Infosys 204
Assignment #4 proved to be a challenge for nearly everyone, and I’m not sure if that is because I reached a little too far in my expectations or that you just didn’t have enough time (or energy) to really give it the attention it required. Still, there were several truly excellent papers, which tell me at least that the job was doable.
Some of you are still failing to pay attention to punctuation issues I noted in Memo #1. Rest assured I took that into account in grading.
I suppose the one persistent problem in this particular assignment is the proper citation format for cases. Although there are many acceptable citation forms, a few features are common. The general rule is that when you mention a case name in the text, you should underline or italicize the name, and you should always abbreviate (even the first time) to the last names of the lead parties. So, a first text reference to the Seventh Circuit opinion would be: “In Corcoran v. Sullivan, the court held….” Subsequent text references could simply name the lead party, e.g., “But Corcoran holds to the contrary.” Again, however, you need always to underline the case name. If two cases have the same lead party (as is the case with an underlying decision and its appeal), you can simply note the first will be Corcoran I and the second will be referred to as Corcoran II.
Whether you are giving the detailed citation in parenthesis or in footnotes/endnotes, your initial citation must, at a minimum, give the case name, the reporter, the court, the year, and the page number. A general reference to Corcoran would be: Corcoran v. Sullivan, 112 F.3d 836 (7th Cir. 1997). If the first reference is to a specific part of the case, you give both the initial and specific page number as follows: “The Seventh Circuit rejected all but Corcoran’s copyright argument as having too little merit to warrant discussion.” (Corcoran v. Sullivan, 112 F.3d 836, 839 (7th Cir. 1997)). Subsequent references can simply give the “pinpoint” citation, e.g., Corcoran, 112 F.3d at 837. You can also use Ibid. and Op. Cit. if you prefer.
As I mentioned before, you only need to give reference to the main reporter, not to all of the reporters that printed the case. Once a case is reported, moreover, cite to the reporter not to the internal court numbering system or a LEXIS/Westlaw interim page number. So for federal cases, always go with the Federal Reporter series; for the Supreme Court, the principal reporter is U.S. Reports (U.S.), but these are usually a year behind the Supreme Court Reports (S.Ct.). For the Supreme Court only, you need not name the Court in your citation, e.g., Feist v. Rural Telephone Service, 499 U.S. 340 (1991).
There were several objectives I had with this assignment. I wanted to see if you could do a close read of a new case without discussing it in class, and integrate the opinions with the statute and identify where they diverged. Next, I wanted you to apply materials from the readings and class discussion to flesh out some of what was missing, both in the facts and the legal analysis, in the two court opinions. I also wanted you to apply the readings to a discussion of what Corcoran could have done differently had he read and understood what you read and understood. Finally, I asked for some comments on the institutional limits revealed in the assignment.
The key here was to apply the material from the class to the Corcoran case. The best papers were specific about the facts of Corcoran’s problem, the materials from the reader that applied, and how the two intersected. (Proper citation of the referenced materials was also a good idea.) Just mentioning cases that you thought were relevant was not enough; you had to say how the case applied (or didn’t). In law school, we teach students to approach exam problems using the “IRAC” process – Identify an issue, cite a Rule from a case, Apply the rule to the particular facts given, and then Conclude. I was looking for something like that throughout this paper. Some papers took the material as an invitation to write very general papers about software developers and their rights and obligations, without making reference to Corcoran’s case or material from the reader. That was not what I was looking for.
Some of you also fell into Corcoran’s own trap of putting all of your eggs in one basket—that is, arguing exclusively about copyright. Copyright is certainly an important element, but it wasn’t the only one, and the best papers disposed of that quickly and moved on to other issues that Corcoran didn’t—the meaning of “destroy,” for example, or the determination of “without authorization.”
The Facts
There is quite a bit we don’t know, of course, and most of you noted that Judge Posner in particular seemed to misunderstand much of what is known, whether or not those misunderstandings or omissions make a difference.[1] But here are a few things we do know:
1. Corcoran was not an employee of MCS--he was a contractor. There is no question that Corcoran owns the copyright to his own software, because he is not an employee (see Wisconsin v. Corcoran, 522 N.W.2d 226, 231 (Wis. App. 1994)) and there was no written agreement prior to his beginning work that his programs would be considered “work for hire.” These are the minimum requirements for MCS to claim copyright in the programs, which is why the state never argues to the contrary.
2. There was no written contract, or if there was, we know nothing of its contents. You could have certainly assumed the contract was oral, and you would have been correct.
3. Corcoran served as his own attorney throughout the case. Both case headings note that Corcoran represented himself “pro se.” Remember that when we read the Aymes case I told you to look up the phrase “pro se” if you didn’t know what it meant.
4. Corcoran was writing macros for Lotus 1-2-3, not programming from scratch or creating a custom database. (See Corcoran, 522 N.W. 2d at 230 n. 10.) The data from the restaurant reviews were being entered into Lotus spreadsheets that Corcoran set up; he also programmed macros to process the data and to produce the reports, all within the structure of Lotus 1-2-3.
Beyond these certainties (which some of you fumbled), there is a lot of doubt as to what did or did not happen. What I wanted you to focus on, however, were the facts (missing or otherwise) that mattered. Did he willfully destroy data that was not his property? Did he have authorization to do so? Understanding the four facts noted above was central to avoiding a lot of wasted analysis. For example, those who said he was MCS’s employee had no basis to argue that anything was his property. And knowing something about Lotus 1-2-3 and its file structure (only one person went down this path) might have helped to sort out whether it was even possible for him to delete his own work without necessarily deleting data MCS entered into the spreadsheets he designed.
Just to clear up some confusion, I suspect what really happened here is that MCS hired a relatively unsophisticated contract programmer to set up the spreadsheets and reports, and Corcoran was unable to manage the project professionally. He told them the spreadsheet forms and macros were ready, and they began to enter data (remember that time was of the essence) while he continued working on the reports, which never came out quite right, causing him to go back to the spreadsheets and macros themselves, which may have required re-entry of information. His communication skills were poor, and MCS, itself an unsophisticated buyer of programming services (remember too this was 1987—pre Internet, for starters), became frustrated and began to feel captive to someone who kept telling them it was OK, he was nearly there, and anyway that they weren’t to pay him until they were completely satisfied.
The frustration on both sides increased and eventually the relationship melted down. Corcoran took matters into his own hands and then, convinced of the righteousness of his cause, compounded error on error by reading the copyright statute literally and holding to that reading (and nothing else) even when there were other ways to avoid conviction.[2]
Legal Analysis
Perhaps the best way to analyze Corcoran’s case and the materials we covered in class would have been to organize your argument around the key elements of the WCCA that he was found to violate: (1) willfully and knowingly, (2) destroys data and (3) without authorization. (You could have also argued—as Corcoran should have—that the amount of damage was less than $2500, but Corcoran has technically waived this argument long before we get to the Supreme Court.)
It seems nearly certain, by the way, that Corcoran’s insistence on holding copyright to MCS’s data is a non-starter. For one thing, as the opinions note, his reasoning would mean that any application software developer would own copyright to every copyrightable output created using that software (apply his description of how his programs “created” the data files to any piece of software you own). In other words, Microsoft owns the copyright to every document created in Word, every spreadsheet in Excel, every PowerPoint presentation. Indeed, following Corcoran’s logic, Lotus (now IBM) owns the copyright to his work, and probably to the MCS data as well.
That can’t be right, and it isn’t. It doesn’t matter if the data gets any copyright in the first place (if the “reviews” are more than just facts), if the compilation gets copyright (the structure of his spreadsheets), if MCS participated in the design of the inputs or outputs—regardless, Corcoran does not own what he deleted, that is, the files with the data (again knowing the Lotus file structure would be informative here). And it doesn’t matter that MCS doesn’t have copyright in them either—just because the facts aren’t protected by copyright does not mean the files are not protected by property laws. The WCCA makes it illegal to destroy someone else’s data, and that is what he intended to do.
The best starting point would have been to step back and ask why the Wisconsin legislature passed the WCCA in the first place. I told you that there was no published legislative history for the act (there often isn’t for state laws), but it should have been pretty clear that the WCCA was passed as an anti-hacker statute; that is, that its chief concern was with (at that time) dial-in hackers getting access to company’s data and injuring it—whether through copying (e.g., credit card numbers) or malicious deletion.
The law was not aimed at people like Corcoran, who was not a hacker at all but a contractor, even if a very bad one. Indeed, one view of the facts might be that had Corcoran not destroyed the data files and had MCS hired someone else to finish his work on the spreadsheets, it might have been more apt to find MCS guilty of violating the WCCA.
With that understanding, however, it won’t surprise you that an argument that relied on the semantics of “delete” and “destroy,” though interesting, was ultimately unpersuasive. Sure, Corcoran didn’t destroy the data—that is, he didn’t burn the copies of the restaurant reviews, nor did he wipe the files off the hard drive (he may not have known how). He issued, though his programs, DOS commands that “deleted” the files in the way that DOS did it, which was simply to clear the file names from the disk directories. (All that discussion about the FAT shouldn’t confuse you into thinking Corcoran did anything clever—that is a description of what DOS did, not what Corcoran did). So he did not destroy the data—he didn’t even delete the files in some sense.
But does it matter? Remember why this law was passed. If “destroy” were interpreted so literally as to mean that Corcoran’s action didn’t “destroy data,” then the WCCA would be pretty toothless. For that would mean that hackers (the real object of the law) could at their leisure dial in to computer networks and delete files all they wanted to, as long as they didn’t go the final step of wiping the applicable sectors of the disk. Think through the logical application of your argument—if it boils down to an absurdity, you can rest assured that opposing counsel will point this out with glee.
Did Corcoran have “authorization” to delete the files? Well, if he owned them, of course he did and likewise, if he owned them, then what he did was neither willful nor knowing destruction in the first place. Since he doesn’t own them, where might “authorization” have come from? He arrogantly insists that he himself provided the authorization, and so refuses to make the alternative argument that if that isn’t right, then MCS gave him all the authorization he needed.
Would that have worked? Certainly he has more authority than the hackers the WCCA is aimed at—who are unknown, unwanted intruders who have only ill intent. By contrast, Corcoran was invited onto the premises and given access to the computer on which the files were stored; he was authorized to create and modify the structure of the files; and since MCS appears to have begun putting data in before he finished the reports, we can assume he had the authority to modify the data if necessary in the course of fixing his reports and possibly the macros in the spreadsheets.
This is a start, but it probably doesn’t get him too far. “DestroyAll” (like “Carnivore”) is a pretty poor choice of names for a procedure, for one thing, leaving Corcoran little room to argue that MCS authorized him to simply erase the files for no reason other than his own revenge.
And yet as many of you noted, data destruction happens all the time in the case of trial copies of software, shareware, and the like, where you enter information into a program knowing that after a certain period of time the software will shut down and perhaps delete itself and your data. Or data is destroyed when the operating system crashes and you lose your unsaved work.
These are good counter-examples to have brought in, but you needed to distinguish them from Corcoran’s case. In the shareware/trial software example, the user accepts an agreement that acknowledges the potential destruction; likewise, with operating or application software, the license agreement acknowledges that any loss of data is not the responsibility of the provider. It’s true that you can’t contract your way out of criminal law, but there are other reasons Bill Gates isn’t sitting in a Wisconsin prison. The license agreement, for starters, surely constitutes “authorization” to modify or even destroy data if the right conditions are met. In the case of bugs, beyond the authorization the “willful and knowing” requirements are not met.[3]
Applying the Reading Materials
In both the discussion of Corcoran’s appeal and what he could have done to avoid prosecution and conviction, I hoped to see lots of specific references to the readings—nearly all of them had something useful to say. The best papers told me which materials were being referenced, explained why they were relevant, told me how they spoke to the Corcoran case, and then drew some conclusion. Some papers, unfortunately, took your lead from Corcoran and stayed focused on just the copyright cases. Worse, some of you demonstrated, to my dismay, that you missed the most basic message of some of the most important cases:
§ From Feist and its progeny, we have the rule that “sweat of the brow” doesn’t apply; that is, you do not get copyright for works simply because you worked hard to create them. By and large, data compilations do not get copyright protection, unless the data represent subjective, creative work expressed as numbers, or if the compilation is itself original enough—but then all that gets protected is the arrangement, not the underlying data. The bottom line is that you cannot get copyright protection for simple and basic facts, period. (You can get property rights to your copies of such data, however, whether they are on paper or a hard drive.) Note, as some of you did, that the law is very different in the European Union.
§ The difference between Aymes and Aetna-Standard is a key difference between copyright and patent. An employer is presumed to be the owner of copyright for works created by its employees; an employer is not presumed to be the owner of patents for works created by its employees. (Copyright is easy to get, patent is hard, and therein the rationale for very different rules in the employment context.) There is no “shop right” for copyright, only for patent, and then only for patents developed by employees, not contractors.
§ “Work for hire” means copyrightable work for which the “author” is deemed to be the company, not the individual who created it. All work by employees operating within the scope of their employment is “work for hire.” Work by contractors may be “work for hire” if, among other requirements, there is a written agreement explicitly acknowledging this fact before work commences, and the contracting party significantly directs or designs the work in question.
I am not going to go through each reading separately; suffice it to say that all of the readings were put to good use in the sum of all the papers, as well as some impressive outside material (though it was not necessary to do so).
One very subtle argument, however, is worth noting, as only one person made it. UCITA, had it been in effect in Wisconsin, may have given Corcoran authority for precisely the sort of self-help he engaged in. See #12 in Jean Braucher’s list of Top Twelve biggest problems with UCITA in the memo included in the readings. Compare § 816(c) with § 605. I was surprised to see so few references in the papers in general to UCITA, but then maybe not.
What Could Corcoran Have Done?
This of course is the bitterest pill to swallow in this case, for we are left (from ProCD and other cases) with the realization that all that stood between Brian Corcoran being convicted rather than vindicated was the lack of documents written by lawyers that would likely have been read by no one. Prior to beginning work, Corcoran should have gotten written agreement from MCS that he would be paid as he was doing the work, which would have solved most of his problems. Beyond that, a license agreement (nothing fancy and no need for shrinkwrap) absolving him of any responsibility for damage caused to data, and authorizing him to delete or disable his software if his bills were not paid or MCS failed to accept his software in good faith or give him adequate specifications, would have solved the rest.
Even worse, despite these mistakes Corcoran could likely have prevailed had he gotten any kind of legal counsel. A lawyer would have helped him understand why his dogged, rigid, formalist, literal and just plain wrong reading of the copyright statute was a non-starter, or at least why it was important for him to argue additional defenses. Most notable is the fact that he did not challenge the amount of damages offered by MCS. The $4000 is suspect because the explanation of Mary Mueller implies that indeed his work was used eventually—some of the money went to reentering the lost files but some seems to have gone to finishing what he started. He could clearly have argued for an offset in the amount for the value of what he did (it was not “worthless” or “full of errors”).
A lawyer might also have tried to argue the broader socio-economic implications of shareware and trial software and the “tradition” of such transactions prior to the WCCA, the fact that WCCA was aimed at a completely different kind of behavior than a contract dispute gone out of control (to date, Corcoran is the only person ever convicted under the WCCA!), and the possible culpability of MCS under the very same statute.
A really good lawyer would have made certain that from trial to appeal to habeas corpus, the judges would have had a clear understanding of the architecture of Lotus and its file structure in particular, what Corcoran’s software really did in the “delete” mode, and why analogies like Posner’s book author erasing pages is one of the worst he’s ever used.
[1] I didn’t expect anyone to notice this, and no one did, but there is a hint as to why Posner’s opinion is of such poor quality. The heading notes that the case was “submitted” on March 4, 1997, rather than “argued.” This means that there was no oral argument for the appeal (see Fed. R. App. Proc. 34); that is, that the court of appeals considered the case solely on the briefs of the parties. The decision to deny oral argument is usually made for cases of little merit, and such cases are subsequently worked on not by the Judge’s law clerks but by someone from the pool of court clerks who work on the lesser cases.
[2] I base some of this synopsis on materials you did not have access to, including excerpts from the trial record and briefs on appeal. Copies of these items will be available when you pick up your papers, and you should read them if you’re interested in knowing more about the specifics. I thought you had enough facts and paper to deal with, which is why I didn’t give you these excerpts in the first place.
[3] Some argued that if you know there are bugs and don’t fix them or inform the users in a timely manner, you have “willful and knowing” destruction of data. I can only say this argument will get you no where in a criminal court of law, and thank goodness.