Thursday, May 3, 2018

Through actions and words, Qualcomm recognizes deficiencies of patents asserted against iOS search

Today's Qualcomm v. Apple hearing in Munich (see my previous post on Apple and Intel's invalidation efforts and my tweet on the postponement of the hearing by 4.5 hours on short notice) started with a surprising move by Qualcomm: a withdrawal of several asserted claims. (Last month Qualcomm also withdrew several asserted claims from its first ITC complaint against Apple.)

Qualcomm's German lead counsel, Quinn Emanuel's Dr. Marcus Grosch, handed out the related motion to dismiss and explained that Qualcomm's intent was to narrow the (in)validity issues in the case, the obvious objective being to avoid a stay of the infringement case (or of the enforcement of a potential injunction) pending resolution of any Apple-Intel nullity actions or the opposition (to a recently-granted EPO patent) I blogged about earlier today.

The Munich I Regional Court has defined some local rules for patent infringement cases that are unique: as opposed to a single trial, an early first hearing is held. Usually it is so early that the parties have just filed, or announced an intent to file, their validity challenges in other fora (mostly the Federal Patent Court, sometimes the EPO, rarely with the German Patent and Trademark Office). Therefore, validity discussions are reserved for the second hearing, which is then--unless further court sessions turn out to be necessary--a trial followed by a ruling. But today's hearing was unusual in this regard. I had never before seen a first hearing in Munich at which validity would have been mentioned and alluded to. Not only did Qualcomm point to it as the rationale for its withdrawal of certain claims it believed to be jeopardized by a prior art reference from Korea but Judge Dr. Zigann also referred to validity in two contexts:

  • When Apple's lead counsel in today's hearing, Freshfields Bruckhaus Deringer's Prince Wolrad of Waldeck and Pyrmont, said an identified vagueness of the patents-in-suit shouldn't disadvantage Apple, Judge Dr. Zigann alerted him to the fact that a breadth of scope running counter to Apple's interests in the infringement context may actually serve it very well on the invalidation front.

    As for vagueness: in a related context, Qualcomm's lead counsel acknowledge a "shortcoming of a definition" in the patent document. That kind of concession is rarely heard in court, but was inevitable in this context. The patent specification (not counting claims and figures) has only about 10 pages and indeed contains various unclear terms and expressions.

  • Judge Dr. Zigann indicated that, after looking at the prior art references cited by Apple and Intel, a stay pending resolution of the parallel challenges to the validity of the patents-in-suit is quite a possibility. And toward the end, he told Qualcomm directly: "Your withdrawal of claims shows that you know you have to do something to avoid a stay."

It obviously wasn't possible for the court to form an immediate opinion on the impact of Qualcomm's surprise dismissal of some claims. Under procedural rules, Apple's counsel could have asked for more time to formally respond, but consented to the motion to dismiss those claims (while reserving the right to seek a partial reimbursement of litigation expenses) and, when asked whether Apple would dispute the remaining infringement allegations without requiring time for further thought, Prince of Waldeck jokingly said: "We can overcome any related inhibitions."

The infringement allegations that Qualcomm dropped related to speed dial (where you can assign a speed-dial number to a contact). What Qualcomm is still claiming is that Apple's iOS devices--formally it's going against Spotlight, but Qualcomm's counsel said that Siri, as incorporated into iOS 11, is also infringing--violate those patents by finding a name, based on entry of a number or letter, and offering different communication channels.

As a former anti-software-patent campaigner it always frustrates me when courts and parties, despite the exclusion of software patents under the European Patent Convention, discuss infringement theories relating to the "user experience" and user-interface software--which is what today's infringement analysis was essentially about.

Things could still change between now and the final ruling, which will come down after a trial scheduled for September 20, but the most likely scenario is that Qualcomm will prevail on infringement, yet the case will be stayed because the validity of the asserted patent claims is too doubtful. In that case, Apple's equitable and antitrust defenses may become moot anyway, which is why the court may never reach them.

I can relate to the court's skepticism of some of Apple's non-infringement theories--with one exception: Citing the Federal Court of Justice's Rangierkatze ("shunting trolley") decision, Judge Dr. Zigann argued that Qualcomm's patents are infringed even if an inefficient way in which some end users presumably organize their phone directory (having a separate entry per each contact number of messenger or email account of a contact) is a prerequisite for achieving the result covered by the asserted patent(s). The Shunting Trolley decision just says that even an uncommon way in which an apparatus is used (and this case here is about apparatus claims) can give rise to an infringement finding because the fact that an apparatus may be used differently in most cases does not serve as a defense to infringement. That's a logical and important principle, but for my taste, looking at it in technical and policy terms, it's highly problematic to deem a search patent infringed if a certain search result and presentation thereof (in this case, it's also about icons) is brought about by a user's way to organize a directory, not a result of automation.

