Via Michael Froomkin and James Grimmelmann, I learn of the lawsuit filed by Thomson/Reuters, proprietors of the closed-source EndNote software used by scholars to manage sources, against George Mason University, home of Prof. Dan Cohen, leader of the project to develop Zotero, an open source Firefox plugin that competes against EndNote. (Prof. Cohen also publishes the great Digital Humanities blog.) The complaint alleges that GMU’s EndNote site license prohibits reverse engineering the software, and that Zotero was developed in part by reverse engineering EndNote to determine how Zotero could convert EndNote’s output styles into Zotero’s output styles. This allegedly amounts to a breach of the site license by GMU, which is a breach of contract. Thomson/Reuters seeks an injunction, not the standard contract remedy, against further distribution by GMU of the illegally-developed Zotero product. (Oops — I almost typed “infringing” Zotero product!)
Grimmelmann’s parsing of the complaint leads him to the conclusion that it’s the work of “a comically inept overbearing bully.” Bullying and overbearing it may be. Not only does the complaint seek an injunction against further distribution of Zotero, but it also asks for an injunction that “terminate[s] the ability of each Zotero user to use and/or further distribute any .csl style files that were converted from the EndNote Software’s proprietary .ens style files using Zotero.” I don’t use the software, so I’m speculating a little bit: It looks like the timing of the complaint, and this extraordinary demand to kill Zotero functionality for Zotero users, is related to the imminent release of a Zotero update that would create a shared Zotero commons. Zotero users could easy share their source data with other Zotero users. The complaint challenges the heart of the scholarly research enterprise, the premise that knowledge should be available to all and shareable by all. Thomson/Reuters would like to say, apparently, that if you put that knowledge into EndNote, then Thomson/Reuters is your gatekeeper. That’s shameful.
But I’m not ready to agree that it’s comic or inept. It’s not an accident, I think, that Thomson/Reuters has carefully pleaded only state law claims based on the language of the site license. If it can persuade the trial court that it should follow the Federal Circuit’s lead in Bowers. v. Baystate Technologies, holding that the Copyright Act does not preempt breach of contract claims involving “no reverse engineering” clauses in software license agreements, then the plaintiff may have a path to victory here. The contract may be valid, even if it’s obtuse. Of course, Thomson/Reuters would still have to prove that GMU (i.e., Prof. Cohen and his team) did the things (reverse engineer and so on) that the complaint alleges they did.
On policy grounds, as in Bowers, that outcome would be comic (or worse), to borrow James’s phrase. Thomson/Reuters is transparent in its effort to use a software license to suppress a competitor in a product market and to interfere with the production and distribution of knowledge. This is a result that it almost certainly could not achieve if it filed a copyright infringement claim in federal court, alleging the same facts. The styles at issue might be non-copyrightable or the subject of only a “thin” copyright that protects only against literal copying. Any alleged reverse engineering might well be treated as fair use. That the complaint actually seeks an injunction confirms that this is a thinly-disguised infringement lawsuit. If it were a real infringement lawsuit, it would and should be a loser.
If the case isn’t settled quickly, expect GMU to remove the case to federal court, and expect the preemption issue to be litigated via a motion to remand. It would be nice, too, if we saw intervention in the lawsuit by a different university or consortium committed to open access and open source principles in higher education, participating on behalf of Zotero-using faculty who object to Thomson/West’s proposed interference with their work.
The excessive damages for a contract complaint (one that admits the harm is irreparable!) is one issue, but there’s a couple of other problems too, one of which is: what does the prohibition on “reverse engineering” prohibit?
Since this is contract law, the question is what the parties intended — and since this is a form contract nobody actually read (because no one ever reads form contracts), that boils down to what the reasonable person would interpret “reverse engineering” to mean. Does “reverse engineering” include just looking at unencrypted EndNote output formats? As far as I can tell, that’s all that’s at issue here. I wouldn’t think that was RE — by that definition I’ve “reverse engineered” various news readers, the time entry program at my firm, some Palm software, and certain videogame configuration files — and I hardly consider myself a hacker.
This definition is sort of like saying you “reverse engineer” your car if you look at the part that’s broken and go into an auto parts store and guesstimate which replacement part you need. Maybe that’s a possible definition, but I don’t think it’s the standard definition.
I agree, but I highlighted reverse engineering in the post because Thomson/Reuters does so in the complaint. The full license term alleged in the complaint is a broader prohibition on “[r]everse engineering, de-compiling, translation, modification, distribution, broadcasting, dissemination, or creation of derivative works from the [EndNote] Software.”
Are Zotero output styles “translations” or “derivative” of EndNote output styles? In light of the “reasonable person” standard for interpretation that likely applies here, is it appropriate to apply the broad Copyright Act definition of “derivative work” to the license? Is that what both parties intended and understood?
Hmm … seems something went wrong with my comment. Let’s try again …
No, they are not.
While Zotero could in theory export Endnote-derived CSL files, the team made the decision not to include this code so as to avoid any potential for just these sort of problems. So the core of this suit is based on incorrect (and, BTW, easily verifiable) information.
