It is probably fair to say that Matt Levy’s views are contrary to mine with respect to many patent related matters, but particularly so with respect to software patents. It, therefore, probably comes as a shock to at least some that we have a very cordial and friendly relationship. We chat whenever we are attending the same industry functions, we get together for lunch periodically in DC, we talk via telephone, and we swap e-mails.
Until recently Levy, who is also the author of Patent Progress, was patent counsel for the Computer & Communications Industry Association. However, Friday, April 14, 2017, was his last day at CCIA. Levy will now join Wiley Rein in the firm’s Washington, DC office.
One of the things I have always liked about Levy is that while we do disagree about much, we do so in a way that is not disagreeable. I also greatly appreciate how he always tries to work with me to figure out at what point in our analysis we go separate ways, which leads to some very interesting and lively conversations (as you might imagine). By doing that, however, we also discover just how much we actually agree, which sometimes is surprising.
Recently we were swapping e-mails and something came up that I found particularly intriguing. I told him I thought we really need to take this on the record and he agreed. What follows is our conversation, which took place on January 23, 2017. The publication of this interview has been delayed for several reasons.
Before proceeding, a disclaimer: The views expressed are Levy’s, who was speaking for himself and not for the CCIA, any of its member companies. That out of the way, without further ado, here is my interview with Matt Levy.
QUINN: Thanks, Matt, for taking the time to chat with me today I really appreciate it.
LEVY: It’s my pleasure.
QUINN: The reason I wanted to reach out and have this conversation was because in a recent email chain you and I were going back and forth and you said that you think we probably agree that innovation is key or critical to the economy or economic growth, and you said that we probably would agree that patents are critical for protecting innovations. But you said that you thought where we may disagree is on whether patents drive or follow innovation. And we can talk about any of this but that’s really why I wanted to go on the record with you, to have a conversation about that point in particular, because I have mixed feelings about that. I think in some instances patents drive innovation, and in other instances patents may follow, but I really want to get your thoughts and let’s just have the conversation go where it may.
LEVY: Sure. I’ve been thinking about that since I wrote it; I think I do believe at least at some fundamental level innovation comes first. I think that human beings are truly inventive, creative beings who’ve invented everything from fire and tools up to writing and all the amazing inventions that human beings have come up with long before the modern computer age. Some things are really incredible if you stop and think about how people came up with them. So people are inherently creative. But there is some point at which even if someone innovates they’re not going to have the drive to do something with it on a larger scale without some sort of protection. And I think that’s where something like patents come in is to enable the development on a larger scale. There may be follow on innovations that happen along the way that arguably would not have happened without the patents. So it’s probably a little too glib to say that all innovation always comes before patenting. I think there may be some — there are probably innovations that come afterwards. But I think at its core innovation comes from human beings’ inherent creativity and the patents really are there to let us do something with those innovations on a larger scale and get them out to people. I think it’s safe to say that in many industries without patents we wouldn’t have the development of products that we do. Pharmaceuticals are a classic example but they’re not the only industry that benefit from patents.
QUINN: That’s right.
LEVY: I guess that’s my starting point.
QUINN: Yes, and there are probably many industries that do, many innovations and many innovators that do. I wonder – this is fundamentally a very difficult question – I suppose in many regards how you feel about this probably dictates to some extent which side of the aisle you line up on when people are saying are you pro-patent or are you more anti-patent. And such is probably not a fair characterization as I’ve come to learn in my conversations with you. You’re a pro patent guy in some real large respect, although I think the perception of you is that you’re against patents. And maybe before we go any deeper let’s address that issue right here. How would you characterize yourself?
LEVY: I think I am pro appropriate patents. I absolutely believe that patents are a proper tool to use to protect innovations where there’s a risk of copying. I don’t want to be too simplistic about it. I think that there are a lot of very positive aspects to patents. The problem in my mind comes when we allow patenting that’s just much too broad. And this is really because of my background in software. If someone patents something at too high a level, too abstract a level, they end up capturing way more than they invented. And so they end up being able to block other people from doing things that they should be able to do. This is really where the software industry, programmers and developers especially, start clashing with the patent system; the way the software industry developed from its early days was through sharing and reusing code. There are even books that describe what are called design patterns, which are these high level patterns that you implement in code. They might give some low level code to implement a few things but generally they’re describing at a high level data structures and algorithms; the implementations are using pretty standard techniques. If we allow people to start patenting things like design patterns, programmers get heartburn and start freaking out, because there’s no way that the person invented every possible use of that design pattern. So patents do tend to clash with software, and that’s where a lot of tension comes in. But I can absolutely see the benefit of patents for many industries.
Pharmaceuticals, as I said, is a perfect example, but there are plenty of other industries. I was watching a show a year or two ago; I was on the road and I was watching some show on household inventions. The host would find small inventors who would come up with a useful household product and then they would pick a couple of them to pitch to an investor, who would then do something with it if they liked it. The winning inventor that night was this guy who had developed a special rim for a glass. The idea was that it was intended originally for a wineglass; it was some pattern in the rim, it was very small but apparently it stimulated your tongue in a way that made the wine less bitter. And he actually had solid patents on the idea. He had spent a ton of his own money on it; the investor decided to go with it because the way he had written the patents allowed them to invest not only in the glasses but to look at other — he had patented some other possible uses of it with other cups. And I was looking at this thinking, well, this is perfect, this is exactly what patents are for. The guy had a brand new idea. He invested time and money into it. It would be something that would be very easy for other people to copy if they saw it, so it’s perfectly appropriate. The patents here let the investors take the risk of developing a new product. Now whether it ever went anywhere I don’t know. I don’t know that I’ve ever seen the glass on the market but still that’s the basic idea of patents, right? It’s not a reward so much as it’s a chance to let the inventor invest in the invention without fear of all that money and effort just going to waste because somebody comes and rips them off.
QUINN: I guess in some respects I wonder how you wouldn’t characterize it as a reward. It feels like a reward in some important ways; a reward for coming forward with the innovation. And just to circle back for a minute before we go down that path with respect to software, which I suspect will take us down a long path. I think we disagree about this to some extent, maybe some very large extent, because I think software should be patented. I think in a lot of respects too much got patented and too many people who didn’t really understand software technology were both representing inventors and were trying to be inventors.
QUINN: And I think you see some – let’s just say very thin specifications. Any time I do a patent search I can almost guarantee that the first time an inventor tells me what their software is I’m going to find it. And if you read them, and I read these things all the time, and most of what I’m reading even from three, four years ago would have almost no chance of being issued if it was filed today. So much more is required. But one of the frustrating things from my perspective dealing with those people who are programmers is usually when they’re talking about patents, and being stopped by patents, the start and end of what they look at is the title of the invention. And they throw their hands up and say oh, how could this be patented even with good patents. And that becomes frustrating. And maybe it’s frustrating simply because the patent system has gotten so complicated that unless you have a PhD in patents you can’t understand what is actually being protected, which is maybe another conversation for another day. But I do think that if you have a software invention that really is an invention and you can really describe it and you explain how to bring it into being from a technical standpoint that should be treated no different than any other invention, which is why I personally am uncomfortable with using 101. I think we have to be using 112.
LEVY: Well, so I agree with the first part of what you said, which is that assuming that there’s something technical here and you describe how to do it and how to bring it into being — I think I wrote this on your site, actually — then I do think that there’s a space for software to be patentable in those circumstances. I disagree as far as whether it’s 101 or 112. So the 112 versus 101 debate is interesting and I’ve had it a number of times and I’ve thought about it quite a bit. I actually think if we started from scratch what we might do is essentially require software or computer implemented inventions to come under means plus function almost automatically so that we just don’t even worry about it. You have to put in the implementation details. You’re going to claim it functionally because that’s actually the clearest way to claim it and if we don’t have to worry about whether you’ve given enough detail to explain how it works, then we’re done. I think part of the problem is that means plus function has not been interpreted properly in my opinion, because that’s been applied so loosely. Now that may be changing, I certainly hope so, but there was a long period where unless you actually literally used the word “means” it was not going to be interpreted as means plus function even though it was supposedly only a presumption. It was a presumption that was never rebutted as far as I could tell in Federal Circuit case law. There were a number of years where it was 100% if that word “means” was there it was means plus function. If “means” was not there it was not means plus function. And that’s problematic for software, because you often have a module for doing something, or a component for doing something. And if I can’t go buy that component, if I can’t look at it and say okay I know what to go buy to do that then that’s probably not enough structure. Again, I’m over simplifying because we’re just having a discussion here. But that’s the basic idea.
QUINN: I don’t know that you are. And maybe I’m going to shock you because I’m going to disagree with you, but I think I’m going to disagree going in a different direction that you might expect. I personally would think the law should be that if you do not use means plus function language or any terms that are legally equivalent then it should not be means plus function. But — but I think when you’re using means plus function what you as the writer have to be doing is thinking to yourself that I am intentionally invoking this treatment and therefore by using language I am reading in the technology that I’ve disclosed into the specification into my claim. So if you don’t use means plus function then you’re not doing that unless you have gone through the painstaking process to define things and then you’re pulling in a definition the way that you would in any other kind of patent, which in the software space is maybe a little bit tougher to do if you’re not using means plus function language. I always tell people that I think you want to write your specifications to support use of means plus function claims, you know, write your specifications so that you could use means plus function claims even if you don’t. That way you have that information there, which will necessarily create a specification with a thick technology discussion.
