« Just How Mythical Is the Rational Market? | Main | Scientific Theories and the Efficient Market »

Tuesday, April 26, 2011

Civil Procedure in Patent Clothing

The Federal Circuit Court of Appeals recently decided an important patent civil procedure case (en banc) about the scope of injunctions. Tivo obtained an injunction against Echostar in a long-running patent infringement case relating to DVR software. The injunction included two features:

1. Don't distribute any new products that infringe, and

2. Disable recording in a particular list of products that have been adjudicated as infringing.

So, Echostar "designed around" the patent, or so it believed. It distributed its modified DVR software in new products and re-enabled recording in the existing products using the modified software.

The district court held that the change by Echostar was insignificant and thus infringing, because its "design around" infringed the patent in a different way. Echostar argued that any attempt to design around the patent warranted a new trial.

The court also held that disable means disable, and reenablement - even if non-infringing - was contempt of court. Echostar argued that the injunction was overbroad, and it was improper to enjoin non-infringing products.

So, the question on appeal was how patent contempt proceedings should proceed. I discuss below.

First, the easy issue  - the disablement injunction. This one is cut and dry. The injunction - though surely overbroad - was not appealed by Echostar, nor did Echostar file a motion to modify the injunction. Quite simply, it misread the injunction (whether intentionally or not is a question I will leave to others, though let this be a cautionary tale).

As such, the court ruled that this was an easy case - if you are enjoined, you must follow the injunction, appeal it, or convince the court that it is overbroad. Overbreadth complaints are not well taken after the injunction is violated.

I'm not a Civ Pro guru, but this seems to comport with my general understanding of injunction law. It is also the rule I followed in practice, which was painful the few times I had a client on the wrong side of a nutty ex parte temporary restraining order in trade secret cases. ("The defendant is enjoined from talking to any of its customers about anything, including the existence of this order.")

The second issue is slightly more interesting - at what point does the defendant get a new trial to determine whether a design around is still infringing? In the past, the court had a bit of discretion to determine whether or not there is a "colorable" difference in products between the original and the design around. Indeed, the first panel to hear Tivo v. Echostar relied on this to say that the district court was within its right to find infringement by other means.

The en banc panel retained the "colorable differences" test with respect to whether a contempt proceeding was proper - if the new product is more than colorably different, then there is no contempt. This consideration looks at the changes to the disputed patent claim elements by the defendant, and considerable changes mean the court cannot even get into an infringement inquiry. In ththis case, the court remanded so the district court could determine whether there are more than colorable differences in Echostar's new product; the language seems like a clear hint that there are colorable differences.

The court added a new wrinkle, ruling that patentees must also show that the new product infringes even if there are no colorable differences.  Of course, this makes a lot of sense - if the injunction is to not infringe, then one must infringe to violate the injunction. But when would there be no colorable differences, but also no infringement? More often than you would think, especially where the infringed element is a very small piece of a big puzzle. 

Further, who decides whether there has been infringement? Should the court make the decision? Is the defendant entitled to a new trial? Here the opinion is not very helpful. Quoted below is the court's announced rule, which seems to imply that this is a judicial function that doesn't require a new trial:

[T]he court is required to evaluate the modified elements of the newly accused product against the asserted claim, on a limitation by limitation basis, to ensure that each limitation continues to be met. In making this infringement evaluation, out of fairness, the district court is bound by any prior claim construction that it had performed in the case. The patentee bears the burden of proving violation of the injunction by clear and convincing evidence, a burden that applies to both infringement and colorable differences.

This seems clear - the district court judge should go element by element to see if infringement remains. Simple enough, right? Except for when the court is not supposed to make the determination. In this very case, for example, the court states: "It is also possible that, in a new infringement proceeding, a fact finder could conclude that the PID filter in EchoStar's redesigned device meets the "parsing" limitation and that the devices continue to infringe the asserted claims, but that should not be decided in a contempt proceeding."

This quote leaves the question unclear - should any infringement be retried? Defendants everywhere would argue that yes, every claim of non-infringement must be retried.

I'm not so sure. I think the best way to read the case is that the retrial of infringement only happens where the court finds colorable differences. In other words, where the redesign is targeted at the specific claim elements that were disputed and proven at trial. If the changes are only small, then the judge can verify that there is still infringement (or not). But where the changes are big and targeted at disputed claim elements, then allegations of continued infringement by other product features must be retried. This interpretation helps reconcile the seemingly contradictory quotes above, and also makes sense in practice.

There are longer term consequences from this rule. First, it might give an incentive to overlitigate, because patentees will now want to prove as many alternate forms of infringement as possible. Of course, maybe they do that anyway. Second, it puts pressure on the rule that winning parties cannot appeal arguments they lost. After all, it is much cheaper to file a contempt claim than to retry infringement in case of a design around, and then appeal the court's errors.

Even so, this seems to be a reasonable marriage between basic civil procedure rules and specialized patent rules about what it means to violate an infringement injunction.

Posted by Michael Risch on April 26, 2011 at 10:56 AM in Civil Procedure, Intellectual Property | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Civil Procedure in Patent Clothing:


On the issue of the injunction, it was a little more complicated than simply misreading it. The original jury verdict found hardware and software claims infringed, and the judge ordered disabling of hardware and removal of software. On appeal, Echostar won reversal of the hardware claims though not the software claims, and the case was remanded. Tivo v. Echostar, 516 F. 3d 1290 (Fed. Cir. 2008) ("In sum, because of a failure of proof of literal infringement, we reverse the judgment of infringement of the hardware claims with respect to all of the accused devices.")

Echostar took the position that the portion of the injunction requiring disablement of hardware was "effectively" reversed, since it was premised on a now-reversed finding of infringement. But Echostar never appealed the injunction specifically, opting instead to appeal just the infringement finding underlying the injunction. The district court decided, and the federal circuit has now affirmed, that Echostar was still bound by the hardware injunction, even though the hardware claims had been reversed.

The civil procedure lesson, therefore, is that you must appeal both the underlying liability AND the remedy premised thereon. That makes little sense to me, but then, nobody asked my opinion.

Posted by: Brandon | Apr 26, 2011 11:23:28 AM

"But when would there be no colorable differences, but also no infringement? More often than you would think, especially where the infringed element is a very small piece of a big puzzle."

I'm having a little trouble imagining this, particularly given the doctrine of equivalents. What's an example?

Posted by: Bruce Boyden | Apr 26, 2011 11:41:28 AM

Not to beat a dead horse, but the colorable differences test asks whether the differences between the infringing/enjoined product and the new product are "no more than colorable." So if the differences are merely colorable differences then you proceed under contempt, whereas if the differences are more than colorable, then you need a new case.

Importantly, it is the difference between the two products (old vs. new) that is measured, and the claims are essentially irrelevant other than to direct you where to focus your comparison. Under this test, it is conceivable and even permissible that a noninfringing product would still be a contempt, if the difference, though merely colorable, nevertheless avoided infringement (say you went just beyond an outer range and DOE was unavailable due to Festo).

Posted by: Brandon | Apr 26, 2011 4:17:34 PM

Brandon 1: You can only appeal judgments and orders - if an injunctions comes as a separate order, belt and suspenders procedure is to appeal it. May seem odd, but that's the rule in every other type of case. That said, if you win a partial reversal on appeal and think that partial reversal eliminates a portion of an injunction, it makes good sense to ask the court to confirm that fact via a motion.

Bruce: I think the doctrine of equivalents is key. The court talks about looking at the particular elements disputed and how they were changed: "Instead of focusing solely on infringement, the contempt analysis must focus initially on the differences between the features relied upon to establish infringement and the modified features of the newly accused products."

Perhaps one line of code was changed, but the argument is that this renders it non-infringing. I would say that is a small difference, but still non-infringing. Another example is a pencil, where the debate at trial is whether the pencil lead is a particular composition. Let's say that the infringer changes the type of wood. Maybe it is a small difference, but also non-infringing. In either case, perhaps the court must send it to trial on a DOE theory, especially because contempt requires clear and convincing evidence, which a court is unlikely to find on the DOE.

Brandon 2: I don't read this case they way you do, and think that is the primary change in the law in this case. The court clearly says: "Conversely, when a court concludes that there are no more than colorable differences between the adjudged infringing product and modified product, a finding that the newly accused product continues to infringe the relevant claims is additionally essential for a violation of an injunction against infringement. Thus, the court is required to evaluate the modified elements of the newly accused product against the asserted claim, on a limitation by limitation basis, to ensure that each limitation continues to be met."

Posted by: Michael Risch | Apr 26, 2011 4:34:27 PM

OK, refreshing my recollection of DOE... If I'm understanding correctly, the test for whether there are substantial differences between the defendant's product and the plaintiff's invention is whether those differences are more than colorable. So a finding on a contempt proceeding of "no more than colorable" differences would seem to address that question. The only thing left for the district court to determine in a new trial would be prosecution history estoppel, is that right? What if there weren't any relevant changes to the element in question during the prosecution though? It would seem the outcome of the new trial would be entirely determined by law of the case. That seems pretty inefficient. Am I missing/screwing up a step?

Posted by: Bruce Boyden | Apr 26, 2011 5:23:55 PM

I think your scenario is one outcome, but I think it's not the only only. There may be a small change that just not equivalent (e.g. different function). Or a change to an element that wasn't litigated below. Or the chance that "more than colorable" is different than "substantial."

Note that this rule might also benefit defendants. After all, it might be better for a defendant to have a finding of small, non-infringing difference - case closed without filing a new lawsuit by the patenteee - than large difference - case open for new trial before the same judge.

Posted by: Michael Risch | Apr 26, 2011 6:17:17 PM

Re 1: I understand the judgment/order issue but where the premise of the order is reversed . . . What if the order had been that Echostar pay an ongoing compulsory royalty rather than an injunction (e.g., Paice v. Toyota)? Would the royalty still be owed even if the underlying infringement determination was reversed? Isn't the right to exclude (i.e., injunction) premised on infringement, and not naked judicial power? It seems like pure form over substance. What happens to the eBay v. Mercexchange application of equitable factors where the patentee cannot show a likelihood of success on the merits? This is a mirror image of eBay and I hope Echostar petitions for cert.

Re: 2, you're right. The test for contempt vs. new case is colorable diffs, but after that, infringement still matters in both.

Posted by: Brandon | Apr 26, 2011 11:55:19 PM

Post a comment