In its weekly roundup, IAM (Intellectual Asset Management magazine) noted that "[p]atent suit numbers are falling in Dusseldorf and Mannheim but are on the rise in Munich as the city’s courts develop a strong pro-plaintiff reputation." I'd like to comment on that, also because I reported on the premiere session of the Munich I Regional Court's third patent litigation division (the 44th Civil Chamber under Presiding Judge Dr. Georg Werner) last week. If you wish to skip directly to the part on software patent-eligibility in light of Solas OLED v. Samsung, please click here.
Munich is definitely an attractive venue for patent holders, and that is so for a variety of reasons not least including the one noted by IAM. The Munich court still does throw out or stay cases that don't meet a certain standard, and that's why the first case the 44th Civil Chamber heard appears, for now, unlikely to be the first in which it will grant an injunction. I'll get to the fundamental weakness of the patent-in-suit in that Solas OLED v. Samsung case again in a moment--the headline of this post relates to it.
If we assumed for discussion purposes that better-than-average chances of winning favorable rulings are a major factor in this, then Munich would simply be beating Dusseldorf at its own game. For many years--actually, a couple of decades--it was no other court than the one in Dusseldorf that consistently and shamelessly set a low bar, only to attract as many patent cases as possible to a city that is anything but a major center of technological innovation, as opposed to the Munich area with BMW, Siemens, Linde, and so many others. Comparing Munich to Dusseldorf is like Silicon Valley vs. Chicago at best, Pittsburgh at worst.
By the time I got involved with patent policy (2004, the European legislative process on computer-implemented inventions aka software patents), Dusseldorf already had a reputation for disadvantaging defendants in different ways. I know this from discussions with patent litigators and patent attorneys in the mid-2000s, and I know there was at least one German media report at the time that discussed the issue.
The biggest part of the problem is that Dusseldorf is the venue that exacerbated the practical impact of bifurcation: the German approach under which alleged infringers don't have a full invalidity defense in the infringement case itself, but must separately litigate validity, with the infringement court merely performing a rough assessment as to the likelihood of success in order to decide on whether to stay the infringement proceedings pending a parallel invalidation action.
It was Judge Dr. Thomas Kuehnen ("Kühnen" in German) who many years back created case law according to which an infringement case should not be stayed even if the patent is reasonably likely to be invalidated: no, according to the Kuehnen doctrine on stays, the standard is so exacting that the patent has to be extremely likely to be invalidated. Dusseldorf didn't become a popular patent venue just because of the quality of its judges: it lowered the bar for patent assertions to succeed, which in turn attracted many cases, and the more cases a court gets in a field of law, the more expertise its judges will build over time. But there's absolutely no indication that those Dusseldorf judges were smarter than their counterparts in Munich or Mannheim at any given point in time.
The economic damage that the Dusseldorf court has dealt to actual innovators over the years due to its unbalanced decisions must be huge. One particularly shocking and embarrassing case is now getting more attention: Schweitzer-Mauduit v. Julius Glatz. I discussed that one in a Halloween post calling it a creepy patent case. Meanwhile, Frankfurter Allgemeine Zeitung, the most famous German newspaper, has also written about it. What I learned from the FAZ article is that the number of layoffs may be closer to 30 or 35 than the 60 I had previously heard. And I heard that the plaintiff, a publicly traded U.S. company represented by Quinn Emanuel, even sued (over its actually invalid patent) the couple that runs the company (and owns many of its shares) personally, seeking 17 million euros in damages from those people, depriving them of the benefits of their company's liability limitation.
Suing executives personally in Germany is a tactic I discussed in a post last year. Those who were pushing for the recent German patent "reform" bill missed out on a huge opportunity, and one of the smaller things they could have accomplished (if that group had not been mostly a mix of amateurs and saboteurs) would have been to do away with executive liability for patent infringement. While the law allows this, the question is, of course, whether a firm like Quinn Emanuel Germany should engage in such practices that the clients it mostly works for (in Germany, that means companies like Google and Daimler) sharply disapprove of. In that SWM v. Glatz case over a cigarette paper patent, QE Germany acted in such questionable ways that I was thoroughly disappointed to find out about it. The Frankfurter Allgemeine article just adds to that unfavorable impression of what they did in that case.
I do, however, wish to clarify that what QE Germany did in that particular case is presumably an outlier. As for my criticism of their performance (i.e., rate of success) when they were defending Daimler against Nokia, it was definitely the poorest set of results I had ever seen from them, and a look behind the scenes revealed that they had gotten some cases stayed, but Nokia was really beating them in the nullity and opposition proceedings and would have won two or three more injunctions pretty soon. The results don't support Daimler's decision to work with QE Germany: they could have worked with just about any other German patent litigation firm (as my post noted), given that they lost big-time and could have had those poor results at a (much) lower cost. I still believe that QE Germany would get better results if they did what every other patent litigation firm in Germany does (at least I'm not aware of any exception): to have patent attorneys involved every step of the way to benefit from their technical knowledge and patent prosecution expertise.
Solas OLED v. Samsung raises fundamental question of software-related patent eligibility
The infringement-related part of the discussion showed an issue that is familiar from other cases involving software architecture patents: compared to physical aspects (such as whether a certain layer is made of a particular material), it's much harder for courts to determine whether a certain structure is actually found in program code. Last week some reference was made to "pointers" being passed along. So, is passing a pointer on from one code segment to another the same as passing the data on? Is a certain object raw data (whatever the capacitive touchscreen measures with its horizontal and vertical wires) or is it the result of some kind of evaluation? Is a "server" a separate computer or could it just be a subroutine of a computer program (Apple's "quick links" patent)? In the U.S., those questions go before juries unless they can be resolved purely through claim construction. In Germany, the courts try to decide them based on the pleadings.
It doesn't appear exceedingly likely that Solas OLED will even win the infringement ruling it is seeking. But in any event, that patent-in-suit is outrageous.
I'll use some Microsoft terminology now because almost all of the coding I've done over the last 20 years has been on the Microsoft .NET framework (where C# is the predominant programming language). There's the important distinction between "design time" (which is when programmers edit code in the development environment--Visual Studio in that case) and "execution time" (which is when the program actually runs on a computer).
Even those favoring a broad scope of patent-eligible subject matter in connection with software should draw the line where the benefits that a patent claims to generate are purely relevant at design time, not at execution time.
To recognize a multitouch gesture like pinch-to-zoom (clearly multitouch because you need two fingers), it is one possible strategy to identify single-touch gestures (such as when each of the fingers taps on the screen) and to then discern a multitouch gesture based on a combination of previously-identified single-touch gestures. That's a step-by-step approach. But it's also possible to proceed directly to the multitouch analysis (based on the raw data)--and that multitouch analysis could, theoretically, even be implemented with AI algorithms, which would just figure out the most efficient path based on experience.
This is the key point now: If the "one-two" approach of firstly identifying single-touch gestures and then recognizing multi-touch gestures based on a number of single-touch gestures (like firstly identifying trees, then saying all these trees are a particular type of forest) had technical advantages at execution time, there might be something patent-eligible about the claimed invention. Not so here. The patent specification itself merely suggests that directly analyzing raw data for multitouch purposes is more error-prone (and harder to test), meaning that humans who write the program code are more likely to deliver quality work under the "one-two" approach than the direct one.
There's no such thing here as inherently superior reliability (and/or efficiency) for technical reasons. It's all about reducing the complexity of a problem for the human mind by breaking up the software design process into two steps. You can't claim the human mind in a patent.
The Munich court knows the position of the Federal Court of Justice on interpreting an issued patent: for infringement purposes (invalidity is another story), courts have to assume that the patent examiner granted the patent because there was some technical benefit somewhere. In this case, the Munich court reasonably says that the patent argues it makes it easier to add support for new gestures. The claims do not refer to the addition of new gestures: they just talk about recognizing gestures (whether old or new), as opposed to referring to (for example) an entry system for defining a gesture. But based on the specification, it is correct to say that if the patent purports to solve any problem, it's the problem of supporting programmers tasked with writing gesture recognition code.
Whether raw data is analyzed directly or whether there is the two-step approach taught by the patent, code can be reliable or unreliable. You can write perfectly reliable and efficient code either way. Or you can leave it to an AI system, which will then figure out over time what works best for its purposes.
If it's all about simpler, cleaner code from a human perspective, then we're not at the level of automation. Even software developer tools may meet the machine-or-transformation test. But there has to be an automated process. You start some program code, and it does something that other code can't do, or it does the same but outperforms the state of the art (by being faster and/or more reliable and/or more secure and/or using less memory, bandwidth etc.).
It's really astounding that whoever examined that patent application at the German Patent & Trademark Office failed to see that there was no such thing as a technical effect. It's about a coding style and mere probabilities of whether programmers will do a better job one way or the other. It's not about a specific process implemented in software and yielding per se technical benefits.
If the mere proposal to break up a complex software development problem into smaller building blocks because they're easier to digest for the human mind and save the effort of reinventing the wheel was patent-eligible without specific execution-time benefits, we'd see a deluge of additional patent infringement cases against the entire technology industry.
Share with other professionals via LinkedIn: