concurring in part and dissenting in part:
I agree with the majority’s analysis of the issues presented in this case, as applied to this case only.1 The government *1016failed to follow either the procedures of United States v. Tamura, 694 F.2d 591 (9th Cir.1982), or those outlined in the approved warrant, which would have required the government to conduct its search of the seized intermingled computer files by using computer technicians, not case investigators.
Further, I would affirm the Cooper, Mahan, and Illston orders because the seized names at issue were not in “plain view” when seized. The plain view doctrine requires evidence of illegality to be “immediately apparent” to the searching investigator. Horton v. California, 496 U.S. 128, 139-140, 110 S.Ct. 2301, 110 L.Ed.2d 112 (1990). Here, the portion of the spreadsheet (located within the “Tracey” directory) which contained the drug testing results, contained both the names of the ten ballplayers who were the subjects of the warrant and the names of many other ballplayers for whom the government did not have probable cause to search and seize. The spreadsheet did not, however, initially display on the agent’s computer screen the results of steroid testing as to the ballplayers. To see the spreadsheet column containing the results, the agent had to scroll right on the spreadsheet, on to another screen. However, once he scrolled to the right, the agent could see not only the testing results for the targeted ten, but also the results for all of the other ballplayers whose results were listed on the spreadsheet.
A valid “plain view” seizure of items that are truly “immediately apparent” would have required the agent to display only the testing results' for the ballplayers for whom he had a warrant, and seize only evidence of additional illegality if such evidence is “immediately apparent” as part of the segregated results for those ballplayers. For instance, the agent could have selected the spreadsheet rows for the ten ballplayers for which he had a warrant, then copied and pasted those rows into a blank spreadsheet.2 If he had done so, he would have seen only those drug testing results for which he had a warrant.
However, as the government conceded at oral argument before Judge Mahan, rather than limiting the scope of his search in any way, or seizing only that evidence of illegality “immediately apparent” on the *1017first screen he called up, Agent Novitsky intentionally and volitionally scrolled right on the spreadsheet, without first having segregated only the responsive files, “to see if there was anything above and beyond that which was authorized for seizure in the initial warrant.” This demonstrates the seized evidence of illegality was not “immediately apparent” nor in “plain view.”
Thus, I agree with the majority and vote to affirm the district courts’ orders. I write separately, however, because I cannot concur in the proposed guidelines established by the majority opinion, for three reasons.
I
First, the proposed guidelines would require police to forswear any use of evidence found in “plain view” if searching computers with intermingled files. Such a rule departs from existing Supreme Court precedent regarding the “plain view” exception to the Fourth Amendment’s warrant requirement, and do so without a single citation to the Supreme Court’s extensive precedent on the subject or an explanation why that precedent no longer applies. In its seminal discussion of the “plain view” exception in Coolidge v. New Hampshire, 403 U.S. 443, 464-466, 91 S.Ct. 2022, 29 L.Ed.2d 564 (1971), overruled in part on other grounds, Horton v. California, 496 U.S. 128, 110 S.Ct. 2301, 110 L.Ed.2d 112 (1990), a plurality of the Supreme Court explained that:
It is well established that under certain circumstances the police may seize evidence in plain view without a warrant. ...
An example of the applicability of the ‘plain view1 doctrine is the situation in which the police have a warrant to search a given area for specified objects, and in the course of the search come across some other article of incriminating character....
What the ‘plain view’ cases have in common is that the police officer in each of them had a prior justification for an intrusion.... The doctrine serves to supplement the prior justification— whether it be a warrant for another object, hot pursuit, search incident to lawful arrest, or some other legitimate reason for being present unconnected with a search directed against the accused — and permits the warrantless seizure. Of course, the extension of the original justification is legitimate only where it is immediately apparent to the police that they have evidence before them; the ‘plain view1 doctrine may not be used to extend a general exploratory search from one object to another until something incriminating at last emerges,
(emphasis added); see also Horton, 496 U.S. at 137, 139-140, 110 S.Ct. 2301 (affirming that Coolidge, is binding precedent).
It may be, as the majority asserts, that it is more difficult — perhaps significantly so — in the electronic search context to ensure a warrant-based search of seized intermingled files is “as limited as possible” before the police may employ the “plain view” exception to seize materials not described in the warrant. Coolidge, 403 U.S. at 467, 91 S.Ct. 2022. But that gives us no legal ground to eliminate in its entirety a constitutional rule established by the Supreme Court, when to be applied to computer searches.
Rather, because the Court has not yet drawn a distinction between “plain view” seizures in a motel room or automobile, on the one hand, and a computer on the other, we must follow the “plain view” exception, that is until the Supreme Court decides otherwise; our task is to ensure the search is as limited as possible. Given the nu*1018merous risks to privacy the majority identifies, we ought to require a magistrate to give exacting scrutiny to the scope of the search, so to ensure the search is as narrowly tailored as possible to the goal of seizing evidence specifically described in a warrant. The magistrate must ensure that a search for particularized evidence based on probable cause and authorized in a warrant does not devolve into the kind of general search the police conducted here. To this end, the ideas behind some of the majority’s guidelines are sensible: specialized personnel, and not investigators, ought to conduct electronic searches, and the search protocol should be designed to uncover only the information for which the government has probable cause. In United States v. Tamura, 694 F.2d 591 (9th Cir.1982), we held the magistrate should oversee the search process; perhaps the instant case counsels that such oversight ought to be quite close. But identifying these important Fourth Amendment goals does not permit us to disregard the Supreme Court’s jurisprudence on “plain view” as to computer searches simply because we prefer policies that conflict with the rules the Court has established.
II
Second, the establishment of guidelines (which are little more than dicta but are nonetheless binding precedent in this circuit, see Brand X Internet Servs. v. FCC, 345 F.3d 1120, 1130 (9th Cir.2003)) in the manner chosen by the majority goes against the grain of the common law method of reasoned decisionmaking, by which rules evolve from cases over time. This is particularly troublesome in a rapidly developing area of law such as this one, as computer search capabilities improve exponentially by the month.
I recognize that although we lack the competitive advantages of Congress to set out specific, bright-line rules through deliberation and testimony from interested parties, it is our obligation, and our obligation alone, to determine what constitutes unreasonable searches and seizures. See Marbury v. Madison, 1 Cranch 137, 5 U.S. 137, 177-78, 2 L.Ed. 60 (1803) (“It is emphatically the province and duty of the judicial department to say what the law is.”). That said, precisely because we lack those advantages of Congress, we must treat our establishment of brightline rules with great care and deliberation; at the very least, amici should have an opportunity to weigh in on the dramatic doctrinal shift the majority’s guidelines contemplate. Here, there has been no briefing whatsoever on whether the plain view exception should no longer apply to computer searches.
By focusing on the “plain view” exception as applied to this case, rather than issuing bright-line diktats, we would be employing the traditional common law method of deciding novel questions of law, which method recognizes the limitations of human ingenuity and wisdom, by limiting our decisions as precisely as possible to the case at hand. The common law method permits us to evaluate different cases over time to discern the most sensible rule given the technologies that develop; I’m afraid the majority opinion short-circuits this process in an area where the capabilities of computer software are still rapidly evolving.3
*1019ill
Third, one reason for adopting a more deliberative, common law approach is because the majority’s guidelines raise substantial practical problems, which the opinion fails to address, with respect to the consequences of the bright line rules the majority establishes. For example, the Supreme Court has held that if evidence seized by police is contraband, it can never be returned:
The Fourth Amendment was designed to protect both the innocent and the guilty from unreasonable intrusions upon their right of privacy while leaving adequate room for the necessary processes of law enforcement. The people of the United States insisted on writing the Fourth Amendment into the Constitution because sad experience had taught them that the right to search and seize should not be left to the mere discretion of the police, but should, as a matter of principle, be subjected to the requirement of previous judicial sanction wherever possible.... [Thus, i]t was error [in this case] to refuse petitioners’ motion to exclude and suppress the property which was improperly seized. But since this property was contraband, they have no right to have it returned to them.
Trupiano v. United States, 334 U.S. 699, 709-710, 68 S.Ct. 1229, 92 L.Ed. 1663 (1948) (emphasis added), rev’d in part on other grounds, United States v. Rabinowitz, 339 U.S. 56, 70 S.Ct. 430, 94 L.Ed. 653 (1950), overruled in part on other grounds, Chimel v. California, 395 U.S. 752, 89 S.Ct. 2034, 23 L.Ed.2d 685 (1969).
Child pornography is contraband,4 and there is no right for a citizen ever to possess it. The opinion does not explain whether a “third party” computer technician — as the majority requires the police to hire if they will not voluntarily foreswear use of the “plain view” exception— who comes across child pornography yet refuses to report it immediately, or returns it as part of data seized and searched, can himself be held liable for the possession of child pornography. Indeed, the opinion does not explain whether the police must destroy illegally seized contraband that implicates a third party — a participant in the contraband pornography still or video — who does not have a property interest in the illegally seized evidence and, therefore, would not have standing to suppress its use in evidence. Such is the problem with issuing bright line rules at the expense of developing the law through the common law method.
Perhaps the test should be whether, under the facts of the particular case, the police have executed the search warrant in a “narrowly tailored” manner. For example, in the electronic search context, a dotcom start-up company may well create software next week or next month that can accurately search through electronic storage media to report only the handful of files most likely responsive to a warrant.5 *1020Then a computer specialist conducting the search would be able to open a small number of potentially responsive files, selected by the software, instead of thousands. In a perfect world, the specialist would find the responsive file on the first click; then, the. specialist must stop. But suppose the responsive file is the second file opened, and the first file contains an image of child pornography. According to the majority, the specialist must ignore the child pornography and cannot use it as probable cause for a future search warrant. But why should that be, given the limitations on the scope of the search the software put in place? Are we certain that such a search automatically and always fails to comport with the Constitution?
Rather, it is more sensible to permit the specialist to report the child pornography image to a police investigator, who may seize the file to use as evidence at a trial. The investigator may also opt to return to the magistrate and explain that, although the government employed the narrowest search protocol possible, the specialist unavoidably “came across”6 contraband, the illegality of which was “immediately apparent.” 7 The magistrate ought to have discretionary power to expand the warrant to permit a search for additional child pornography, after deciding whether the specialist’s search of the data that uncovered the first child pornography file was truly based on the most limited possible search parameter, or if the specialist took some volitional step to view the data beyond what was necessary to procure the evidence in the warrant. If the latter, the seized file would be subject to a motion to suppress, or, if non-contraband, a motion for return of evidence under Rule 41(g). In general, the magistrate is best suited, and is perfectly capable, of deciding whether a “plain view” seizure is based on a sufficiently narrow search of seized intermingled files to comport with the Fourth Amendment, given the specific facts of the case.
For the reasons discussed above, I concur in the majority’s analysis of the facts and law as applied to this case. I do not, however, concur in the majority’s proposed guidelines. It would be nice to give magistrates clear “guidance” when possible. But that is true in all cases, and yet we still approach that goal by issuing rulings on the facts before us, and nothing more.
. I agree, however, with Judge Callahan and Judge Ikuta that the later-filed Cooper Order does not have preclusive effect on the earlier-*1016filed Mahan Order. That said, the majority reaches the correct result in this case even if the Cooper order is not given preclusive effect.
. The record reveals the spreadsheet was in Microsoft Excel format. Agent Novitsky viewed the spreadsheet on or about April 8, 2004. By that time, Microsoft Excel was the most widely used spreadsheet application available for Microsoft Windows and Mac OS X operating systems. See http://en.wikipedia. orgAviki/MicrosofLExcel. Microsoft Excel became commercially available in 1985, and currently sells for $229 retail. See http:// office.microsoft.com/en-us/excel/FX 102464391033 .aspx.
In Microsoft Excel, to avoid scrolling to the right and viewing the results column for all of the ballplayers instead of just for the targeted ten, all Agent Novitsky had to do was the following: While depressing and holding the Control key, he would click on the numbers on the left side of the spreadsheet that corresponded to the rows that contain the names of the targeted ballplayers. The rows containing those ballplayers’ names would become highlighted. Novitsky would then release the Control key. He would next go to the top of the screen, click on the "Edit” menu, and choose "Copy.” Then, he would click on the "File” menu at the top of the screen, and choose "New Blank Workbook.” When the new blank spread-sheet appeared on the screen, he would click on the "Edit” menu in the new blank spreadsheet and choose "Paste.” The rows of the ten targeted ballplayers selected in the original spreadsheet and only those rows — would appear in the new spreadsheet. Novitsky would then scroll to the right in the new blank spreadsheet and would see only the testing results for the targeted ballplayers for whom he had probable cause to search and seize.
. Further, if the majority's guidelines are now the law of our circuit, they conflict with the more cautious, common law-style approach of the Tenth Circuit, which has implicitly recognized the "plain view” exception exists in the context of electronic searches but has not delineated its precise scope. See United States v. Carey, 172 F.3d 1268, 1274 (10th Cir.1999) ("Although the question of what constitutes 'plain view' in the context of computer files is intriguing and appears to be an issue of first impression for this court, and many others, we do not need to reach it here. Judging this case only by its own facts, we conclude the items seized were not authorized *1019by the warrant. Further, they were in closed files and thus not in plain view.”). Here, in my view, the testing results of the ballplayers other than the targeted ten were also in "closed files and thus not in plain view.” The investigating officer “opened” the files through his volitional act of scrolling to the right on the spreadsheet, an act he knew would produce testing results not only for the targeted ten, but also for the remaining ballplayers listed on the spreadsheet. We do not really have to go beyond Carey.
. See United States v. Mack, 164 F.3d 467, 473 (9th Cir.1999).
. See, e.g., Verne Kopytoff, Google Reveals Tool That Seeks Similar Images, S.F. Chron., April 21, 2009, at C3 (noting that Google "introduced an experimental tool Monday [April 20, 2009] that allows users to narrow their search results to photographs that are alike in terms of their content, perspective and color”).
. Coolidge, 403 U.S. at 465, 91 S.Ct. 2022.
. Id. at 466.
. Federal Rule of Criminal Procedure 41(g), which was previously codified at Rule 41(e), states as follows:
Motion to Return Property. A person aggrieved by an unlawful search and seizure of property or by the deprivation of property may move for the property's return. The motion must be filed in the district where the property was seized. The court must receive evidence on any factual issue necessary to decide the motion. If it grants the motion, the court must return the property to the movant, but may impose reasonable conditions to protect access to the property and its use in later proceedings.