I don't read Shunting Trolley as saying that manual steps can simply substitute for automation when a patent is all about an apparatus that automates a process. If a user operates a device in an unusual way, that's one thing; but letting the user do the job that technology is supposed to do is irreconcilable with the concept of automation. An essential step--for example, the key inventive step of a patent--must not be performed manually. If one thinks it through, it would mean that a patent on a machinery cutting paper in certain ways would be deemed infringed by a less advanced machine, or even a simple knife, if the effect can be achieved by folding the paper accordingly before cutting along a certain line. Or, getting back to search, any search patent would then be infringed by any mobile device if a user could bring up Google and produce a certain type of result (I'm not sure whether that could be said for today's patents-in-suit, but it's possible, given that Google, when accessed from mobile devices, provides an icon for calling a phone number).

Apple's counsel argued that the patent claims say the device must be "configured to" present certain types of search results (and only those results) in a certain format, thus Apple's non-infringement theory isn't foreclosed by Shunting Trolley.

They have more qualified people than me to devise their strategy, but if anybody asked me how to state the issues for an appeals court after an infringement finding based on Shunting Trolley in this particular case, I would argue

  1. that an essential step, especially the primary inventive step, of a technology patent must be performed automatically, not manually, also including manual preparation that, in connection with the claimed technology, brings about a certain result, and that such an automation requirement is distinct from the Shunting Trolley concept that even rare forms of use can infringe;

  2. that, at a minimum, the proposed rule must apply to automated steps that an apparatus claim requires a device to be "configured to" perform (as opposed to broader terms than "configured to," such as the mere potential capability of achieving a result); and

  3. that the patent claims at issue involve a search filter --i.e., "only" (the word appears in the asserted claims) certain results (and not others) should be shown --, and if filtering is essential to the inventive step, then the mere fact that the narrower set of results found in a patent claim is a true subset of the totality of results listed by the search technology does not constitute an infringement unless there is evidence that the related filter is actually applied in the process of identifying the results (this is then about the inner workings of the search technology, not just about what comes out at the end).

Judge Dr. Zigann noted that the Federal Court of Justice is patentee-friendly. Others say so, too, and it's plausible. It may very well be that today's United States Court of Appeals for the Federal Circuit is more balanced than the Federal Court of Justice of Germany. But there are serious issues when a patent involves search filters and is all about automation, yet an infringement is identified where a claimed filter doesn't appear to be in play and a result depends on users organizing their data in a way that comes down to them doing the index-building or iteration that a database or search engine should perform.

Obviously, if an infringing product utilized a patented filter and additionally collected search results in other ways, such as by combining the results of multiple filters or by adding results from an unfiltered search to the results of a filtered search, then using the filter means that the related claim limitation is practiced in the process, even though the result doesn't suggest it because the result of filtering gets mixed with other results. But if no filtering ever occurs, the limitation just isn't practiced.

While Shunting Trolley says that even the "coincidental" achievement of a result can constitute an infringement, the performance of an essential function by a user as a substitute to an automatic function is no longer "coincidental." If a user organizes data in such a way that filtering happens, then the user performs a key part of what the patented invention is meant to do. The Shunting Trolley logic, the way I read it, is that even if coincidental circumstances trigger a patented process (in that case, a mechanical one), the fact that the patented process may occur under such circumstances constitutes an infringement. The concept of a filter ("only") and the notion of coincidence just don't mix in my view. A filter requires an iteration or an index; coincidence doesn't perform the task of filtering even if the result may look in some cases like what a filter would have delivered. Just think of physical filters and it's clear a physical filter isn't infringed because someone enters something pre-filtered at the top.

All of this may never play a role if the patents-in-suit go down the tubes of invalidation. However, there are four patents-in-suit, asserted against different Apple entities in two different lawsuits per patent (one suit per patent targeting Apple Inc., the other one two European Apple companies), and while Apple and Intel may have pretty good chances of defeating all four patents, there is that remaining risk of maybe one claim from one patent surviving--and then an infringement finding would hurt.

The next Qualcomm v. Apple hearing in Munich will be held in two weeks. I don't know yet what the patent-in-suit in that case, or group of cases, is about.

Share with other professionals via LinkedIn: