Saturday, September 15, 2018

In Munich iPhone case, Qualcomm wants thin air to infringe software patent claim limitations: dangerous precedent possible

The name of the game is the claim. There's no German equivalent, at least none that rhymes, but I vividly remember Quinn Emanuel's Dr. Marcus Grosch stressing this point when he was defending Motorola Mobility against Microsoft, particularly in the Munich appeals court. Now, as counsel for Qualcomm in a German lawsuit against Apple that will be tried on Thursday and involves the iPhone's built-in Spotlight search technology, he's espousing the very opposite position. The former anti-software-patent campaigner in me dreads and hates the notion of software patents being given overbroad scope in litigation--way beyond any reaonable claim construction.

Whatever may come out of that Munich case (technically, a host of cases asserting different members of the same patent family and targeting different Apple entities) is rather unlikely to move the needle in the intercontinental Qualcomm v. Apple dispute. Based on how a first hearing went in May, even an infringement finding wouldn't result in an immediate injunction: the case(s) would most likely be stayed over validity concerns. But Munich is one of Europe's most important--and fastest--patent litigation venues, and Presiding Judge Dr. Matthias Zigann one of Europe's influential patent judges (among other things, he's a contributor to a leading reference on patent law). An infringement holding could set a dangerous precedent affecting many other cases before the Spotlight matter, if ever (since the patents-in-suit might all die anyway), would be appealed.

The issue here is related to the nature of patent-eligible inventions, to claim construction, and, ultimately, to the principles of infringement analysis.

Half a century ago, the Federal Court of Justice of Germany defined, in its Red Dove ("Rote Taube" in German) decision, what constitutes a patent-eligible invention: "Patent protection is available for a teaching of a planned course of action involving the use of controllable forces of nature in order to achieve a result with an identifiable causation." (emphases added)

If the forces-of-nature part was still applied, most software patents would be invalid anyway. But that's a different story. The issue in the Munich Spotlight case(s) is that Qualcomm's infringement allegation is based on the mere possibility of a screen output, in some edge cases, "only representing the usable communication channels [of a given user]" (emphasis added), without a planned course of action and an identifiable causation.

Qualcomm is now seeking to prevail on the basis of a total misreading of the 2005 Federal Court of Justice ruling in Shunting Trolley ("Rangierkatze" in German). As a result, even thin air could infringe key elements of a patent claim.

Here's my translation of the Shunting Trolley doctrine:

"A patent is definitely infringed if the limitations of the patent claim are practiced and the accused embodiment is objectively capable of delivering the features and effects covered by the patent. An infringement holding is not avoided because an apparatus is normally operated in different ways and customers therefore do not usually make use of the relevant teaching, even in cases in which a vendor explicitly recommends using its apparatus differently, provided that there is a possibility of the patented teaching being utilized."

I already vented outrage over Qualcomm's approach in my post-hearing commentary in May. One actually just has to take the very first sentence of the Shunting Trolley doctrine seriously in order to reject Qualcomm's infringement theory: "if the limitations of the patent claim are practiced". Not so here. You can find an exemplary set of those claims on Google Patents. The key passage here is in the penultimate paragraph of claim 1:

"presenting, on the display of the first communication terminal, a plurality of selectable items only representing the usable communication channels, responsive to sensing input of a title" (emphasis added)

The word "only" is key here. Qualcomm's counsel says: if a user only has usable communication channels (suzch as only one usable channel per user) in an address book, the search result will be limited to only usable communication channels. Unfortunately, there's no Markman proceeding in Germany, where claim construction would be determined first and where disputed claim limitations would have to be interpreted by the court in writing. Otherwise I have no doubt--no doubt whatsoever--that this claim element here would have been deemed a filter (a word that, therefore, appeared multiple times in my May post on this matter).

Therefore, under general principles of patent law and the way the first sentence of the Shunting Trolley doctrine restates one of them, Qualcomm would actually have to prove the existence of a filter. The fact that Spotlight normally displays information related to a user that does not constitute a "usable communication channel[]" strongly suggests that there is no such filter. (Theoretically, the display could aggregate information retrieved with multiple filters, but at least at the May hearing, Qualcomm didn't even make that allegation and Apple's counsel, Freshfields Bruckhaus Deringer's Wolrad Prince of Waldeck and Pyrmont, denied the presence of a filtering operation of the claimed kind.)

The discussion in May had me so concerned about how software patents might get overbroad scope in future German infringement cases that I read the entire Shunting Trolley opinion. As I had assumed, I found that it stops far short of the basis on which Qualcomm seeks to prevail here:

The mechanical patent asserted in Shunting Trolley covers a certain kind of braking system: a particular type of rail brake. Even the defendant couldn't deny that the accused product contained such a rail brake, nor could the defendant deny that under some rare circumstances, more or less accidentally, the rail brake could be activated, but stressed that a different brake, called "operational brake," always got activated when a user hit the brake, while the rail brake was not intended for normal use:

So the case turned on the relevance of either brake to the infringement analysis. The Shunting Trolley defendant naturally wanted the analysis to be limited to the non-infringing one of the two brakes:

The Federal Court of Justice, however, sided with plaintiffs and deemed the trigger irrelevant (even a mere coincidence is sufficient), and applied the usual rule that the additional presence of non-infringing components is irrelevant:

Qualcomm's position is correct on the left side of that chart, and at the top of the right side (deeming the trigger irrelevant), but throws the baby out with the bathwater by also deeming the claimed cause and causation (see the pink color in the image below) irrelevant, limiting the entire analysis to the effect:

The legal error in the pink part is that the Federal Court of Justice never said a mere coincidence could serve as a substitute for practicing a claim limitation (which takes us back to the identifiable causation it already stressed half a century ago). Nor did Shunting Trolley say that you only had to look at whether a braking occurred, ignoring what caused it and how. If we now apply the only reasonable interpretation of "only" in the given context--a filter--to the Qualcomm v. Apple case and correctly map it to the Shunting Trolley doctrine, we find that the user input is irrelevant to the analysis, but there must still be a filter, in the absence of which there simply is no infringement:

Thin air is not a filter. Otherwise, thin air could infringe a patent on a coffee filter if filtered coffee comes out at the bottom because filtered coffee is poured in at the top. It would be absurd, but it would be particularly bad for software innovation, given that patents on data processing often involve an output that can meet certain criteria with a particular set of input data, even if no patented technique is utilized. For another example of an analysis confined to a mere digital output, a screen displaying a different random number per second could be deemed to infringe a patent on a digital thermometer because, if you watch it long enough, it will sooner or later happen to display the actual temperature.

I still hope Qualcomm's infringement theory will fail. If not, then maybe some other defendant in some other case will make use of the above and get a total misreading of Shunting Trolley overturned...

Share with other professionals via LinkedIn: