Once inside, I waited in line to pass through the metal detectors, as a guard watched from a balcony overhead.  Fortunately I had left my pocketknife in my desk a few blocks away.  Otherwise I would no doubt have been denied entry, lest I rush the bench, flashing an inch and a half of cold steel.

Finally I took the elevator to the 27th floor, where I found the Harry A. Blackmun Rotunda, featuring photographs and memorabilia of the late Justice.  He had served in the 8th Circuit Court of Appeals from 1959 until 1970, when President Nixon appointed him to the Supreme Court.

The furnishings and fixtures were still fresh and gleaming.  The south wall was all windows, affording an excellent view of Busch Stadium.  You could easily watch a ball game from up here, if you didn't need a hot dog.  In the courtroom, however, the proceedings were shielded from such distractions by floor-to-ceiling curtains.  A curving wall of dark and polished wood stood as backdrop to the bench.

After a couple of unrelated hearings, Paul Grewal of Day Casebeer spoke for the appellants, the developers of bnetd.  The judges asked several questions during his presentation.  By contrast the attorney for Blizzard, Stephen H. Rovak, gave what amounted to a monologue, with little or no interruption from the bench.  Finally Paul Grewal presented his rebuttal.

On Groklaw AllParadox has already given his account of the proceedings.  I won't try to replicate it with another blow by blow recounting, which I could hardly do in any case, because I couldn't take notes fast enough to keep up.  Instead I shall give a more general summary of the issues as I understand them, and my overall sense of what was going on.

I am no lawyer, and I have not followed this case the way I have followed SCO v. IBM.  I haven't researched the legal aspects much either, and I'm sure I'll get some details wrong.  I hope that others will catch at least the most egregious of my blunders.

What's left of the case, after the lower court rulings, poses two main questions, which I shall address in turn.

The first question: when the defendants agreed to the click-through EULA that came with the game software, did they thereby waive their right to reverse-engineer the communications protocol?  The EULA says they did, but is that provision of the EULA enforceable?

Grewal emphasized that he was not contesting the enforceability of click-through licenses in general, which is well established.  Rather, his argument was that the enforceability of EULAs was governed by state law, which -- in this case -- was preempted by Federal law in the form of the Digital Millenium Copyright Act (DMCA).

While the overall intent and effect of the DMCA was to strengthen the rights of copyright holders, Congress saw fit to carve out a special exception for reverse engineering, narrowly defined.  This provision was designed to foster innovation and competition.  I'm not sure what that narrow definition is; it may refer to reverse engineering for the sake of interoperability.  In any case Grewal maintained that this exception covered the activities of the defendants.

In making this case Grewal had to face a major obstacle.  The Bowers v. Baystate Technology decision upheld a clickwrap contract prohibiting reverse engineering.  To get around that obstacle he had to take a different tack.

According to Grewal there are three ways that federal law may preempt state law: express preemption, conflict preemption, and something else that I didn't catch.  The Bowers decision was based on express preemption, but Grewal's argument was based on conflict preemption.  I don't know what this distinction amounts to, but it enabled him to maintain that the Bowers case was not on point.  Instead he preferred to rely on an earlier ruling in the Vault case, which was based on conflict preemption.

Rovak (the Blizzard attorney) responded that the Bowers decision cited the Vault decision but rejected it as a precedent.  On the face of it that observation could favor either side.  Maybe the Vault decision was wrong, as Rovak implied.  Or maybe the Vault decision was not on point in the Bowers case, but still on point for the Blizzard case.

The second question received less attention: did the defendants violate the DMCA by circumventing the security features of the Blizzard software?  Blizzard's own servers authenticated the clients in order to deny services to pirated copies of their games.  The bnetd developers offered to provide such authentication, but they needed Blizzard's cooperation to do so, and Blizzard wouldn't cooperate.

This part of the argument is hazy to me.  If I crack encryption in order to rip a DVD, I'm accessing information whose owners tried to protect it through technical means.  That's a violation of the rules (and whether I like the rules is another matter).

However the bnetd developers weren't hacking the Blizzard servers or otherwise accessing anybody's data without permission.  They only accessed the data of their gaming clients, who -- by connecting to the bnetd server in the first place -- had implicitly authorized that access.

Apparently Blizzard contends that the reverse engineering itself was a circumvention prohibited by the DMCA.  The defendants respond that the DMCA makes a specific exception for this kind of reverse engineering.

One argument I didn't hear anyone make: if the exception carved out by the DMCA could be so easily defeated by the terms of a EULA, then what is the point of the exception in the first place?

All in all, Grewal's presentation was systematic, methodical, and disciplined, if sometimes a bit rushed.  Rovak's presentation, on the other hand, was an exercise more of rhetoric than of law.  He dredged up irrelevancies, he rambled, he huffed and puffed with wounded indignation.

In particular, he tried to tar the defendants with a broad brush of software piracy, even though they had legal, non-pirated copies of the Blizzard game software, and had no way to determine whether other parties (their clients) were using pirated software.

Likewise Rovak complained of copyright infringement.  Since the bnetd server performed the identical services as Blizzard's own servers, he said, that's copying, even if the code was written separately.  This argument is easily refuted:

1. The bnetd server was interoperable but did not behave identically -- in fact it offered a number of enhancements, for example for tournament play.

2. Even if it did behave identically, identical behavior does not constitute copyright infringement.  Patents cover function; copyright covers expression.  If the bnetd code was developed independently, without access to the Blizzard code, then it did not infringe copyright even if it turned out to be substantially similar or even identical to Blizzard's code.

3. The lower court had already dismissed all of Blizzard's copyright claims, with prejudice.