I listened to This American Life’s new show on patents today; it’s a followup to the original, widely heralded show from two years ago. I think it was well reported, and worth a listen. Despite concerns about NPEs being worse now than two years ago, I thought the show was oddly more balanced than the last time (even though nearly half the show is just a replay of a portion of the original). The part of the show I want to focus on is a patent granted to Crawford based on a 1993 filing date, which claims a large part of the online backup business model/technology (depending on how you look at it).
In the original show, it was revealed that Intellectual Ventures (IV) purchased the patent in 2007, sold it to a shell company in Marshall, Texas, called Oasis in 2010, but retained a “back end” percentage of any settlements/litigation winnings. Oasis sued a bunch of online backup companies, settled with most, and then lost against the holdouts when a jury found that Crawford actually co-invented the patent but failed to include the other inventor on the patent application. In this part, we learn that IV’s back end percentage is 90%, and that Crawford got about 18% of IV’s take. In short, Oasis got 10% as an enforcement agent. I have a lot to say about this topic, but only a short amount of time now, due to deadlines on three different papers (who says professors take the summer off). I did want to note three takeaways:
1. The Intellectual Ventures part of the story is actually the least surprising/interesting part of the story. I think it’s a great piece of reporting; IV made a real gaffe by pointing to Crawford as its model inventor, and This American Life dug and found some answers. But the answer isn’t all that new if you’ve been following this area. At the time Oasis filed its suit, the conventional wisdom was that IV was mounting several lawsuits through shell companies. Maybe this was due to earlier IV statements that they preferred not to litigate, much like people never die at Disneyland.
It appears that the Oasis lawsuit featured in the show is one of those shell lawsuits, though I have no idea whether IV ceded control when it transferred the patent or retained control about who pursue. To this, I ask the same question I asked in my Wired op-ed: does it matter? Would we view this lawsuit or the patent differently if it were brought by IV? If IV had a 10% backend instead of 90%? If Crawford brought the suit? If Crawford’s company survived to bring suit? My answer is generally not, other than having concerns about misinformation and secrecy. Perhaps others feel differently.
2. Many software patents are too broad. This is nothing new, of course, but the show illustrates it nicely. I’ll discuss the Crawford patent below, but the first part of the show nicely illustrates this point with a segment on the podcasting patent. The show argues that the inventors never could make podcasting work, and resorted to audiocassette distribution; the argument is that the inventor merely invented the “idea” of podcasting.
This type of attack on the patent falls under a few different legal rules: written description (claiming more than the inventor actually invented), enablement (claiming more than the inventor actually taught), and functional claiming (claiming the function divorced from the way it was implemented). Further, in Life After Bilski we argue that an idea that is so broad that it is unmoored from any specific application might be an unpatentable abstract idea.
I don’t know where the podcasting patent stands, but my first read of it makes me think it’s pretty broad.
3. My favorite part of the show was the story about how Crawford obtained the patent. While the ending is not so great (invalidity due to non-joinder of inventor), the story itself is one that is rarely told in today’s patent bashing news cycle. It turns out that Crawford and friends came up with the idea for online backup with passwords back in 1991, as part of a failed business venture to solve a very real problem that consumers faced – an easy way to back up data offsite. The story describes how the friends swapped floppy disks, but remember that 1991 was before writeable CDs, and floppy drives were 1.44MB. The average 20MB hard drive (can you imagine? how did we live like that?) would take many floppy disks.
Thus, they came up with a solution. Was it obvious? I don’t know. The first When Patents Attack show implied as much, but the patent survived whatever challenges the litigants brought on that front. For all the talk about many other patents in the same area, it might just be that those patents were improvements on this pioneering patent. It wouldn’t be the first time, but I don’t really have any idea.
The point I want to make, though, is that this is patent law working exactly as it was supposed to. It was people getting together to solve a problem, coming up with a solution, and attempting to protect it. (Well, there was one hiccup, which is that the inventors didn’t get together and file, apparently – but it was invalidated for that reason. This isn’t the case most of the time.) Indeed, the show confirms a key point that IV has always maintained, that a significant amount of money was paid to the inventor. That money might stimulate further investment in the invention in the future.
But another thing happened that often happens; the inventors were too early. They came up with a solution that wasn’t commercially viable yet. The world of 2400 baud modems where users pay for bandwidth by the minute was simply not ready for the online solution, and the business failed to materialize. Indeed, my own firm tried something similar in 1997 working with a Sparcstation and a client who was a Sun engineer; it was still too slow to use in-house, let alone online. Technology needed to catch up.
This is one reason why we often see patent assertions much later in a patent’s life, especially in software. Many really great software solutions–both in patents and in research–implement ideas nobody ever thought of on equipment that’s just barely ready to handle it. It’s not until years later that the solutions seem obvious because the enabling technology makes it commercially feasible.
The question is, what do we do about this? In other areas of patent law, we tolerate long delays. We would prefer not to have them, of course. Indeed, one side benefit of the current assertion trend is that patentees are getting more savvy about commercializing earlier. For example, Crawford and friends could have approached folks with fast networks and computers – universities – with their solution back in the early 1990s. Maybe today they would have, and might have found some early adopters.
But even if patents are not exploited, other patent areas treat this as a feature of the system; most calls to reform the system to avoid this problem have been to commercialize earlier, not to kill off  the patents (so long as the patent is not hidden for too long). Software patents seem to be different, though. The collective wisdom seems to be less tolerant of this delay for software. I think the reason for this stems from my second point, above. Because software patents can be so broad, the concern is that a single broad patent –even a valid one– can render an entire industry infringers. Furthermore, patents filed in the early 1990s expire seventeen years after grant, not 20 years after filing. This effectively extended their life, perhaps for too long.
I’ll end by saying that I share these concerns about software patents, but I’m not so sure that they are so different from the rest of the patent world. There have been plenty of broad overarching patents in a variety of industries over time. Some had poor outcomes that stunted industries, and some fueled industries. People disagree about the software industry, I think.
These were my primary takeaways. I wish I had a clever solution, but I don’t; the show is worth listening to, however, with an open mind.
The patent filed back in 1991 was definitely trivial — at least so far as it was patented. If the patent included specific methods for solving the slow technology of the day to actually make it work, then it would be non-trivial. But it would be restricted to solving the problem for the AS4000 and 2400 baud modems, as described in the episode. Once broadband and faster computers and storage systems are invented, the challenges of inventing that kind of thing completely disappear and their research means pretty much nothing.
What was patented was basically the IDEA of doing a backup. That’s an obvious idea, and shouldn’t have been patented, not only because it wasn’t obvious, but because they didn’t get it to work.
It’s as if I patented the airplane in 1860, but never built a working one, and I was able to sue every airplane company that built a plane that flew. Sure, I came up with the idea of using two wings first, but the idea is obvious, and I didn’t solve any of the inventing challenges that make a plane actually work — and the technological advances that made it possible weren’t around when my patent was made. These guys weren’t “ahead of their time”, they claimed they invented something when it was technologically impossible to even invent it at the time!
The truth is that the Patent Office didn’t understand software when it was coming out — or it was lobbied to allow patents that weren’t actual inventions. Now we’re stuck with this, and the billions of dollars that these companies make is stifling real innovation.
Thanks for reading, and for the comment. I don’t know the true facts of what happened, nor do I know that all the facts were reported. That said, I heard the episode a bit differently. What I heard was that they bought the AS400 and made a working system, but the venture fell through when they realized it was not commercially viable. Indeed, if the solution was really as trivial as folks think, it’s not hard to believe that they implemented it.
Regardless of how we heard it, your point about patenting the idea is a good one. This is what I meant about software patents being too broad. But I think the issue is more nuanced than “it was just the idea of…” The problem is that we have long allowed patents on ideas. The Wright Brothers patent on tail rudder control was for the idea of moving the rudder; the court even allowed infringement long after people figured out how to do it manually rather than with the automatic version in the patent. Many parts of Morse’s telegraph were mere ideas but upheld. This blog post does a good job of highlighting Morse and other famous ideas upheld by the Supreme Court throughout history: http://www.bilskiblog.com/blog/2013/05/the-historical-cases.html
So, I get the view that we should limit these early patents to just the specific mechanism, but it brushes against the goals of patent law. If folks figure out how to do something when it is hard, why should their inventions become valueless when the enabling technology makes it easier? I get that we don’t like it when earlier patents block later improvements, but we also like early inventions, early disclosure, and early expiration.
In the past this was less of a concern, because enough time passed to render the patent moot. Software, however, moves a bit faster, and a long prosecution time extends the patent even more. So maybe the answer is shorter duration – some have suggested this.
A final note: the patent, as far as I know, did not go into reexamination, nor was a summary judgment motion filed on novelty or obviousness. I take that as at least some evidence of non-obviousness under the patent law that we might view differently with 20/20 hindsight more than 20 years later.
I tend to agree with you that the implementation was trivial, even back then (at least based on my recollection of from the time). The patent even mentions several online drive mounting services. This raises another point I often discuss with software patents. We’ve granted patents in the past to trivial implementations of non-trivial ideas. So, if online backup with interactive userid/password setting and encrypted transfer was really a leap at the time, it might be considered non-obvious. Software engineers hate, hate, hate this notion, but there you have it. I don’t know whether the idea was a leap at the time; certainly few were offering interactive signups for many services over the internet in 1993, but simply tacking interactive signup onto any service doesn’t seem earthshattering.
Again, it’s all about ideas versus actual engineering. The *idea* of encrypting data stored on a central server is really obvious. But the *method* of doing so might not be. Patents need to reward inventors for their work, not for their dreams. And that makes almost all software patents “too broad” because the Patent Office didn’t know what they were doing. Many advocates for getting rid of most software patents will point to one patent as an example FOR software patents: the RSA encryption algorithm. That’s not only something that wasn’t trivial, but it was a groundbreaking moment in cryptography history.
The line between “idea” and “invention” is a lot blurrier than people realize, so really this whole argument is about broadness. And that’s why we have to resort to terms like “trivial”. The notion of backing up data is trivial. Any software engineer that has had to worry about data loss has desired this solution. But to do it over a crappy modem with hardware from the early 90s is pretty difficult. We need to reward inventors for their hard work, not just the first computer nerd to run to a patent office with the patent of “like this thing in real life, but online” — which make up a huge number of the patents being enforced by patent trolls.
And yes, often technology outpaces invention. Imagine if in addition to the plane design, the Wright Brothers had invented a technology to cut wood to make it lighter but still maintain the integrity needed to fly. Then that patent would be rendered obsolete when lighter metals such as aluminum became feasible to use. In the crazy world of software patents, they would be granted a patent on all lightweight materials, and would’ve been able to sue the aluminum makers.
Patent law should be pragmatic, and help whatever industry needs it, and right now it isn’t helping software at all.
I think we’re generally agreeing. The question, for example, is whether Crawford disclosed the non-trivial method of getting encrypted data to a server (somehow I doubt it, but I haven’t read the 97 pages of the patent).
More generally though, it’s a hindsight issue. Every idea looks trivial in hindsight: if only the technology were better at the time, of course everyone would have done x, y, or z. The question is whether we should grant patents to the ideas that no one would have ever thought of at the time because the technology was not better. Crawford is not a good example, I don’t think, given the prevalence of virtual drives. But that’s a key question. I think there can be such non-obvious ideas that come about before their time – but many other people don’t.
This leads to your plane example. Patent law has always allowed for expansion into new materials/technologies that were unforeseen. If the Wright Bros. get a patent on a plane weighing less than x pounds at a time when the only way to do that is with thin wood, they get to enforce that patent against later aluminum planes weighing less than x pounds, That’s how enablement works – they call it the “temporal paradox” for a reason.
This is one reason why there’s often not a meeting of the minds when people discuss software. The concepts that drive software folks crazy are ones we have allowed in non-software since the beginning of patent law. Maybe it was wrong to allow it before. Maybe it just can’t apply to software. But it’s not new to software.
Comments are closed.