1997 B.C. Intell. Prop. & Tech. F. 041501

Misunderstanding RAM:
Digital Embodiments and Copyright

Kristen J. Mathews  fnA

April 15, 1997


In the opinion of the United States federal courts, digital software embodied in a computer's Random Access Memory (RAM) is sufficiently fixed to constitute a "reproduction" under the Copyright Act.1 As a reproduction, the creation of the RAM embodiment, or the loading of software into RAM, is a potential copyright infringement. However, a close reading of the Act and its legislative history reveals that a digital work embodied in RAM should not be considered a reproduction of the work. Furthermore, including digital works embodied in RAM as reproductions is a poor fit in light of the policy behind the Copyright Act. This would mean that every time a person opens a computer program, he or she might be infringing a copyright. The courts' widely criticized finding can be explained, at least in part, by early law makers' confusion about computer memory and inability to fit RAM into previous constructs. Courts and law makers have built on each others' flawed or non-existent analysis of RAM embodiments as reproductions since the 1976 Act was being drafted. These approaches ignored the purpose behind the fixation requirement when interpreting it. Since digital embodiments in RAM do little harm to a copyright holder, their categorization as reproductions is not consistent with the policy behind the fixation requirement for reproductions.2

In response to the intuitive incorrectness of the current law and its contradiction with the policy behind copyright law, critics have made several arguments against finding an embodiment in RAM to be an infringement.3 Critics argue that RAM embodiments should be afforded one of the defenses of fair use, copyright misuse, or per section 117(1), that the copying is an essential step in the program's utilization. This paper, however, focuses on another argument born of the same policy considerations: That a RAM embodiment is not sufficiently fixed to be a reproduction under the Copyright Act's definitions.

The Reproduction Right

One of a copyright holder's five rights under the Copyright Act is the exclusive right to make and to authorize the making of reproductions of the copyrighted work.4 The Act defines "copies" as "material objects . . . in which a work is fixed by any method now known or later developed, and from which the work can be perceived, reproduced, or otherwise communicated, either directly or with the aid of a machine or device."5 For the purposes of the Copyright Act, a work is 'fixed' in a tangible medium of expression when its embodiment in a copy or phonorecord, by or under the authority of the author, is sufficiently permanent or stable to permit it to be perceived, reproduced, or otherwise communicated for a period of more than transitory duration."6

"Fixation" is a requirement for two different substantive copyright provisions. One, a work must be "fixed" to be eligible for copyright protection. 7 Second, in order for a copy of the work to be an infringing reproduction, it must also be "fixed."8 The definition for "fixed" provided by the act is tailored to the requirements for copyrightability as opposed to those for reproduction. This is evident by the definition's use of the words, "in a tangible medium of expression" which mirrors the act's requirement for copyrightability in section 102, not the Act's reproduction provisions.9 Nevertheless, law makers and courts have applied the section 101 definition of "fixed" to the reproduction provisions as well. This application is clearly supported by the Act's legislative history. Both the House and Senate Reports refer to the section 101 definition of "fixed" in their discussion of the reproduction provisions.10 Furthermore, the Commission on the New Technological Uses of Copyrighted Works (CONTU) report, to be discussed further below, also uses section 101's definition of "fixed" in its reproduction analysis.11 However, although using section 101's definition of fixed for both the copyrightability and reproduction provisions is appropriate, interpreting the definition of "fixed" the same way for both provisions is not appropriate. The two provisions serve different purposes, and so should their fixation requirements. Interpretation of each fixation requirement should be in the context of its respective provision.
Because of the different contexts, consideration of the purpose behind the fixation requirement for both the copyrightability and reproduction provisions shows that drawing the same line in both provisions for what constitutes "more than transitory duration" is not appropriate. Nevertheless, courts and law makers have, thus far, interpreted the two provisions interchangeably.12 In both provisions, fixation serves the purpose of requiring that the physical embodiment lasts through time. However, several differences exist. For copyrightability, fixation serves the purpose of requiring that a work have a physical embodiment that lasts long enough to need protection, and that in exchange for protection, the author contributed something of lasting value to society.13 The fixation requirement also brings the Copyright Act within the power afforded to the Congress by the Constitution. The Constitution gives the Congress the power to "promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries."14 By "writings," the drafters envisioned something physically materialized. Further, fixation facilitates enforcement of copyrights by allowing exact identification of the copyrighted work in an infringement action. With these purposes in mind, the fixation requirement for copyrightability should be interpreted to require that a work last long enough to need protection, be identifiable, and contribute to society enough to deserve copyright protection.

For reproduction, on the other hand, the fixation requirement ensures that, for a copy to be an infringement, it must harm the copyright holder. Otherwise, no protection is needed because copies that do not harm copyright holders do not deter their creation of new works.15 With these purposes in mind, the fixation requirement for reproduction should be interpreted to require that an embodiment last long enough to be harmful to the copyright holder.

Understanding Computer Memory

A detailed understanding of computer memory is necessary to apply copyright reproduction provisions to RAM embodiments.16 Computer memory is far more complex than courts and law makers have recognized, and its use varies greatly with the architecture of a particular computer, operating system, software, and with the type of digital work in question. The term "computer memory" is often misused. Technically, "computer memory" refers only to the RAM of a computer. However, in common usage the term "computer memory" is often used to refer both to RAM and to permanent or semi-permanent forms of storage such as hard drives or floppy disks, because both "memorize" information. In fact, much of the legislative history relied on was written before hard drives were common, and thus before the distinction was well known. This ambiguity has added confusion to copyright law's treatment of RAM because law makers and courts have not used clear and explicit language when referring to RAM and hard drive "memory."18 When the address is reallocated or the electricity cut off, the embodiment is lost completely. Data in RAM, both secondary and primary, is stored electronically, not magnetically as on a hard disk. In other words, the zeros and ones that compose digital data are merely high or low electronic field states. Once the electricity is gone, the registers can no longer sustain the high field states and the data is lost.
Different kinds of digital files are treated differently in RAM according to the kind of file, the instructions of the application program being used, and the computer¹s architecture. This varying treatment effects how transient a RAM embodiment is, yet courts have not addressed this complexity. They have relied on the premise that a program remains in RAM as long as it is open. However, if a system uses virtual memory, portions of a program not in immediate use may be stored in a temporary file on the hard drive instead of RAM. Secondly, when a file is "opened," or loaded into RAM, the application program being used may govern how much of the file is put into RAM. The fact that RAM might only embody currently used portions of a software means that the RAM embodiments last for less time. Each portion of a computer program lasts only as long as it is in immediate use, not as long as the entire program is in use.

So far, courts and law makers have not addressed the different ways that digital files might be stored in RAM. They have treated all RAM embodiments the same way, regardless of the particularities of the computer system. Moreover, courts and law makers have relied on the premise that data persists in RAM until the program is quit or the computer is turned off. While this may have been true in some cases, attorneys should be ready to argue that the RAM embodiment in their particular machine was more transient than those in the case precedents, and therefore not "fixed", due to different system architecture.

How the Law Got This Way

Modern law's conclusion that digital embodiments in RAM are reproductions can be attributed to law makers' early confusion and unclear statements about computer memory, subsequent changes in computer architecture to which courts have not adapted, and to courts building on each other's flawed or absent analysis. In the 1970's, during the drafting of the 1976 version of the Copyright Act, computer technology was still very new, and law makers did not fully understand how computer memory worked. Moreover, the first legal treatments of computer memory pre-dated the appearance of hard disks. Therefore, early legislative history does not distinguish between permanent or semi-permantnt storage and the RAM of a computer.19 Regardless of this lack of distinction, courts have since relied on this legislative history. The failure to distinguish between hard drive and RAM persists today in a recent case which made the same mistake.20

A computer's permanent hard disk is a completely different animal than its RAM, and they should be treated separately under copyright law. It is well established that copies of a digital work on a semi-permanent or permanent computer memory device, such as a computer's hard drive or Read Only Memory (ROM)21, are fixed under the Copyright Act.22 They are intended to, and do, remain in storage for more than a transitory duration, until they are actively deleted. Digital data embodied in RAM, however, is not permanent but in fact very transient. RAM is specifically designed to maximize speed at the expense of storage ability, moving data quickly through a computer's processor so that its instructions can be implemented rapidly. A portion of digital work might only remain in RAM during its immediate use, not as long as the file, as a whole, is "open. " The transient nature of RAM distinguishes it from more fixed storage and necessitates special analysis under the reproduction provisions.

The legislative history of the 1976 Copyright Act, in reference to fixation as it applies to reproduction, as opposed to copyrightability, sheds no light on whether the Congress intended RAM embodiments to be reproductions. Although the legislative history does not mention computer memory, it says that "showing of images on a screen or tube would not be a violation" of the reproduction right.24 Troublesomely, the legislative history's reference to fixation for the sake of copyrightability says images on a screen, in a tube, and in computer memory are not fixed, but its reference to fixation for the sake of reproduction only says images on a screen or tube are not fixed.
Congress' silence and apparent confusion about whether embodiments in computer memory are sufficiently fixed to be reproductions, nine pages after stating that works embodied in RAM are not copyrightable, accompanied by its parallel treatment of screen and tube images in both contexts, can be interpreted in two different ways. Either the Congress was sloppy in its legislative history and intended RAM embodiments to be neither copyrightable nor reproductions, or the Congress specifically left computer memory out of its discussion of fixation of reproductions, thereby regarding them as fixed. The legislative history is simply unclear.

The two legislative statements about fixation, the former regarding reproduction and the latter regarding copyrightability, have received mixed reactions among law makers, courts, and commentators. Law makers and courts have completely ignored this legislative history citing instead the CONTU report, discussed below.25 Conversely, their critics have referred to the legislative history's statement about fixation regarding copyrightability, instead of reproduction, to support arguments that embodiments in computer memory are not reproductions. Neither of these approaches is appropriate. Neither should courts and law makers ignore legislative history, nor critics support their arguments with inapplicable legislative history.

In 1974, as Congress was revising the Copyright Act, it realized that computers and photocopy machines presented complex issues in copyright that needed to be addressed by experts. Congress created the National Commission on New Technological Uses of Copyrighted Works (CONTU) which spent three years analyzing copyright as applied to these new technologies.26 CONTU's final report, released in 1978, recommended amendments to the Copyright Act. CONTU's recommendations were adopted by the Congress without additional comment. Courts regard the CONTU report as a form of legislative history.27

In a footnote reference to the Congressional Report, quoted above, which says that embodiments in the memory of a computer are not copyrightable, the CONTU report makes another ambiguous statement. It says, "Insofar as a contrary conclusion is suggested in one report accompanying the new law, this should be regarded as incorrect and should not be followed, since legislative history need not be perused in the construction of an unambiguous statute."28 CONTU's use of an ambiguous pronoun, "this," prevents readers from knowing what should be regarded as incorrect in the event of a contradiction, the legislative history's or CONTU's conclusion. Furthermore, CONTU's reference to the legislative history's language is inappropriate because the language regards copyrightability, not reproduction.

The CONTU report's statements about computer memory are too ambiguous to derive conclusion. The CONTU report states that "the placement of any copyrighted work into a computer - or the 'inputting' of it - is the preparation of a copy" and therefore a potential infringement of copyright."29 The CONTU report, written well before hard drives became prevalent, fails to draw the crucial distinction between the permanent or semi-permanent storage and RAM. If the Report refers to the installation of a digital work onto a computer's hard drive, then its statement says nothing controversial. However, if it refers to loading a digital work into the RAM of a computer, then it apparently surpasses the protection afforded by the Copyright Act. The Report does not recognize that RAM is far more transient than hard drive, and thus avoids the question of fixation. As a result of this general misunderstanding courts, relying in part on unclear legislative history, have established that loading copyrighted works into the RAM of a computer constitutes reproduction under the Copyright Act.
The CONTU report describes an instance when a work loses its ability to be copied, and is thus no longer an infringement. Unbeknownst to the CONTU, the instance it describes is precisely what happens to data when it is embodied in RAM. The report says, "Only when the program is inserted, instruction by instruction, into the processing element of the computer and electrical impulses are sent through the circuitry of the processor to initiate work is the ability to copy lost."30 This process is like that done in RAM, as data is sent through a computer's central processing unit. This language, therefore, seems to imply that RAM embodiments are not reproductions.

Although CONTU's intention for RAM embodiments is not clear from its Report, its general intentions for the use of computer programs is clear. In order to ensure that the rightful possessors of software are able to use the software in accordance with their reasonable expectations, CONTU recommended that the Congress adopt section 117 of the Copyright Act.31 Section 117, as recommended by CONTU, allows the rightful possessor of a computer program to make copies or adaptations if the copying is an essential step in using the program or for archival purposes.32 This provision allows a rightful possessor to load a program into the RAM of a computer, and thereby use it, regardless of any contractual restrictions imposed by the copyright holder. It is clear from section 117 that CONTU at least intended software possessors to be able to load their software into RAM. One commentator suggests that CONTU might have set up a nonexistent problem of RAM embodiments as reproductions to justify its recommendation of section 117.33

Review of the line of cases dealing with digital embodiments in RAM of a computer illustrates the flawed way in which the current law came into place. The first time a court faced a defendant who made a copy into the RAM of a computer, in 1986, neither the court, the attorneys, nor the defendant realized the important difference between permanent memory and RAM.34 In Bly v. Banbury Books, Inc., the defendant ran an early form of a printer driver off a floppy disk. This entailed copying the driver into the RAM of his computer, not to a hard drive or another floppy disk. Even though there was no case law on point at the time, the defendant conceded copyright liability, leaving the court to decide only damages.35 The defendant's concession was a mistake. The cases offered by the plaintiff in support of liability involved copies made onto permanent storage devices, not RAM.36 The court even commented on the deficiency of these authorities.37 In one authority, an audio-visual work was found fixed for the purpose of copyrightability where the work was embedded in read-only-memory, a permanent storage device.38 In the other, copies were found to be reproductions where they were fixed onto floppy disks, semi-permanent storage devices.39 The defendant might have prevailed by distinguishing these cases and arguing that embodiment in RAM is not a reproduction. His failure to do so has lead to a series of cases that disregard proper analysis and build on each other's faulty logic to create the rule that RAM embodiments are reproductions.

Two years later, a court faced a RAM copy again. Here too, the defendant did not argue that the embodiment in RAM was not a reproduction. In Vault Corp. v. Quaid Software Ltd., the plaintiff manufactured computer disks with copy protection features (preventing a user from creating a usable duplicate of software placed on the disk) used by software companies seeking to prevent the piracy of their programs.40 The defendant created a program that unlocks the plaintiff's protected disk so that users could create a usable copy of software housed on the plaintiff's disks. In order to create its product, the defendant copied the plaintiff's program into the memory of its computer to analyze the manner in which the program operated. The Court of Appeals for the Fifth Circuit found that when the defendant copied the plaintiff's program into the memory of its computer, it "reproduced" the program.41 However, the defendant's reproduction was exempted from the copyright act under section 117.42 Because the defendant did not argue that the copy in the RAM of its computer was not a reproduction, and relied solely on the section 117 defense, the court did not discuss or analyze the law behind reproduction or fixation at all, and cited no authority for it. It merely granted the defendant a defense under section 117.

The next time a court faced a defendant who copied software into the RAM of a computer, the court reacted similarly. In ISC - Bunker Ramo Corp. v. Altech, the United States District Court for the Northern District of Illinois found that a defendant who diagnosed, repaired, and refurbished equipment necessarily loaded operating software and diagnostic utility programs into RAM, thereby creating an infringing reproduction of the software.43 The court provided no analysis of its own but merely cited Nimmer on Copyright,44 and two cases. One of the cases it cited was Bly, discussed above, in which the court did not find that embodiment in RAM was a reproduction. In Bly, the defendant conceded liability and the court merely decided damages.45 The other case cited by the ISC court as authority, Micro-Sparc, Inc. v. Amtype Corp., was also not on point. It involved a defendant who input the source code of a program into a computer and saved it on floppy disks.46 It is not controversial that this kind of permanent copying, onto a floppy disk, is fixed enough to be a "reproduction." Micro-Sparc does not support the fact that a temporary copy into RAM is a reproduction. The court's use of both Bly and Micro-Sparc was inappropriate.
Two years later, in MAI Systems, Corp. v. Peak Computer, Inc., the Ninth Circuit stated clearly that copies made from a permanent storage device into a computer's RAM constitute reproduction.47 Here, the defendant provided maintenance service on the plaintiff's computers which had been sold to third parties. In servicing the computers, the defendant turned the computers on, loading the plaintiff's operating system onto the customer's computer for long enough to view an error log, and thereby determine what was wrong with the computer. The defendant in this case was the first defendant to argue that the copy made into RAM was not "fixed," and therefore not a reproduction. However, by 1994, the courts had assumed for eight years, without analysis or argument, that RAM embodiments are reproductions, and so the law was too settled.

The defendant's argument failed for two reasons. One, the court found that the defendant had not provided enough support for its argument to avoid a summary judgment.48 Two, the law had already coalesced, although haphazardly, enough to consider embodiment in RAM to be a reproduction. The court, acknowledging a lack of authority for its assertion that a copy made in RAM is fixed, nevertheless found support for its conclusion in three authorities.49

First, the court reasoned from Apple Computer, Inc. v. Formula International, Inc., focusing on that decision's analysis of the section 117 defense.50 In Apple, the court found that the defendant's copying of a program onto permanent, silicon chips was not an "essential step" in using the program, as required by section 117, because the defendant could have instead copied the program into RAM, without making a permanent copy.51 The Apple court reasoned, "[t]his would be only a temporary fixation. It is a property of RAM that when the computer is turned off, the copy of the programs recorded in RAM is lost."52 Focusing on this language, the MAI court construed Apple as authority that a copy into RAM is fixed.53 However, this analysis is faulty. Careful reading reveals that the Apple court used RAM as an example of a transient embodiment, not a fixed one, in contrast to a permanent embodiment fixed in a floppy disk.54

Next, MAI cites Vault as support for its conclusion. In Vault, as discussed above, the defendant did not offer the argument that embodiment in RAM is not fixed and the court provided no analysis on this point. The court instead granted the defendant the §117 defense.55

Finally, the MAI court cited the CONTU report as authority for its assertion that the embodiment in RAM is a reproduction. As discussed above, the CONTU report does not state this conclusion. Predating the wide use of hard drives, it fails to distinguish adequately between a computer's hard drive or floppy disks, which are surely fixed, and a computer's RAM, which is more transient. The Report's context reveals, however, that it probably referred to semi-permanent storage, not RAM.56
Recognizing that the authorities relied upon on do not clearly distinguish between a computer's hard drive and its RAM, the MAI court analyzed for the first time a RAM embodiment under the copyright law's reproduction provisions. Focusing on the first phrase of the Act's definition of "fixed," the MAI found that a RAM embodiment can be "perceived, reproduced, or otherwise communicated," but neglected to apply the limiting clause, "for a period of more than transitory duration."57

Less than one year later, the same party, MAI Systems Corp., was involved in a law suit with other service companies that serviced MAI's computers. The District Court for the Eastern District of Virginia, in Advanced Computer Services of Michigan, Inc. v. MAI Systems, Corp., followed MAI in finding that embodiments in RAM are reproductions.58 The defendant argued that RAM embodiments are too ephemeral and transitory to be fixed, but the court rejected the defendant's argument. It cited the aforementioned cases, each flawed in various ways, as authority for finding a RAM embodiment to be a reproduction. In addition, the court also mis-cited Apple Computer, Inc. v. Formula International, Inc. for a proposition that it did not make.59 While the Advanced court cited Apple for the proposition that "inputting a program into memory constitutes a copy," the Apple case actually says that embodiment in ROM, not RAM, is sufficiently fixed for copyrightability purposes, not reproduction purposes.60

The Advanced court was the first to address the "duration" clause. It found that embodiment in RAM for "minutes or longer" constitutes reproduction.61 The court did not assert any conclusions for embodiments lasting less than "minutes," and the court provided no basis or explanation for the line it drew. The court merely decided that "minutes" was enough. Had the court applied the purpose behind the fixation requirement, the line for fixation might have been different.

Just one month after Advanced, the District Court for the Northern District of California faced a case with facts similar to both MAI and Advanced. In Triad Systems Corp. v. Southeastern Express Co., the defendant serviced the plaintiffs computers, and in doing so loaded copyrighted software into RAM.62 The court did not cite Advanced, presumably unaware of the very recent case from the Eastern District of Virginia. Relying on MAI, Vault, and the CONTU report, the Triad court found that the embodiments in RAM were reproductions.63 The court rejected the defendant's argument that the RAM embodiments did not last long enough to be fixed ignoring, like the MAI court, the requirement that an embodiment be fixed "for a period of more than transitory duration."64 The court dismissed the duration clause saying, "the copyright law is not so much concerned with the temporal "duration" of a copy as it is with what that copy does, and what it is capable of doing while it exists. 'Transitory duration is a relative term that must be interpreted and applied in context. " The court held that regardless of how long a work is embodied in RAM in a particular computer system, "an ephemeral RAM copy of [plaintiff's] operating system software is the functional equivalent of a longer lasting copy in other computer systems. As a result, the kind of temporal distinction [defendant] is attempting to draw is not probative of the fixation question . . . . " In other words, the court dismissed the duration requirement of the copyright act as unimportant because any copy in RAM could potentially remain in RAM for any amount of time. In sum, the Triad court took a different approach than the Advanced court. The Triad court ignored the duration requirement while the Advanced court applied the requirement drawing a line for duration that is not based on the purpose of the requirement.

Seventeen years after the CONTU report, in 1993, the government formed the Information Infrastructure Task Force to "articulate and implement the Administration's vision for the National Information Infrastructure (NII).65 The Task Force, in turn, created the Working Group on Intellectual Property Rights, chaired by Assistant Secretary of Commerce and Commissioner of Patents and Trademarks Bruce A. Lehman, to "examine the intellectual property implications of the NII and make recommendations on any appropriate changes to U. S. intellectual property law and policy."66 The Working Group published the "White Paper" in 1995 which assesses current copyright law as applied to the internet, and recommends some changes for the copyright law's better application to the internet.
The White Paper purportedly relies on the Copyright Act, its legislative history, the CONTU report, and repeated court holdings for its assessment that loading a work into the RAM of a computer constitutes reproduction under the Copyright Act.67 However, neither the Copyright Act, its legislative history, nor the CONTU report say this. The Copyright Act requires that a copy must be fixed for a period of more than transitory duration, which does not necessarily include a RAM embodiment. The Act's legislative history is unclear in its treatment of RAM embodiments as reproductions. The CONTU report is ambiguous in its use of the term "computer memory. " Lastly, the numerous courts that have applied this rule have followed it blindly without thorough analysis.

The White Paper made another assessment about the current state of law regarding RAM embodiments which is unfounded. The Paper states that "when an end-user's computer is employed as a "dumb" terminal to access a file resident on another computer such as a BBS or Internet host, a copy of at least the portion viewed is made in the user's computer. Without such copying into the RAM or buffer of the user's computer, no screen display would be possible."68 A court, in early 1995, has already faced this situation and did not have enough evidence to reach such a conclusion. In NLFC, Inc. v. Devcom Mid-America, Inc., the defendant accessed and printed the plaintiff's copyrighted source code using dedicated telephone lines and a dumb terminal.69 The court was unable to determine whether the defendant's actions actually copied anything into RAM. It stated that the defendant did not "download the software onto its computers in order to work on it. Instead, [defendant] directly accessed the laboratory computers, via leased or dedicated phone lines and the use of a dumb terminal." In general, remote access using a dumb terminal does not embody data into RAM.70 The White Paper's assessment, therefore, has no basis in law or fact.

It is this line of cases and authority that has indoctrinated the RAM embodiment as a reproduction under the copyright law. These cases use inapplicable authority, confuse RAM with hard disks, ignore the duration clause of the fixation requirement, and build on each other's flawed holdings. By now, the doctrine might be too ingrained to be reversed. Therefore, critics have looked to the fair use, section 117, and copyright misuse defenses to preserve a person's right to load digital works into RAM. However, the decision that RAM embodiments are reproductions is flawed from the start.

A Proper Analysis of RAM Embodiments As Reproductions

Consideration of the purpose behind the requirement that reproductions be fixed shows that RAM embodiments should not be considered sufficiently fixed to be reproductions. The fixation requirement is meant to ensure that reproduction must be somehow harmful to the copyright holder to infringe the copyright holder's exclusive right to make reproductions. If a reproduction does not harm or compete with a copyright holder, then it does not deter creators from creating, and need not be prohibited. Meanwhile, the reproduction that does not harm the copyright holder may greatly benefit society by allowing people to use and enjoy other peoples' creations.

RAM embodiments do not threaten the benefits that copyright holders derive from their work. In order to create an embodiment in RAM, a user must have a permanent copy of the work to load into RAM. The copyright holder benefits from the sale of this permanent copy. Once the RAM embodiment is made, it cannot be easily transported, cannot survive without a constant flow of electricity, and has no market value. Although technically, one could load a copyrighted piece of software into the RAM of a portable, battery operated machine and distribute it to others, the software would be lost forever as soon as the battery runs out or is taken out of the machine to be replaced. The only way redistribution could be feasible is if the RAM embodiment is copied into a permanent form which itself would be a copyright infringement. In summary, the RAM embodiment only has one reasonable use - it is the only way that a digital file can be used or read. This use of a RAM embodiment does not harm, but actually helps, the copyright holder. If not for RAM, no one would be able to use the copyrighted work, and therefore, no one would purchase it.
When advancements in technology render the copyright law ambiguous, as it does with RAM embodiments, we are instructed by the Supreme Court to consult the basic purpose of the Copyright Act.71 When balancing the interest in providing incentive and reward to creators against the interest in public availability of creative works, the Supreme Court has found the public good to be the ultimate aim of Copyright law. In tune with the Supreme Court's instructions, the treatment of RAM embodiments under the reproduction provisions should favor the public's legitimate interest in being able to use the software that they purchase, even without a license allowing it.

The one way in which RAM embodiments can harm the copyright holder is one from which a copyright holder should not be protected. That is, the copyright holder's monopoly on servicing computers. Under the current law, which holds that RAM embodiments are reproductions, the holder of the copyright in software is the only party who can provide any service which requires loading up that software. Although section 117 was intended to prevent this result, software venders easily avoid section 117 by licensing, instead of selling, their software.72 Therefore, with RAM embodiments being reproductions, copyright holders effectively gain a monopoly in the service industry that should not accompany their copyright in the software.

Implications of the MAI Rule

Whether correct or not, the MAI line of cases has established that loading software into the RAM of a computer is sufficiently fixed to constitute reproduction under the Copyright Act. Since the RAM of a computer is essentially the same as the RAM in many other electronic devices, and is no less transient, the courts' holding has potentially broad implications. RAM, found in anything from a computer to a digital wrist watch, is nothing more than a series of registers that stores information temporarily by the flow of electricity. Any electronic device that has anything more than an on-off switch usually contains some form RAM. For example, a computer printer has RAM which usually embodies one page of a document at a time as it is printing. Similarly, a fax machine has RAM which embodies an image line-by-line as it transmits data over a phone line. A computer modem has RAM as well. Computer cards, such as video cards or microprocessor cards have RAM that is also no different than the RAM in a computer.

Electronic devices unrelated to computers also have RAM. For example, some of the features in a modern television require RAM. Picture-in-picture uses RAM to "remember" the coordinates at which the inner picture should be placed. RAM is also used in a TV's auto-replay feature. Modern VCR's have RAM which allows programming capability and time-keeping. When the electricity to a VCR is cut off, the VCR loses it's time and programming. The same goes for modern stereos which "remember," for example, to what station it was tuned last. Even a digital wrist watch has RAM, consisting perhaps of a single register, which keeps the time as long it has electricity.

The RAM in all these devices works the same way. Further, the different RAM does not vary in the degree to which it is transient. Therefore, although the courts have so far limited their holding to computer software embodied in the RAM of a computer, there is no way to distinguish this holding from RAM embodiments in other devices. For example, if a work is transmitted by FAX or modem, or spooled to a printer, an embodiment is made in RAM that is no less transient than that in a computer's RAM - even if the work is never saved into a file or printed onto paper. The RAM in a VCR or wrist watch is unlikely to ever embody a copyrighted work. However, the RAM used in the auto-replay feature on a television embodies a copyrighted work every time it is used.
Courts have already applied the MAI rule to different kinds of RAM. In DSC Communications Corp. v. DGI Technologies, Inc., the court, citing MAI, found that loading software into the RAM of a microprocessor card constitutes reproduction under the Copyright Act.73 This decision is necessary under the MAI cases because the RAM of a microprocessor card is just as fixed as the RAM in a computer. Similarly, in In re Independent Service Organizations Antitrust Litigation, the court applied MAI where the defendant loaded diagnostic software into the RAM of a Xerox machine.74 Early on, a court faced another kind of RAM embodiment. In Secure Services Technology, Inc. v. Time and Space Processing, Inc., the defendant placed a "protocol analyzer" between two of the plaintiff's fax machines in order to discover the handshake protocol necessary to produce a fax machine that could communicate with the plaintiff's fax machines.75 In doing so, the defendant embodied the handshake protocol, in the form of digital data, in the memory of the protocol analyzer. Since the court found that the handshake protocol was not copyrightable, it did not need to decide if the embodiment was sufficiently fixed to be a reproduction.76

The MAI holding has serious implications where the internet is concerned.77 Many internet transmissions involve embodying data in RAM without embodying it on hard disk. For example, real-time chat uses only the users' RAM.78 Moreover, transmissions, as they pass through the nodes of the internet to their destination, are saved temporarily in RAM.79 In addition, browsing the internet, depending on the configurations of the web browser, might involve only RAM embodiments. Under MAI, these transmissions are reproductions even though data is not copied into permanent form. The court in Religious Technology Center v. Netcom On-Line Communication Services, Inc., in dicta, confirmed this conclusion about internet browsing.80 It noted that "browsing technically causes an infringing copy of the digital information to be made in the screen memory [or screen RAM]. " Acknowledging that this result is the functional equivalent of creating an exclusive right to read, the court predicted that the fair use defense would probably apply. The White Paper re-states that MAI clearly applies to innumerable internet transactions.81 MAI's implications for the internet are somewhat alarming. They mean that much of what makes up the internet is potential copyright infringement. If this application persists, without an applicable defense such as fair use, the freedom of information that the internet offers will be thwarted.


After ten years, MAI's holding that digital embodiments in the RAM of a computer are reproductions, although misplaced and naive with respect to the crucial differences between RAM and hard disk copies, might be too indoctrinated in case precedent to be overturned. On the other hand, the MAI holding has been highly criticized by almost all the law review articles that discuss it.82 It seems intuitively incorrect that an evanescent embodiment in RAM is considered a potential infringement. The holding is not based on any clear legislative intent, and courts have never given the proposition thorough legal analysis based on copyright policy. The fixation requirement of reproduction should be interpreted with its purpose in mind. That is, a copy should not be an infringing reproduction unless it harms the copyright holder, and thereby deters creation. Additionally, courts have failed to recognize that the degree to which a RAM embodiment is transient depends on the particular system architecture and the type of file being embodied. Moreover, the holding has serious implications for the use of other kinds of non-computer RAM, and for the internet. As technology and internet usage increases, the MAI holding must be reevaluated or narrowed.


fnA. Boston College Law School, class of 1998.

1. MAI Sys. Corp. v. Peak Computer, Inc. , 991 F. 2d 511, 518-19 (9th Cir. 1993), cert. dismissed, Peak Computer v. MAI Sys. Corp. , 510 U. S. 1033 (1994); Triad Sys. Corp. v. Southeastern Express Co. , 1994 U. S. Dist. Lexis 5390, 12-18 (N. D. Cal. 1994); Advanced Computer Servs. of Michigan, Inc. v. MAI Sys. Corp. , 845 F. Supp. 356, 363-64 (E. D. Va. 1994).

2. 17 U. S. C. § 101 (1994).

3. Ira L. Brandriss, Writing In Frost On A Window Pane: E-Mail and Chatting on RAM and Copyright Fixation, 43 Journal Copyright Society U. S. A. 237 (1996); Niva Elkin-Koren, Copyright Law and Social Dialog on the Information Superhighway: The Case Against Copyright Liability of Bulletin Board Operators, 13 Cardozo Arts and Ent. L. J. 346 (1993); Katrine Levin, Intellectual Property Law: MAI v. Peak: Should Loading Operating System Software Into RAM Constitute Copyright Infringement?, 24 Golden Gate U. L. REV. 649 (1994); Jessica Litman, The Herbert Tenzer Memorial Conference: Copyright in the Twenty-First Century: The Exclusive Right To Read, 13 Cardozo Arts and Ent. L. J. 29 (1994); Bradley J. Nicholson, The Ghost In The Machine: MAI Systems Corp. v. Peak Inc. and the Problem of Copying in RAM, 10 High Tech. L. J. 147 (1995); Comment, Copyright Infringement Was Never This Easy: RAM Copies and Their Impact on the Scope of Copyright Protection for Computer Programs, 45 Catholic U. L. Rev. 181 (1995); Michael E. Johnson, Note, The Uncertain Future of Computer Software Users' Rights In the Aftermath of MAI Systems, 44 Duke L. J. 327 (1994); Carol G. Stovsky, Note, MAI Systems Corp. v. Peak Computer, Inc.: Using Copyright Law to Prohibit Unauthorized Use of Computer Software, 56 Ohio St. L. J. 593 (1995); Brian J. Murphy, Case Notes: Loading Software Into RAM Creates a "Copy," 10 Computer High Tech. L. J. 499 (1994); David Post, New Wine, Old Bottles: The Evanescent Copy, N. Y. L. J. , May 1995; Letter from James Boyle, Professor of Law at American University, to Bruce Lehman, Assistant Secretary Department of Commerce, (Apr. 19, 1996).

4. 17 U. S. C. § 106.

5. 17 U. S. C. § 101.

6. Id.

7. 17 U. S. C. §102.

8. 17 U. S. C. §101.

9. 17 U. S. C. § 102.

10. H. R. Rep. No. 1476, 94th Cong. , 2d Sess. 62 (1976); S. Rep. No. 473, 94th Cong. , 1st Sess. 57-58 (1975).

11. Final Report of the National Commission on the New Technological Uses of Copyrighted Works 55 (1978) [hereinafter CONTU report].

12. Craig Joyce ET AL. , Copyright Law 50 n. 3 (3d ed. 1994).

13. H. R. Rep. No. 1476 at 52-53; S. Rep. No. 473 at 104-05; Brandriss, supra note 3, at 242-45.

14. U. S. Const. art I, § 8, cl. 8.

15. CONTU report at 29 (Copyright should not grant anyone more economic power than is necessary to achieve the incentive to create).

16. The following explanation of computer memory is derived from the following sources: Thomas C. Bartee, Computer Architecture and System Design 244-315 (1991); Interviews with Danielle M. Berven, Electrical Engineer, in Boston M. A. (Fall 1996); Telephone Interview with George Berven, Mechanical Engineer, in Boston, M. A. (Dec. 2, 1996).

17. Religious Technology Ctr. v. Netcom On-Line Communications Servs. , 907 F. Supp. 1361, 1367-68 (N. D. Cal. 1995); CONTU report at 55.

18. 17 U. S. C. § 101.

19. CONTU report at 55 ("The text of the new copyright law makes it clear that the placement of a copyrighted work into a computer - or, in the jargon of the trade, the "inputting" of it - is the preparation of a copy. " This language, as well as the rest of the Report, does not make clear whether it refers to RAM or hard disk).

20. Religious, 907 F. Supp. at 1367-68 (court used MAI, a case that involved embodiment in RAM, as support for the fact that works embodied on hard disk were reproductions. Thereby, the court failed to recognize a distinction between RAM and hard disk).

21. Read Only Memory (ROM) is a chip containing in a permanent form the few basic instructions necessary to start up the computer.

22. Williams Elecs., Inc. v. Artic Int'l., Inc. , 685 F. 2d 870, 874 (3rd Cir. 1982).

23. H. R. Rep. No. 1476 at 62.

24. H. R. Rep. No. 1476 at 53.

25. No court that has faced RAM embodiments as reproductions under the Copyright Act has cited the Congressional Reports. The White Paper also does not mention them.

26. Public Law 93-573 (1974).

27. 2 Nimmer on Copyright § 8. 08[B][1] (1996).

28. CONTU report at 55 n. 131.

29. CONTU report at 55.

30. CONTU report at 56.

31. CONTU report at 29-30 (The Congress adopted CONTU's recommendation with only one change. It replaced the term "rightful possessor" with the term "owner. " The Congress did not provide any explanation for this perplexing change).

32. 17 U. S. C. § 117.

33. Johnson, supra note 3 at 334.

34. Bly v. Banbury Books, Inc. , 638 F. Supp. 983,984-85 (E. D. Pa. 1986).

35. Id.

36. Id.

37. Id.

38. Williams, 685 F. 2d at 874.

39. Micro-Sparc, Inc. v. Amtype Corp. , 592 F. Supp. 33, 34 (D. Mass. 1984).

40. Vault Corp. v. Quaid Software Ltd. , 847 F. 2d 255, 256 (5th Cir. 1988).

41. Id. at 261.

42. Id.

43. ISC-Bunker Ramo Corp. v. Altech, 765 F. Supp. 1310, 1332 (N. D. Ill. 1990).

44. 2 Nimmer on Copyright at § 8. 08 (1989).

45. Bly, 638 F. Supp. at 985.

46. Micro-Sparc, 592 F. Supp. at 34.

47. MAI, 991 F. 2d at 519.

48. Id. at 518.

49. Id. at 519.

50. Id. at 518-19.

51. Apple Computer, Inc. v. Formula Int'l. , Inc. , 594 F. Supp. 617, 622 (C. D. Cal. 1984).

52. Id.

53. MAI, 991 F. 2d at 518-19.

54. Apple, 594 F. Supp. at 622.

55. Vault, 847 F. 2d at 261.

56. CONTU report at 56.

57. MAI at 519.

58. Advanced, 845 F. Supp. at 363-64.

59. Id. at 364.

60. Apple Computer, Inc. v. Formula Intl. , Inc. , 562 F. Supp. 775, 779-80 (C. D. Cal. 1983), aff'd, 725 F. 2d 521 (9th Cir. 1984).

61. Advanced, 845 F. Supp. at 363.

62. Triad, 1994 U. S. Dist. LEXIS 5390, 1.

63. Id. at 12-17.

64. Id. at 16-17.

65. Intellectual Property and the National Information Infrastructure: The Report of the Working Group on Intellectual Property Rights 1 (1995) [hereinafter White Paper]

66. White Paper at 2.

67. White Paper at 65.

68. White Paper at 66.

69. NLFC, Inc. v. Devcom Mid-Am. , Inc. , 45 F. 3d 231, 236 (7th Cir. 1995).

70. Interviews with Danielle M. Berven, in Boston, MA (Fall 1996).

71. Sony Corp. of Am. v. Universal City Studios, Inc. , 464 U. S. 417, 431-32 (1984), reh'g denied, 465 U. S. 1112.

72. 17 U. S. C. § 117 (code language, in contrast to CONTU's recommendation, only gives "owners" of software §117 rights).

73. DSC Communications Corp. v. DGI Technologies, Inc. , 81 F. 3d 597, 599 (5th Cir. 1996).

74. In re Indep. Serv. Org. Antitrust Litig. , 910 F. Supp. 1537, 1541 (D. Kas. 1995).

75. Secure Servs. Technology, Inc. v. Time and Space Processing, Inc. , 722 F. Supp. 1354, 1358 (E. D. Va. 1989).

76. Id. at 1362.

77. White Paper at 64.

78. Brandriss, supra note 3 at 265.

79. White Paper at 66 n. 205.

80. Religious, 907 F. Supp. at 1378 n. 25.

81. White Paper at 64.

82. Supra note 3.

Front Page Comments © and Disclaimer About IPTF The Intellectual Property and Technology Forum