Intellectual Property Bulletin Summer 2006

Patent Rights and Open Source—Can They Coexist?

By: Laura A. Majerus


A patent gives its owner the right to exclude others from making, using and selling the claimed invention. Thus, patent rights give a patentee great control over who uses his or her invention. In contrast, the basic idea behind distributing software under an open source license is that anyone should be able to view and use the "source code" of the computer program and modify it for their own purpose. (The source code is the human-readable version of the software.) The author of the program must make the source code available to others. Anyone can modify the source code without obtaining permission from the author. A business decision to distribute software under an open source license affects how the author of the software may be able to use his or her patent rights, but does not affect whether one can or should apply for patent protection. A business decision to release software under an open source license (or to incorporate such software in a proprietary product) may grant certain patent licenses to people who receive the software, and the patent owner cannot control who these receivers will be.

Defining Open Source Licenses

It is difficult to talk about open source licenses in the abstract. A large number of licenses have been identified as open source (see, e.g.,, and each has its own quirks. One of the most frequently referenced open source licenses is the Gnu Public License (the GPL). A copy of the GPL can be obtained at

The GPL takes the open source concept that people should be able to view and use source code one step further, by requiring that any software that is a derivative work of the GPL'd software also be distributed under the GPL. The GPL is a license of copyright rights owned by the original author of a software program. When an original author releases his or her software under the GPL, he or she licenses those who receive the software to copy, distribute and produce derivative works of the software, providing that they then, in turn, release the software and derivative works of the software under the GPL and make available the source code of the software and the derivative works.

The GPL has been interpreted as defining a derivative work of the GPL'd software as modifications of the software and as any software that "links" to the GPL'd software. Thus, the GPL'd software potentially "infects" all software into which it is linked and forces that software also to be distributed under the GPL. Although this definition of derivative work has not been addressed by the courts, it lends enough uncertainty to the use of GPL'd software that many commercial software developers do not use GPL'd software developed by others because of the possibility that doing so would force distribution of the source code for the commercial product under the GPL. There is also a version of the GPL called the Library, or Lesser, GPL that does not have this infection property.

Certain entities choose to distribute their original software products under the GPL. For example, the Linux operating system is distributed under the GPL. Thus, companies and individuals can obtain the source code for the Linux operating system and can modify the software as long as they distribute the modified software under the GPL as well.

Patent Rights of the Author or Modifier

If a company plans to distribute its software under the GPL, does it make sense to obtain patent protection as well? Open source licenses, and the GPL in particular, do not bar a software author from obtaining patent protection on inventive aspects of his or her software. Nor does the GPL bar a programmer from including one's patented features when he or she modifies some one else's software previously distributed under the GPL. But patent enforcement rights are severely curtailed when software embodying the patented technology is distributed under the GPL.

The preamble of the GPL states: "Finally, any free program is threatened constantly by software patents. We wish to avoid the danger that redistributors of a free program will individually obtain patent licenses, in effect making the program proprietary. To prevent this, we have made it clear that any patent must be licensed for everyone's free use or not licensed at all."

Once a patent is granted, a patent owner typically has the right to exclude others from making, using and selling his or her patented invention if they so choose. Setting aside any arguments that the preamble of the GPL is somehow not a part of the license, it seems clear that an author or modifier who distributes software under the GPL cannot assert his or her patent rights against subsequent users and redistributors of the GPL'd software. Thus, there is at least an implied license to those who receive the GPL'd software in any patents covered by the software.

Why then, would anyone want to obtain a patent on an invention that is going to be distributed under the GPL? There are several reasons: 1) the author may plan to license the patent to others to produce a revenue stream; 2) the author may want to assert his or her patent rights against redistributors who do not conform to the GPL license terms (for example, by failing to redistribute under the GPL); 3) the author may want to have patent rights to use as an offensive or defensive weapon against infringers who are not using the GPL'd software; and 4) the author may plan to also distribute a non-GPL'd version of the software. For example, while the author may not be able to use his or her patent rights against people who receive and redistribute his or her GPL'd software, the author can use these rights against his or her competitor who sells a competing product that incorporates the invention that is not a derivative work of the author's original code. According to one reasonable interpretation, the GPL only precludes the patentee from asserting his or her rights against people who are practicing the invention by using one's GPL'd software. People who independently create other software are not subject to this implied license. Interestingly, the author could assert his or her patent rights against a competitor who is releasing independently developed software under the GPL, as long as it is not based on the original author's distribution. The fact that the infringer distributes under the GPL is irrelevant as to whether one is infringing patents of others. The original author has not given permission for his or her competitor to use the patented technology.

Existing Licenses

What if the original author or modifier is a patent licensee under an existing license? For example, the original author might have a license from a third party for an invention included in the GPL'd software. The original author may, for instance, release Internet software that incorporates an encryption technology covered by a patent licensed from a third party. If the original author of the GPL'd  Internet software has been granted the right to sublicense, one may be able to include the patented technology in his or her GPL'd software, because he or she can give a sublicense to subsequent users. Conversely, of course, the author cannot distribute his or her software if he or she cannot sublicense subsequent users. The GPL addresses this issue:

If you cannot distribute so as to satisfy simultaneously your obligations under this License and any other pertinent obligations, then as a consequence you may not distribute the Program at all. For example, if a patent license would not permit royalty-free redistribution of the Program by all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Program.

Thus, if an author does not have the right to sublicense or if the sublicense would require the payment of royalties, the author must choose between not including the patented technology and not releasing the software. Companies planning to distribute GPL'd software should consider, therefore, whether the software is subject to existing patent licenses. If so, and if sublicensing rights cannot be obtained, the patented technology should not be included in the GPL'd distribution. If existing formal license agreements cover technology in software that a company plans to distribute under the GPL, the existing license should be carefully reviewed to determine whether the license includes the right to sublicense. The converse, of course, is that patent owners who license to another party who has stated intent to distribute under the GPL should consider whether they want to grant the right to sublicense. In theory, there is no limit to how many other parties will eventually receive the GPL'd software and become sublicensees of the patent.

What is the Limit of an Implied License Under the GPL?

As discussed above, users and modifiers of GPL'd software have at least an implied license to use patented inventions included in the GPL'd software. The users probably do not, however, have a blanket license to use the invention in other circumstances. Thus, if a programmer who uses GPL'd software develops software that uses the patented invention outside the scope of the GPL'd software, one arguably does not have a license to use the invention in this manner. If a user copies a portion of the GPL'd software and includes it in this software, he or she may arguably have a license because his or her new software is based on GPL'd software.

Patents in Non-U.S. Countries

There is no such thing as a worldwide patent. Patents are granted on a country-by-country basis. The GPL states:

If the distribution and/or use of the Program are restricted in certain countries either by patents or copyrights the original copyright holder who places the Program under this License may add an explicit geographical distribution limitation excluding those countries, so that distribution is permitted only in or among countries not thus excluded. In such cases, this License incorporates the limitation as if written in the body of this License.

The GPL seems to be referring to the situation where the patent is owned by a third party. It does not address the possibility that a patent holder and author of the program are the same. This assumption seems correct because, if the patent holder and author are the same entity, the users of the program would have an implied license to use the patented invention and would have an implied license to distribute and modify the software, because this is the whole purpose of the GPL. Under those circumstances, the distribution or use would not be restricted.


Because there are no decided court cases interpreting the GPL, there are business risks involved both in releasing software under the GPL and in using software received from others under the GPL. Typically, disputes involving the GPL are settled informally out of court. It remains to be seen how courts will interpret the rather vague references to patents in the GPL, but it seems certain that, as more commercial entities begin to incorporate GPL'd code into their products, it is only a matter of time before court cases involving the GPL will begin to appear. While some of these cases will involve the copyright law questions raised by GPL, others will be patent cases, where the GPL is used as a potential defense by infringers who received GPL'd software incorporating the asserted patent. Corporate counsel should be aware of the risks involved in the GPL before making decisions concerning this license.

Going to Videotape: A Juror's Introduction to the Patent System

By: Heather N. Mewes and Darren E. Donnelly

Patentees enter every jury trial with one distinct advantage: an official, ribboned copy of the patent at issue in the case.

This trial exhibit comes embossed with the seal of the Patent and Trademark Office and signed by the director of that office. It also includes a declaration that "[t]he requirements of law have been complied with, and it has been determined that a patent on the invention shall be granted under the law." The official copy of the patent is a visual and vivid reminder to the jury of the presumption that a patent is valid.

The presumption of validity can appear as an impossible hurdle for a defendant to overcome. A jury is commonly (and often repeatedly) instructed that the defendant must show invalidity by "clear and convincing evidence." See, e.g., AIPLA's Model Patent Jury Instructions (2005).

This burden is worse than it sounds. In plain English, "clear and convincing" simply means that the defendant must show that it is "highly probable" that the patent is invalid. See, e.g., Federal Circuit Bar Association Model Patent Jury Instructions (instructing that "clear and convincing" is not the same as proof "beyond a reasonable doubt"). However, effectively and convincingly explaining that the patent office is not infallible and that the jury (as nonexperts) should second-guess its judgment is an important issue to a defendant asserting invalidity in front of a jury.

Defendants often turn to patent law experts to explain the practices and procedures of the patent office — and to make it clear that the patent office can make mistakes because of limited time and resources. However, such experts are increasingly finding a chilly reception in some district courts. See, e.g., Minebea Co. v. Papst (D.D.C., Jun 2005) (excluding patent law expert from testifying regarding practices and procedures of the patent office); Syngenta Seeds Inc. v. Monsanto Co. (D. Del., Sept 2004) (same).

These district courts are instead turning to a Federal Judicial Center video entitled Introduction to the Patent System. This 17-minute video was released by the Federal Judicial Center in October 2002 and is designed to be shown to jurors in patent trials, for example, during preliminary jury instructions. The Federal Judicial Center claims that "[s]pecial care was taken to ensure that [the video] provides an impartial and objective view of the patent process." However, since its release, this video has attracted criticism from litigants on both sides even while gaining acceptance from district judges.

The video is a combination of narration and dramatization. It contains several segments, generally covering: (1) what patents are; (2) invention and filing of a patent application; (3) parts of a patent application; (4) examination of a patent application in the patent office; (5) enforcement; (6) the defense of invalidity; (7) the jury's role; and (8) the differing burdens of proof for infringement and invalidity. A sample patent used in the video is also available for the jury's reference.

Quantitatively, the video devotes the largest share of its time to patent examination (about 5 ½ minutes ). After that, the segments, on invention and filing and the parts of a patent application are each given about 2 ½ minutes. The first segment, covering what patents are and the "patent bargain" runs slightly longer (about 2 minutes) than the fifth segment about the defense of invalidity (about 1 ½ minutes). The remaining substantive segments are relatively brief (about 30 seconds each).

Qualitatively, the segments on invention and filing and examination are both dramatized via repeating characters — the inventor, his or her patent attorney and the patent examiner — acting out their respective roles behind descriptive narration. Even the segment on what patents are is accompanied by stirring imagery: the U.S. Constitution, an official, ribboned copy of a patent and a montage of industrious scientists and archetypal inventions including Edison's lightbulb and the Wright Brothers' plane. In contrast, the segments on the defense of invalidity, enforcement, the jury's role and the burdens of proof are primarily just the narrator.

Thus, criticism of the Federal Judicial Center video has focused largely on the imagery rather than the narration. Indeed, the actual narration is relatively noncontroversial, and similar explanations are found in patent-specific jury instructions. See, e.g., N.D. Cal. Model Patent Jury Instruction A.1 (2004) ("What a patent is and how one is obtained"). For example, on the burden of proof segment, the video takes a middle road — it does not use the "clear and convincing" language that many patentees favor (rather, it uses "highly probable"). However, it does explain that this higher burden is due to the fact that the patent office is presumed to have done its job correctly.

The imagery is more controversial. Both patentees and defendants can point to imagery in the video that is favorable to their position. For example, patentees can point to repeated shots showing conscientious, hard-working examiners and emphasize the extensive examination process illustrated in the video. Patentees can also point to the many favorable associations in the video to well-known inventors and inventions and other positive imagery.

Defendants, on the other hand, can point to the mounting piles of applications in the examiner's office and the scene of endless rows of files at the patent office (think Raiders of the Lost Ark) with an implication that examiners are overwhelmed and overworked. The video even explains in showing this imagery that one of the reasons that the jury is asked to decide invalidity is that the patent office may make mistakes.

Some, nonetheless, perceive the video as creating a disadvantage for defendants through an extended and relatively favorable portrayal of patent office examination and only fleeting images of piles of files (there being no vivid imagery of the public being freed from improvidently granted patent monopolies). The video does devote substantial time to the process of examination — more than any other segment.

To be fair, the purpose of this extended segment is to explain a complicated process that many in the jury have no experience with, but which may be very relevant to the issues in dispute. However, the video's extended illustration on patent office examination can convey the impression that this process is more extensive than it may be in a given case. Moreover, scenes showing the examiner carefully reviewing the specification and claims and searching prior art databases have the impact of visual dramatization. Qualifying information — for example, that an examiner may not have access to all of the prior art or may only have limited time to spend on any one application — is often only narrated. While the video balances this with other content, a defendant confronts the relative cognitive impacts of narration versus actors' portrayals.

Finally, there is also some criticism surrounding what the video does not cover. As it is only intended as an introduction, the video does not comprehensively cover all defenses that can arise in a patent case. The jury would presumably be separately instructed on these defenses (and even on the defenses covered in the video). However, a defendant must overcome any impression in the jury's minds that omission from the video means that a defense lacks merit or requires more-convincing proof.

Without explanation, the use of different media for instructing the jury on the background of the patent system could have the unintended effect of demoting the importance of other instructions. This is particularly of concern where a long (and often boring) list of jury instructions is delivered orally to the jury — jurors may not retain this information as well as they retain the video presentation, resulting in undue weight being given to the video.

Further, there is no guarantee that either a patentee or a defendant will be able to supplement the information in the video with testimony from a patent law expert – even though the video does not cover everything about the practices and procedures of the patent office. For example, in Minebea, the court found that testimony from a patent law expert on this topic would be irrelevant or cumulative. See Minebea; see also Syngenta Seeds ("I have determined that this video is a sufficient basis for instructing jurors"). This can have the effect of preventing a patentee or defendant from making an additional point about the examination process not covered in the video. For example, many defendants want the jury to understand how much time an examiner typically spends examining an application, a fact the video alludes to but does not address explicitly.

While courts are naturally drawn to neutral, approved sources for instructing juries on complex and contentious matters, finding ones that allow advocates to fairly present the messages important to their cases can be challenging. The Federal Judicial Center video is a unique experiment in that the patent litigation bar played an important role in developing it and thus could recognize such issues. As experience with it increases, the patent system evolves and issues with the Federal Judicial Center video or other sources come to light, courts and litigants will have to weigh their options carefully in choosing how to introduce the jury to the patent system.

Quick Updates

Court Compels Google to Release Search Engine Data

In a much-awaited legal decision, Judge James Ware of the Northern District of California granted in part, and denied in part, a sweeping U.S. Department of Justice subpoena for access to information regarding the Google search engine. The Justice Department requested that Google provide 5,000 search queries and a 50,000-URL data-sample from Google's search index. In Gonzales v. Google, Inc. (N.D. CA., Mar 2006), Judge Ware ordered Google to provide the 50,000-URL data-sample, but not the 5,000 search queries.

The government subpoenaed Google, and other search-engine providers (America Online, Yahoo! and Microsoft), to produce search-engine information for a Pennsylvania case involving the Child Online Protection Act. Although the subpoenaed companies are not a party to that case, the Government argued that the information was necessary to determine the effectiveness of content filters for screening sexually explicit material from minors. While all other search engines met the Government's request, Google objected, citing concerns over trade secret disclosure, user privacy rights, and potential user backlash for disclosing user information.

In his Order, Judge Ware considered whether the Google information sought was relevant to the Pennsylvania case, as required by Rule 26(b) of the Federal Rules of Civil Procedure ("FRCP"). The judge chastised the Government for a striking failure to explain its intended use for the information even though the Government had "already been to the pond and fished" with data received from other search engines. Nonetheless, given the broad Rule 26(b) definition of relevance and the narrow scope of the subpoena, the judge concluded that the Google information was relevant.

Judge Ware also considered the burden placed upon Google in producing the information requested and whether the Government showed a substantial need for the information, as required under Rule 45(c)(3)(B) of the FRCP. The judge considered the confidentiality of the Google information and found it to be "somewhat commercially sensitive." The judge further expressed concern that allowing a narrow sample of Google's proprietary information could become the "thin blade of the wedge" enabling broader, future disclosure of Google trade secrets. Still, Judge Ware concluded that, given Google's position as the search-engine market leader, the Government's study might be "significantly hampered" without at least the 50,000-URL sample. The judge found the request for 5,000 search queries duplicative, however, and concluded that the "marginal burden of loss of trust by Google's users" outweighed the Government's request for the search queries.

Judge Ware further raised sua sponte user-privacy concerns that disclosing search-query information might expose personal information included in text strings of some search queries (such as social security numbers, personal interests, and so on). Although the Government assured the judge that requested Google data would be used only in the Pennsylvania litigation, Judge Ware noted that in some cases the Government could be obligated to pursue that information for other purposes (e.g., to investigate queries like "bomb placement white house"). Ultimately, however, Judge Ware made no further findings regarding these privacy concerns since Google was not ordered to produce the search queries.

Federal Circuit Resolves Split on Waiver of Work Product Immunity in Advice of Counsel Cases

In an order granting writ of mandamus, the Federal Circuit clarified that a defendant who asserts an advice of counsel defense to willful infringement does not waive immunity for work product never communicated by counsel to the defendant. In re Echostar (Fed. Cir., May 2006).

When TiVo, Inc., sued EchoStar Communications Corp. asserting willful infringement, EchoStar defended the charge of willfulness by asserting a noninfringement opinion from its in-house counsel. EchoStar also sought an opinion from outside counsel after the suit commenced, but ultimately chose not to rely on that opinion. The district court held that by asserting the advice of counsel defense, EchoStar waived both the attorney-client privilege and work product immunity relating to advice from any counsel regarding infringement.

The Federal Circuit first affirmed the broad waiver of attorney-client privilege found by the district court. In particular, the district court's order required EchoStar to produce all communications with both in-house and outside counsel regarding infringement. Also, the order emphasized that waiver of attorney-client privilege extends to all communications with any counsel relating to the subject matter of the opinion (usually,whether a patent is valid, enforceable and/or infringed), and also encompasses communications made after the filing of the complaint, at least where ongoing infringement is at issue in the litigation. Thus, caution is warranted in communications between defendants and their litigation counsel for the duration of a litigation, even where defendants employ separate opinion counsel and litigation counsel.

The Federal Circuit next reversed the portion of the district court's order requiring production of work product that was not communicated to the defendant. Resolving a split among district courts on the scope of work product waiver, the Federal Circuit emphasized that waiver focuses on what the alleged infringer knew and believed, not the state of mind of opinion counsel. Opinion work product "deserves the highest protection from disclosure" and is discoverable when communicated to the client only because it becomes evidence of a nonprivileged fact — namely, what the client knew and believed about infringement. A counsel's legal opinions and mental impressions never communicated to the client do not acquire such factual characteristics and thus are not within the scope of waiver.

Note, however, that documents reflecting work product can be discoverable even if they are not communicated to the client, if they discuss an attorney-client communication. Accordingly, the Federal Circuit cautioned the parties to take special care in such instances to redact any legal analysis and opinion work product that was not communicated. Going a step further, counsel may be well advised to limit references to attorney-client communications in internal documents.

The Trademark Dilution Revision Act of 2006

Last month, long-anticipated revisions to federal trademark dilution law took another significant step as the Senate passed the Trademark Dilution Revision Act of 2006 ("the Act"). Once reconciled with a similar bill passed by the House of Representatives in 2005 and formally enacted, the Act will substantially modify the Federal Trademark Dilution Act of 1995 (FTDA), to the benefit of owners of famous marks.

Enacted in 1995, the FTDA affords trademark owners a remedy for the dilution of famous marks, even in the absence of a likelihood of confusion between the owner's use of the mark and another's use of the same or similar mark. In Moseley v. V Secret Catalogue, Inc. (U.S. Sup. Ct., Mar 2003) however, the Supreme Court held that the statute required proof of actual injury to the economic value of a famous mark, rather than merely a "likelihood of dilution." The court reached this conclusion based primarily on the text of the statute itself, which stated that "[t]he owner of a famous mark shall be entitled… to an injunction against another person's commercial use in commerce of a mark… if such use causes dilution of the distinctive quality of the mark…."

In response to this decision, both versions of the current legislation seek to overturn the Moseley decision, establishing a "likelihood of dilution" standard and permitting relief "regardless of the presence or absence… of actual economic injury."

Both versions of the Act also clarify that the distinctiveness requirement for protection under federal antidilution law may be met by showing either inherent or acquired distinctiveness. In so doing, Congress rejects the holding of TCPIP Holding Co. v. Haar Communs., 244 F.3d 88 (2nd Cir. 2001), in which the Second Circuit held that only inherent distinctiveness is sufficient to meet the distinctiveness requirement under the FTDA. Thus, even marks that are deemed to be "descriptive" and not inherently strong may be protected by the revised federal dilution statute.

Still to be resolved is the breadth of the exemptions under the law. The House version, while containing various exemptions from liability for comparative advertising, parody and criticism, drew fire from free speech advocates and other groups by apparently abolishing the "commercial use" requirement that currently appear in the FTDA. The Senate version answers these concerns, at least in part, by reinserting this limitation and exempting from liability "[a]ny noncommercial use of a mark."

Fair Use and the Historical Nature of "Groovy" Expression

A recent Second Circuit copyright decision underscores the importance of context (and perhaps the passing of time) in determining whether a given use falls within the fair use safe harbor. In Bill Graham Archives v. Dorling Kindersley, Ltd,  (2nd Cir., May 2006), Bill Graham Archives sued the publisher Dorling Kindersley for reproducing seven copyrighted images of Grateful Dead concert posters and tickets in a commercial, coffee-table book. Titled, Grateful Dead: the Illustrated Trip, the at-issue publication featured a timeline chronicling important dates in the band's history. The timeline incorporated explanatory text and images, including the seven allegedly infringing images, which were reproduced in thumbnail format. Affirming the district court's summary judgment in favor of the defendant, the Second Circuit explained that the otherwise-expressive images now tucked within the confines of an illustrated historical narrative constituted a transformative fair use.

The outcome of a fair use analysis typically depends on how a court weighs in on the fourth fair use prong—the effect of the use on the market or potential market for the copyrighted work. However, the Bill Graham decision rested the bulk of its argument on the first fair use prong: the purpose and nature of the use. That factor requires a court to consider, among other things, the extent to which a would-be infringement is transformative, that is, whether a given use meaningfully alters the underlying work, adding something new and socially useful to society's already existing stock of creative works, or whether the use merely exploits the original expression.

In Bill Graham, the Court found the publisher's use was transformative. While the reproduction of images was designed in part to "capture the eye," the images served more prominently as cultural references and thus lacked the expressive purpose of the original images. The court explained, "Originally, each of the [original] images fulfilled the dual purposes of artistic expression and promotion. The posters were apparently widely distributed to generate public interest in the Grateful Dead and to convey information to a large number of people about the band's forthcoming concerts. In contrast, [Dorling Kindersley] used each of [Bill Graham's] images as historical artifacts to document and represent the actual occurrence of Grateful Dead concert events featured on Illustrated Trip's timeline."

The court continued to employ a species of the transformative use argument when analyzing the fourth fair use factor. When Bill Graham argued that Dorling Kindersley's use of the images harmed its licensing revenue, the Court explained, "[i]n a case such as this, a copyright holder cannot prevent others from entering fair use markets merely by developing or licensing a market for parody, news reporting, educational or other transformative uses of its own creative work."

Patent Injunctions Now Under Greater Scrutiny

In a recent decision, the United States Supreme Court held that prevailing plaintiffs-patentees seeking permanent injunctive relief must satisfy a four-factor equitable test before injunctive relief will be granted. eBay, Inc. v. MercExchange, L.L.C. (U.S. Sup. Ct., May 2006). An analysis of the decision is available at .

Fenwick & West's Intellectual Property Group offers comprehensive, integrated advice regarding all aspects of the protection and exploitation of intellectual property. From providing legal defense in precedent-setting user interface copyright lawsuits and prosecuting software patents to crafting user distribution arrangements on behalf of high-technology companies and implementing penetrating intellectual property audits, our attorneys' technical skills enable the Firm to render sophisticated legal advice.

The contents of this publication are not intended and cannot be considered as legal advice or opinion.

© 2011 Fenwick & West LLP. All Rights Reserved.

We appreciate your feedback!

If you have questions, comments, or suggestions for the editors of the IPB, you can e-mail them to

For subscription requests and address changes, please e-mail​​