Friday, November 30, 2007

Installing Click-Wrap Software Is Copyright Infringement

If we subscribe to the view that software is licensed, not sold, then an interesting paradox arises surrounding "click-wrap" software. The end-user gets caught in a Catch-22 in which he or she must accept the EULA before using the software, but cannot do so without first committing an act of copyright infringement.

The problem is that click-wrap EULAs require the user to run the installation program in order to display and accept (or reject) the EULA. Pursuant to section 117 of the Copyright Act, only the owner of the copy of software is authorized under copyright law to use the software with a computer (using the software with a computer necessarily makes copies). This applies just as equally to the installer program.

Under this theory the end-user does not own the copy of the installer program, and is therefore not allowed to even run the installer. Doing so would be a clear act of copyright infringement. How then can the user accept the EULA and begin lawfully using the software?

The only way to avoid this Catch-22 would be to allow the end-user to read and accept or reject the EULA without having to first run any licensed software. This could be done in a couple of ways:
  1. The installer program could be sold outright to the user, so that the user is the owner of the copy of the installer and may therefore legally use the installer.
  2. The end-user could be required to accept the EULA before the software is even handed over to them.
The first option would require the installer program to reside on its own media, because the legal definition of a "copy" means the physical media on which the copy is recorded. The end-user can't own the installer and not own the other software if both the installer and other software are on the same disk. If they are on the same disk, then either the end-user owns both or the end-user owns neither.

The second option would make licensing software more like entering into normal, enforceable, contracts. For instance, the end-user would be presented with the license at the point of sale and would be required to accept it before taking possession of the software. The seller would be required to make acceptance of the software publisher's EULA a condition of the contract for sale. If the buyer refuses to accept the EULA, then the seller refuses to sell the license to the buyer.

Practically no retail software is ever sold in either of the above two ways. Accordingly, if the courts want to decide that software is licensed, not sold, then they must also come to the conclusion that virtually every person who has ever installed "click-wrap" software is a copyright infringer.

The other more sensible option, of course, is to come to the conclusion that the software is sold outright.

Monday, November 12, 2007

ProCD v. Zeidenberg

One of the more important decisions regarding software "license agreements" has been a 1996 Seventh Circuit Court of Appeals case known as ProCD v. Zeidenberg. It has become one of the most cited cases whenever the legitimacy of EULAs is questioned. On appeal, the court ruled that software EULAs are valid contracts, upholding the enforceability of their terms. It's an interesting case particularly because the court upheld the EULA even though the operating assumption was that the software was sold outright, not licensed.
The good news is, however, that there are easily identifiable problems with the court's opinion in this case. These problems should leave enough room for future cases to come to a different conclusion than ProCD.

First, a bit of background about the ProCD case. ProCD sold software at retail outlets. Matthew Zeidenberg was a customer who purchased a copy of ProCD's software from a retailer. Note that Zeidenberg did not purchase the software directly from ProCD. Later, Mr. Zeidenberg engaged in activity that was contrary to the terms of the "license agreement" and ProCD successfully sued Zeidenberg for his breach of the "license agreement".

It's important, when analyzing the court's opinion, to understand that the court treated the transaction like a normal sale of goods:
... we treat the licenses as ordinary contracts accompanying the sale of products, and therefore as governed by the common law of contracts and the Uniform Commercial Code. Whether there are legal differences between "contracts" and "licenses" (which may matter under the copyright doctrine of first sale) is a subject for another day.
This is an important distinction because this means that Article 2 of the Uniform Commercial Code (UCC) applied to the transaction in which Mr. Zeidenberg purchased his copy of the ProCD software. Indeed, the court relies heavily upon Article 2 of the UCC in formulating its opinion. The text of the opinion makes numerous references to various subsections of UCC Article 2, too many for me to quote all of them here. The question the court decided it needed to address, was whether or not the EULA in ProCD's software constituted a valid contract for sale between ProCD and Mr. Zeidenberg. The court found that it did, based on its interpretation of the UCC.

But this is a case where the legal definition of certain key terms was completely overlooked. It's similar to Apple's licensing blunder, where they've overlooked what the legal definition of what a "copy" is. In the ProCD case, the court overlooked the definition of "contract for sale", which is defined in Sec. 2-106(1):
(1) In this Article unless the context otherwise requires "contract" and "agreement" are limited to those relating to the present or future sale of goods. "Contract for sale" includes both a present sale of goods and a contract to sell goods at a future time. A "sale" consists in the passing of title from the seller to the buyer for a price (Section 2-401). A "present sale" means a sale which is accomplished by the making of the contract.
The key definition in this subsection is that of "sale", the passing of title from the seller to the buyer. This definition tells us that the "contract for sale" covers interaction between the buyer and seller and that the parties to the contract are the buyer and seller. The court in ProCD v. Zeidenberg never answered the question of who, exactly, were the parties to the contract for sale. Since ProCD was not involved in the actual sale, but instead a retailer was, the parties to the contract were in fact the retailer and Mr. Zeidenberg. However, the court repeatedly implied that ProCD was a party to the contract for purposes of interpreting applicable sections of the UCC. This is was an oversight that invalidates much of the court's opinion since they erroneously presumed ProCD into the contract for sale, and based much of their reasoning on this one assumption. All of the court's reasoning that involved the UCC was pretty much misapplied. Here is one such example:
What then does the current version of the UCC have to say? We think that the place to start is sec. 2-204(1): "A contract for sale of goods may be made in any manner sufficient to show agreement, including conduct by both parties which recognizes the existence of such a contract." A vendor, as master of the offer, may invite acceptance by conduct, and may propose limitations on the kind of conduct that constitutes acceptance. A buyer may accept by performing the acts the vendor proposes to treat as acceptance. And that is what happened. ProCD proposed a contract that a buyer would accept by using the soft- ware after having an opportunity to read the license at leisure. This Zeidenberg did. He had no choice, because the software splashed the license on the screen and would not let him proceed without indicating acceptance. [emphasis added]
Here the court infers ProCD into the position of the seller, when in fact the seller was the retailer. It was the retailer that made the offer to sell, not ProCD.

A proper analysis of the ProCD case, if you're going to assume that it was an outright sale as the court had done, requires us to view the contract for sale and the EULA as two separate contracts: one between the retailer and Mr. Zeidenberg, and the other between ProCD and Mr. Zeidenberg. The EULA does not automatically become part of the contract for sale just because there is a label on the box indicating that there is an EULA inside. The language was not added by the retailer and was not a term that was agreed upon as part of the sale.

Interestingly, the court does partly agree that the EULA is separate from the contract. Contradicting it's earlier argument that the EULA became part of the contract for sale, the court begins to treat the EULA as part of the product itself, not as part of the contract:
In the end, the terms of the license are con- ceptually identical to the contents of the package.
and
Terms of use are no less a part of "the product" than are the size of the database and the speed with which the software compiles listings.
The court uses this argument, that the EULA is actually a part of the product, as grounds to argue that Mr. Zeidenberg had the opportunity to inspect and reject the EULA pursuant to UCC Sec. 2-602(1), which gives the buyer the right to inspect and reject the goods. Arguing this, that the EULA is part of the product, puts Mr. Zeidenberg in a highly unusual position. This is because EULAs universally assert that ownership of the software remains with the software publisher. An essential result of the court's argument is that Mr. Zeidenberg was buying these goods only so that he could waive his ownership. This whole notion, that by purchasing goods a consumer is simultaneously waiving their ownership rights is absurd. No reasonable consumer would ever agree to such nonsense. I can hardly dream up a more unconscionable term in a contract for sale. After all, the purpose of purchasing goods is to obtain ownership in them.

Nevertheless, this situation where, by purchasing a product, the consumer must give up his ownership in the product, is an inescapable consequence of the court's reasoning. This is exactly why normally software publishers insist that their software is licensed, not sold. By selling permission to use software, publishers avoid this situation where the consumer is purchasing something meaningless and absurd. Instead of buying an EULA for something already owned, software publishers say, consumers are buying permission to use something they never owned in the first place.

In summary, the court made at least two errors which, if corrected, would completely alter the course of their reasoning:
  1. ProCD was not a party to the contract for sale, the retailer and Mr. Zeidenberg were the parties to the contract. As such, the EULA must be treated as a separate contract.
  2. The EULA is clearly not a part of the goods which must either be accepted or rejected. It would be unconscionable for a consumer to buy something for which they must give up all their ownership rights in order to use.
Addressing these two problems, we quickly find that Mr. Zeidenberg fulfilled all his obligations under the contract for sale and was the rightful owner of the software the moment money changed hands. The software publisher's attempt to force Mr. Zeidenberg to agree to the EULA before being able to use a product that he owns, is an illegal power-grab. The right to regulate use of a copyrighted work is not one of the exclusive rights granted to copyright holders. ProCD had absolutely no right to dictate how Mr. Zeidenberg used the software that he was lawfully entitled to use.

Wednesday, November 7, 2007

Apple's Attempt to be Clever Backfires

I recently had the opportunity to read over one of Apple's alleged "software license agreements". In an apparent attempt to work around the many problems of claiming to license software, when in fact they are selling it, Apple has added the following line to their "licenses":
You own the media on which the Apple Software is recorded but Apple and/or Apple's licensor(s) retain ownership of the Apple Software itself.
This is the first time I've ever read an Apple "license agreement", so I don't know how long they've been using this line. But I bet there are plenty of software pirates out there that would just love to use that line as a defense. "I didn't copy the software, your Honor, I just copied the media that it was recorded on."

Seriously though, one might interpret their line to mean they own the "copyright" in the "Apple Software itself", but I do not think this is what they mean. Copyright law already firmly addresses that issue so there is no need for Apple to reiterate that.

Apple must mean what they appear to be saying, that a copy of software can be separated from the media on which it is recorded for the purposes of transacting a sale. I can only guess that this is an attempt to find a solution to the problems with the tired old "licensed, not sold" blurb. After all, SoftMan so eloquently illustrated how obvious it is that software publishers are in fact selling something. With nowhere to hide, Apple has apparently decided to try to change what it is they are selling. The "logic" behind this change is that the end-user still doesn't own a copy of the software, because Apple owns the software. The end-user only owns a round plastic disc.

Apple's lawyers obviously haven't been doing their homework. What they seem to fail to understand is that the physical media is, by legal definition, a copy of the software:
"Copies" are material objects, other than phonorecords, 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.
That is the word of 17 USC Sec. 101 so you can't really argue with that definition. In the case of Apple software, that "material object" would be the disk, CD, DVD or whatever media it happens to be recorded on.

Here's where Apple's clever little plan backfires. By admitting that they are selling the media right there in their "license agreement", they are admitting that Apple end-users own copies of the software. For purposes of interpreting section 117, the end-user is the owner of the copy. This defeats the whole purpose of claiming that Apple is licensing the software, not selling it. I can't imagine a bigger blunder and if I were Steve Jobs I would immediately can whichever lawyer was responsible for adding that term to the "license agreement". Apple would be much better off if they removed that term ASAP. At least then they could still try to argue that they are only licensing software. As it stands right now, they have outright admitted that they are selling it.

Monday, November 5, 2007

Adobe v. One Stop Micro

As I mentioned in the preceding post, the courts have not yet arrived at an agreement as to whether software is sold or licensed. We've already looked at SoftMan v. Adobe, where the court found that Adobe was selling their software. On the other side of the fence is Adobe v. One Stop Micro. In this case the court ruled that Adobe was licensing their software.

The line of reasoning used in the One Stop Micro case isn't very persuasive, especially when viewed in light of the more recent ruling in SoftMan. Let's look at how the judge in One Stop Micro came to the conclusion that Adobe was "licensing" their software, not selling it.

One Stop Micro was selling "Educational Versions" of Adobe software, and Adobe claimed that One Stop was violating Adobe's reseller license agreement, the OCRA. Adobe argued that the OCRA specifically made the transaction a license, not a sale. One Stop Micro, of course, argued that the OCRA was actually a contract for sale. One Stop Micro pointed out that the OCRA itself used terminology such as "purchase" and "own". But the judge observed that the OCRA also used other language indicating that it established a license. Since the OCRA itself seemed contradictory and ambiguous, the judge deemed that it would be necessary to find other evidence by which to determine whether the OCRA constituted a license or a contract for sale.

One piece of supposed "evidence" that supported the position that the software was licensed, was Adobe's inclusion of an EULA with the software. The judge wrote:
"It would be incongruous to conclude that educational resellers are owners of the Adobe educational versions, while the end users who the resellers distribute to are granted a mere license."
How the judge jumped to the conclusion that under the "sale" theory, Adobe sells a copy of the software to the reseller, but the reseller must then sell a license to the end user is unclear. And inexplicable. Clearly, if the reseller "owns" the copy they purchase, then when the reseller sells that copy to the end-user, the end-user would now "own" the copy for the same reasons that the reseller owned the copy previously. Adobe's claim to be licensing to the end-user via the EULA is by no means evidence that supports their other claim that they are licensing to resellers via the OCRA. That's just one dubious claim supporting another virtually identical dubious claim. Nevertheless, the judge found this reasoning to be, and I quote, "particularly compelling".

Next, the judge relies on a statement from Adobe's Vice President and General Manager for North America:
"Adobe's intent in drafting and signing these documents to create a license, rather than a sale of our software." (sic)
He uses similar statements from others, including at least two other Adobe resellers, and one or two other experts who all testify to the effect of, "It's standard practice in the industry to license software instead of selling it. Everybody does it this way." The judge presents these statements as evidence that supports Adobe's position. But there is never any evidence shown as to how these transactions within the industry at large are indeed licenses and not sales of goods.

By this line of reasoning, used car salesmen could sell their cars claiming them to be "new" so long as every other used car salesmen does the same. Obviously this would never happen, because the fact is the cars are used. What the salesmen claim doesn't change reality, even if they all collectively claim the same thing. Just because every software publisher in the industry claims they license their software, it doesn't make their claim any more true than the used car salesmens' claim.

The judge's biggest error here is that he fails to look at the economic realities of the exchange in question. His line of reasoning is so weak, and neglects to address the most important questions (i.e. does the transaction fit the definition of a lease or a sale under the Uniform Commercial Code), that it's no wonder the judge in SoftMan v. Adobe wrote this about it:
"To the extent that the court in One Stop found that the transaction at issue was in fact a license, and not a sale, this Court simply declines to adopt that analysis."
I think that's the legal way of politely saying, "What the hell were they thinking?"

SoftMan v. Adobe

Is there any case law that supports my theory that software is actually sold, not licensed? Yes, there is, although the courts are apparently not necessarily in complete agreement. Let's first present a case that rules in favor of "sold, not licensed". In a future post, we'll look at a conflicting case and point out some flaws in their reasoning.

In 2001, the case SoftMan Products Co. v. Adobe Systems Inc. went before a U.S. District Court in California. In this case, one of the arguments Adobe used to try to support their position was that they do not sell their software, but instead merely "license" it. The judge dismissed this argument stating, "It is well-settled that in determining whether a transaction is a sale, a lease, or a license, courts look to the economic realities of the exchange." The judge then pointed out many ways in which the economic realities plainly indicated that Adobe's software is sold, not licensed:
  • The purchaser obtains a single copy at the time of the transaction and for a single price
  • The transfer of possession is indefinite with no terms for renewal
  • Distributors and end-users pay full value for the software and accept the risk that copies in their possession may be damaged or destroyed.
The judge also included a quote from an expert on the subject:
The pertinent issue is whether, as in a lease, the user may be required to return the copy to the vendor after the expiration of a particular period. If not, the transaction conveyed not only possession, but also transferred ownership of the copy.
Clearly this is at odds with the view that software is licensed. As such, the judge ruled that Adobe was selling their software, not licensing it. Most other commercial software sold at retail is still sold in exactly the same way that Adobe was selling their software in 2001. Yet more evidence that most software is sold, not licensed.

Thursday, November 1, 2007

Why "Sold, Not Licensed" Matters

Some people may be wondering, "What's the big deal? Why does it matter if software is licensed instead of sold?" The main reason it makes a difference is because of some of the technical details of copyright law. In order to use a computer program, you necessarily must make one or more copies of the software. Some examples: copies are made on the hard drive, copies are made in RAM, and copies may be temporarily made in caches and other places. Normally, copyright law would supposedly prohibit you from making these additional copies. Only the copyright owner, in this case the software publisher, is allowed to make copies.

At some point, Congress decided that copyright law needed to make an explicit exception for computer software. So section 117 was added to U.S. copyright law. This section recognizes the fact that computers can't make use of software without making these incidental copies. As such, the section permits these copies to be made as long as the copies are necessary in order for the software to work on a computer.

But there's a catch. According to section 117, only the "owner" of a particular copy of software is permitted to make these incidental copies. In other words only the owner of a copy of software is allowed to actually use the software on a computer. This is where "licensed, not sold" comes from.

If we subscribe to the view that software is licensed, not sold, then the software publisher is still the owner of any copies of the software. End-users do not own any copies, even those copies that the end-users may have in their possession. So, under copyright law -- even with section 117 -- end-users have no right to actually use copies of software on computers. Instead, the software publisher gives end-users permission to use the software and to make certain copies. The EULA is the agreement that grants this permission. In order to accept the permission given to use the software, the end-user must accept all other terms laid out in the EULA, many of which typically restrict the users' rights far beyond what copyright law would otherwise allow.

If, on the other hand, we subscribe to the view that software is sold, not licensed, then the end-user takes ownership of any copies of software that they buy. As such, end-users are then free to use the software without the publisher's permission. Section 117 allows the end-user, as the owner of the copy, to make any copies necessary in order to allow the software to be used on a computer. Hence the end-user does not need permission from the software publisher to use the software and there is no need for the end-user to agree to an EULA. When software is sold, end-users are afforded all the rights that copyright law provides.

This brings me to one of the most bothersome consequences of the practice of "licensing" software instead of selling it. The use of EULAs has become so widespread, so ubiquitous, that virtually all commercial computer software produced today is supposedly bound by an EULA. If we accept the assertion that EULAs are valid and that the software is indeed licensed and not sold, then section 117 of U.S. copyright law has been rendered meaningless. There would be virtually not a single copy of software that would be affected by section 117 because virtually all copies of software would be owned by their respective copyright owners. Section 117 only has meaning when there are copies of software that are owned by people other than the copyright owners. In other words, software publishers are attempting to unilaterally reverse a law that Congress created for the express purpose of allowing end-users to use software created by others. And that just makes me angry. It should make the courts and Congress angry too.

The Sales Tax Problem

As I mentioned in this blog's inaugural post, there are problems with the line in EULAs that states "this software is licensed, not sold." The first such problem that we'll explore is that of sales tax. Let's examine a concrete example.

I live in the state of Massachusetts. This state has instituted a law that requires all retail sales of tangible property or services to be taxed at a rate of 5%. Not surprisingly, any time I buy retail boxed software from CompUSA (or any other software retailer in the state) I pay that 5% sales tax.

Why am I paying that sales tax? First, we need to decide if I'm buying tangible property or if I'm buying a service. It doesn't seem to fit the law's definition of a service: "a commodity consisting of activities engaged in by a person for another person for a consideration." I'm not paying anybody to engage in an activity for me. I'm paying to take possession of a box with some stuff inside it. Sounds more like tangible property to me.

So what this all means is that, because I'm paying the 5% sales tax, both the State of Massachusetts and CompUSA must believe that I am buying tangible property. That tangible property can only be the box and whatever is inside it. When I look inside retail boxes of software that I've bought, I usually find these two things:
  1. A copy of the software (usually on CD or DVD).
  2. One or more manuals and various other literature.
I have never purchased a retail box of software and found a tangible "license" inside of it (this would actually be impossible, I'll touch on that later). So apparently, I'm buying all of these physical things including a copy of the software. The State of Massachusetts and CompUSA both seem to be in agreement on this point, hence the 5% sales tax. This is in clear conflict with what EULAs usually say, that the software is "not sold".

If the EULA is correct then I'm not actually buying a copy of the software. In fact, according to the EULA, the only thing I "own" is the license. All the rest of the stuff in the box is property owned by the software publisher, and they are just allowing me to use it. Since I'm not buying any of it, I shouldn't be paying sales tax on it.

Software publishers can't simply throw a piece of paper in the box with the word "License" scrawled across it to work around this problem. Why? Because that's not what a license is. By definition, a license is intangible. It is an agreement, or more accurately, it is permission to do something. A piece of paper that says "This is a license to do X" is merely evidence of the license, not the actual license itself. This is why it would be impossible for a publisher to include a license "in the box".

Does this observation, that sales tax is applied to retail boxed software, absolutely prove that retail boxed software is usually sold, not licensed? Probably not. But it's just one of many other pieces of evidence indicating that, despite what EULAs say, software is in fact usually sold outright.