Thursday, June 3, 2010

Harmfulness ranking of ways to use software patents

Even though I'd prefer for software patents not to be granted in the first place, it's a reality that ever more of them get granted. And enforced: litigation costs related to software patents have for a long time been growing at a faster rate than software sales themselves (as an economist said at a governmental roundtable I attended years ago).

That trend isn't difficult to imagine. Just look at the current situation surrounding smartphones, a field in which there's now a number of lawsuits and countersuits among big players as well as different non-producing entities ("patent trolls") targeting large vendors.

In light of all that's going on, which ways to use software patents are more harmful than others? After giving it some thought, I have arrived at this harmfulness ranking:
  1. most harmful: malicious strategic patent holders pursuing exclusionary/anti-competitive objectives

  2. second-most harmful: non-producing entities ("patent trolls")

  3. least harmful: cooperative strategic patent holders granting licenses to entire portfolios on acceptable terms

Exclusion of competitors is the worst way to use software patents

Software is a field of incremental innovation: not only do new ideas build on old ones but also does a complete computer program contain a large number (depending on size, hundreds, thousands or even tens of thousands) of potentially patentable elements.

It's practically impossible to write any computer program of a decent size entirely on the basis of new ideas, on which one could obtain new patents, without potentially infringing on -- and theoretically requiring licenses to -- large numbers of patents already held by others.

Until some patent reform possibly limits the rights of software patent holders to the collection of royalties (or damages, in the event of litigation), they will continue to be entitled to what the legal world calls "injunctive relief": the right to demand and, if necessary, have a court of law order that someone cease and desist from an action (in case of software patents: desist from using a patented technique in a program). But if every patent holder out there decided to enforce his patents against every infringing computer program on an unconditional you-must-cease-and-desist basis, software development as a whole would grind to a halt.

There's only one type of patent holder who would seek injunctive relief without being willing to grant a license on acceptable commercial terms: a strategic patent holder who wants to monopolize a market or maintain a gap between the functionality of their products and those of the competition.

There are some large patent holders who are hypocritical about their use of patents. They pretend to be willing to negotiate licensing but that is only the case with respect to patents that are of secondary importance to them: they would never engage in any (reasonable) negotiation when the relevant patents are key to their core business.

From the perspective of the "infringer" (I put that word in quotes because software patents are in most cases "infringed" on a completely unintentional basis), dealing with a large player who has that attitude is even worse than facing a troll. I will look at trolls in more detail in the next part of this post and they are indeed a big problem. But a troll "only" wants to maximize his income from the patents he holds. Once you pay him enough, he'll leave you alone and target the next victim.

By contrast, if you face a major strategic patent holder who wants to defend a business that may be worth tens of billions of dollars, there's just no way that you can pay him enough that he'll be satisfied with your check. Even if you're a bigger competitor, you might not be able to offer enough (unless you're able to buy out that entire company), and as a smaller player, you can forget even trying. It all comes down to someone trying to shut you out of a market, or to at least reduce your functionality and make you significantly less competitive.

In a field where innovation is incremental and strong network effects already constitute a major barrier to entry all by themselves, refusing to grant licenses is the most pernicious use of patents. That's why it tops the list.

In some cases, antitrust law can come into play if patents are used anticompetitively. But that is an option only under special circumstances. as I explained on different occasions (such as in a speech at a recent conference).

"Trolls" are a feature -- not a bug -- of the software patent system

The above subhead is a summary of a statement that Carlo Piana, a leading European FOSS lawyer, recently made on

If one believes that certain general ideas should be "monopolizable" through patents, then it's a natural consequence that some will obtain (or acquire) patents and try to derive commercial benefits from them without ever creating their own products. Far be it from me to defend the concept of "patent trolls" -- I just want to point out that it wouldn't be practical to impose an obligation on every patent holder to make actual products. At the most it might be possible to limit the procedural rights of a non-producing entity to the right to be indemnified (excluding injunctive relief).

Alternatively, one could conclude that certain patents don't relate to legitimate inventions deserving of a 20-year monopoly, but in that case one has to oppose all patents of a certain category regardless of whether or not their holders actually build products.

The term "troll" is meant to allude to nasty behavior. However, conduct or misconduct aren't necessarily a matter of having products on the market or not, nor a matter of size. Still, "patent troll" is a common synomym for a non-producing entity asserting patents (sometimes quite aggressively) against producing entities and/or their commercial customers.

What many people recall in connection with the word "patent troll" is the non-producing entity that asserted patents against Research In Motion and only hours before a court-ordered shutdown of the BlackBerry wireless email service in the US reached a $612.5 million settlement. People were shocked that a patent could have disrupted an email service that is considered critical by business users and politicians (such as then-Senator Barack Obama), and the amount of money that changed hands also seemed exorbitant.

However, it's a rare exception that non-producing entities actually get such a deal. Many cases get settled for much less, and even if a large corporation is sued, settlements rarely exceed $20-30 million. At some point, the "troll" will determine that a very big bird in the hand is better than ten or twenty birds in the bush.

In retrospect, Research In Motion not only survived that settlement but did extremely well since then, financially speaking.

If the same kind of dispute had occurred between Research In Motion and a large strategic patent holder defending his turf against an innovative new entrant, the outcome might have been the end of the BlackBerry company as we know it. Even $612.5 million wouldn't have been able to solve that problem. For an example, the larger competitor could have used such leverage to force RIM out of the wireless email business or (only gradually better) to "persuade" RIM's shareholders to sell their company at a fraction of what it was worth under normal circumstances.

Cooperative strategic patent holders are the (relatively) smallest problem

If a real company (that has its own products) decides to license its patents on reasonable terms to third parties, including its fiercest competitors, it's a much better party to deal with than a troll (and hugely better than a strategic patent holder pursuing exclusionary objectives).

A big problem with "trolls" is that they don't have to be choosy about how they assert their patents. They might go after the vendor of a product, or (alternatively or additionally) after commercial customers. Since they don't have products on the market and don't have a brand reputation to defend, "trolls" can act in very unpleasant ways without that fact affecting their product business. In fact, the nastier they are, the more successful they will probably be. By contrast, a company operating in the same market would have to think very hard about which moves to make, and targeting customers would probably only happen under the most egregious of circumstances because usually the patent holder will want to have an open door for selling his own products to the same customers (if and when possible).

In terms of the amounts of royalties asked for, a company with a cooperative approach will usually be reasonable because it knows very well what it's like sitting on the other side of the negotiating table. In fact, if it talks to a competitor today about licensing its own patents to that other organization, it might a year or two later need to license patents from that competitor as well.

A troll can be much more reckless because if you don't have products on the market, you'll never need anyone else's patents. You can just maximize the income you derive from your own patents. The law is the only limit in that case (hopefully, at least).

Unless a market leader takes the strategic decision to defend a market leadership or monopoly at all costs, it will also want (unlike a troll) to maintain a reasonably good reputation.

All those differences considered, a company with a product business would very likely have struck a license agreement with RIM on acceptable terms early on, without the dispute escalating to the point where a shutdown of the BlackBerry service was imminent.

It certainly hurts a company from a strategic perspective to have to pay patent royalties to a competitors. Every dollar you pay requires you to raise your own prices, which makes you less competitive, and your competitor can then spend every such dollar on further product development (and/or marketing). But whether or not that is reasonable again depends on what one thinks about the legitimacy of a category of patents. If the 20-year monopoly seems justified and the ratio between the breadth of the patent and the achievement behind it appears reasonable, then it's legitimate that the patent holder gains a significant competitive advantage. If one doesn't think it's reasonable in a particular case, then one may be against a patent or an entire category of patents, but even then it's better to deal with a cooperative licensor than with a troll, let alone with someone whose objective is the elimination of competition.

The comparison in a nutshell

To sum it up with the BlackBerry example, a non-producing entity ("patent troll") endangered RIM's business and walked away with $612.5 million. A malicious competitor defending his turf probably wouldn't even have accepted a hypothetical payment in the billions of dollars. One way or the other, that kind of aggressor would have caused damage going way beyond that amount which RIM was ultimately able to afford. By way of comparison, a cooperative licensor probably wouldn't have pushed the envelope to that extent and would more likely have reached an agreement with RIM rather quickly and for a fraction of the amount, especially since that competitor would have had to factor in (unlike a troll) the possibility of one day also needing licenses to patents held by RIM.

There would certainly be no need to license any patents from anyone if there weren't any patents in the relevant field. But even RIM, despite its $612.5 million lesson and other problems with patent holders, never once made a political demand for abolishing patents in that area. All they ever advocated was some kind of patent reform concerning the examination process and, especially, litigation. I met with RIM's chairman less than two months after that famous settlement, and he explained their priorities to me.

Since there isn't significant support for the abolition of software patents outside the FOSS community (as I explained in my initial post to this blog), it's key for us to ensure companies at least use them reasonably. That's why I started this blog, and it will continue to be the number one priority. This includes defensive initiatives such as the Defensive Patent License (DPL), which could be a first-rate opportunity to deal with the problem.

If you'd like to be updated on patent issues affecting free software and open source, please subscribe to my RSS feed (in the right-hand column) and/or follow me on Twitter @FOSSpatents.