I have to disagree with part of your conclusion: I don’t see this being removed to federal court. GMU is a state school and “arm of the state” of Virginia. This goes back to copyright and trademark, with the whole State Street Bank nonsense. I still have trouble understanding how complete preemption does not adequately state abrogation of immunity, despite all of the circular Eleventh Amendment arguments.
If I were a complete cynic, I’d also assume that the plaintiff wants to be in Virginia state court rather than the Eastern District of Virginia because the state-court judges are less likely to delve into the nuanced differences between a contract to license and a contract to sell, or to understand the difference between “reverse engineering” and “data compatibility.” <sarcasm> Far be it from me, however, to be cynical about the litigation strategy of the company that gave us West v. Mead and Hyperlaw. </sarcasm>
The connection is even more tenuous than that. The question is whether the Zotero output styles are “translations” or “derivative works” of the EndNote software. I.e., the whole thing.
Having used both applications, but never creating the output format for Zotero, I am not sure about this creation process, but… the strength of EndNote at present is their ability to deliver thousands of different output styles. Given the fact that APA style is a APA style, it might be difficult to prove that Zotero took the shortcut of looking at the output style of EndNote, i.e. the apa.ens file and converted it to the Zotero output style, but …. from my experience of modifying an EndNote *.ens file — it isn’t easy to create one and apply all of the rules necessary to get the output into the correct style. So I am sure it took lots of man hours to create these *.ens files. Don’t get me wrong, I love Zotero, so if someone takes the time to create all of these output files from scratch then more power to them, but I can also see ISI’s point if Zotero is taking their labor to shortcut the process in creating these output styles.
It all falls down to whether the Zotero cls files were made by someone with the APA Publications manual or utilizing the EndNote *.ens file, unless I missed the point and there is something technical in the coding which the programs use to create or read these template files which is at the heart of the matter.
INAL, but I don’t think it’s that direct even. I think the argument is either a) the Zotero styling system is reverse-engineered from Endnote’s, or b) all Endnote style files are a part of the Endnote product (and hence covered by the license).
The first is wrong, and the second so ridiculous as to be insulting.
@H. S. McMinn
One might be able to look at things such as:
Does Zotero contain a style for every EndNote style? No.
Do EndNote and Zotero make the same mistakes or choose the same quirks in citations? Probably not.
Is there some proof of who developed styles & changes made to those styles? Yes–Zotero has a public version control system. There are discussions in the forums and points in their issue tracker corresponding to some wrinkles with certain styles.
This is not happening. CSL is a better technology for citations. It is textual, which means that people can easily tweak a pre-existing style to suit the needs of a closely related style. There is an online style generator.
I think there is some confusion about the facts here. Consulting the PDF linked in Mike Madison’s original post at the top will reveal that Zotero does *not* include thousands of output style files converted from EndNote.
It only includes about 15 styles.
What it *does* include, according to the complaint, is the facility to *convert* EndNote output style files into Zotero output style files. It is this conversion facility that is at issue in this suit.
@Robin:
The suit alleges that Zotero:
and that they are:
Zotero includes thoustands of styles.
Finally, it does not have the ability to convert EndNote styles to Zotero styles–it could only use EndNote styles in a read-only fashion (though the suit alleges otherwise).
I’ve been following this law suit, and am just as outraged many of the readers here.
It’s ironic that just today, I recevied an email from Thomson Reuters requesting that I participate in market research.
I wrote up a post about it urging academic who receive this email to NOT PARTICIPATE.
http://www.andrewcullison.com/2008/10/thomson-reuters-market-research/
If you receive the same email I did, maybe you could voice your concerns about the law suit.
From Mike’s post:
After perusing the complaint (for a post over on the Marquette Faculty blog), I have a different theory. It looks to me like the complaint was timed to block the addition of a new functionality to Zotero: the ability to convert .ens files to .csl files. (See paras. 20-21.) Maybe the sharing is also part of it, but the complaint is pretty explicit about the “harm,” and it didn’t mention the sharing of .csl files, only the translation ability, unless I missed it.
@Bruce Boyden
Sharing is mentioned explicitly several times:
P2 is the first section that I quoted, above.
P15 has “distribution or transfer of the [EndNote] Software” emphasized, implying that such distribution occurred.
P23 is the second section that I quoted, above.
P36 again discusses a breach by “distributing and transferring the EndNote Software’s proprietary .ens style files.”
Relief C seeks that Zotero “terminate the ability to …. further distribute any .csl style files that were converted from the EndNote Software’s proprietary .ens style files to Zotero’s .csl style files.”
Also: many paragraphs that complain about the conversion mention that CSL files are “freely distributable.” This is true–currently, there is no EULA for Zotero or various CSL generators that attempt to place a license on files created by end users. It is misleading, though: Zotero had essentially let people use ENS files as-is in a read-only fashion, without generating stand alone CSL files that people could distribute.
Comments are closed.