort in the belief that appellate review will ultimately *293be had, the court only fails again to appreciate the profound effect that any decision in a litigation of this sensitivity inescapably has upon the delicate psychological balance that can determine victory or defeat as much as can combat itself. For even the temporary imposition of sanction in a case of this delicacy and significance to the war against terrorism can, at the very moment when psychological advantage could prove determinative of the conflict or its direction, embolden the enemy in that war and weaken the resolve of those who are charged with prosecution of that war and with protection of our homeland. One can only imagine the encouragement that even a short-lived order dismissing the charges against the appellee in this case would provide the terrorist network and its allies around the world, and the demoralizing setback that such would in turn represent for the men and women who have been charged to track down, capture, and bring to justice those who have brought fear to the international stage. Such an order would resonate throughout the world, and, its belief to the contrary notwithstanding, this court simply could not calibrate its actions finely enough or react swiftly enough to prevent such a sanction from having unanticipated deleterious effect on the Nation’s security.
IV.
The panel opinion of which further review has been sought rested squarely, as the court openly stated, on a conclusion that the “order of the district court is a discovery order like any other.” The district court’s order, however, is not like any discovery order, and, as the government suggests, it may in fact be -unlike any other discovery order. The district court’s order of deposition, whether right or wrong under law, was extra-ordinary in every sense, affecting, as the government compellingly explains, “an ongoing military operation by demanding the disruption of efforts to obtain intelligence and providing an admitted terrorist with access to an enemy combatant detained overseas in the midst of war,” see Petition for Panel Rehearing or Rehearing En Banc at 14.
Because I believe that, under law, the United States is entitled to immediate review of this extraordinary order entered by the district court, I would grant the motion of the United States to recall the mandate issued by the panel, grant the petition for rehearing en banc, and hold that interlocutory appeal lies in this court to hear, on the merits, the government’s appeal of the district court’s order of deposition. To proceed differently, as the court does today, is to play Russian roulette with the security of the Nation.
For these reasons, I dissent.