LEVY: Right. I think that’s good advice. So what I was going to say is that I think that what has happened, at least historically and I’m hopeful that this is changing, is that many people, applicants would — well, patent attorneys let’s be honest — would draft language that was basically functional but not invoke the means plus function language by not using the word “means”; then they manage to persuade an examiner essentially that it was means plus function but it’s not in the record. That way they can then argue that the claim is a much broader version and it’s not means plus function. Of course, the judge following Federal Circuit law says, oh, well the word “means” isn’t there and you didn’t clearly say that it was means plus function so therefore it’s not. And we get these absurdly broad claims this way. That’s the root of a lot of this problem, the whole 101 versus 112 debate.
QUINN: It is. But you know, I wonder who are these examiners because so many of us don’t deal with the examiners that –
LEVY: It has changed. Yeah, but this has changed over time so that now the office trains them to be more aggressive in just forcing applicants to be clear about whether they’re invoking means plus function or not. I think that may be a huge part of the solution: just require it to be in the record whether these claims are specifically getting means plus function treatment and which ones are not. That makes it easier for the examiner, makes it easier for people later once the patent issues. This is probably another topic for another conversation but clarity of the record would solve a lot of the problems that you and I are talking about, just making sure that we can look at a patent and easily know, oh, okay, that’s means plus function even though they didn’t use the word “means” so I don’t have to look outside the spec to figure out what this means. It should be right there.
QUINN: Right. But you know the thing is there has been this cottage industry for a very long time of getting the broadest claims you can get and patting yourself on the back for getting ridiculously broad claims.
LEVY: You and I agree on that. Yeah, you and I agree on that completely. I think it’s bad for your client, I think it’s bad for the patent system. You should be getting a patent to protect what you invented and reasonable variants. It shouldn’t be that someone can just make a little tiny tweak and get around the patent if there’s something fundamental there. You should be getting basic some sort of reasonable equivalents, and what those are obviously depends on the field and the invention itself. But otherwise you should get what you invented and not a lot more.
QUINN: Right. And I see some of these claims and they’re so broad and it does kind of make you wonder. And then by the same token I see some claims that are getting shot down as patent ineligible and they don’t seem really particularly broad to me. They don’t seem particularly abstract to me and I don’t even know how you could begin to know whether they’re abstract because — and I guess I’ve struggled to articulate this and I think I may have recently here figured out what’s really bothering me in a way that I can articulate it, maybe in an elevator pitch. So let me try it on you. These 101 decisions are being made without any prior art but the Alice/Mayo framework forces the decision maker to decide whether or not there’s something inventive and to see whether there is only that which is conventional, because that which is conventional can’t contribute to patent eligibility.
QUINN: But once you start going down that path now you’re considering things that have historically come up only under 102 and 103, but now we’re not even applying any prior art. We’re just subjectively determining whether or not it’s conventional or it’s not conventional and I think that that’s a huge mistake. I think that question has to come up under 102, 103 and then under 112 we ask if it is actually described properly. Otherwise we are left with a test that is nothing more than subjective, which is why all the previous tests in the software industry have ultimately collapsed under their own weight. Once judges came to the realization that look you seem to be against these types of patents so every decision you make using the same legal standard that the rest of use is a negative. And I use the same legal standard that you use and every decision that I make is that they are patent eligible. That’s really what ultimately killed the Freeman Walter Abele test.
LEVY: Right, right. Well, so you know an interesting point that you raise about the term “conventional means.” There’s a practical reason, as I’m sure you know, why 101 has taken on so much power: it’s because at the moment it’s the only real question of law that you can use to challenge the validity of a patent. Everything else is at best a mixed question of fact and law which means you’re going to have trouble arguably getting even summary judgement on it, much less judgment on the pleadings. It’s incredibly expensive if you have to wait that long and go through discovery.
QUINN: Right, I get that. The problem I have with that is you still have to interpret the claim. I mean, you’re deciding at the very beginning in these Alice/Mayo framework questions what is the claim directed to. Well, how can you have any hope of knowing what it’s directed to and whether or not it is directed to a judicial exception if you haven’t interpreted the claim? And I guess where this starts to really become extremely problematic, because admittedly some of these cases that started coming up over the initial couple of years people would look at that and nobody was getting real agitated about because the party that lost the patent shouldn’t be expected to be able to get that patented. So we kind of shrug and wait for another case. But when you see the PTAB say that an MRI machine is not patent eligible that’s when a lot of people say, whoa, whoa, whoa, hold the phone! It’s an MRI machine. And what they did there was to look into the body of the claim and say that it seems to them that the new feature was this classification system, which is software related, so therefore the claim is directed to this software related thing despite the fact that the claim is explicitly written to cover a machine that would include the software. But when you get to that level of dissecting the claim I don’t know you can in an honest way – and by that I mean honest to claim interpretation precedent – go through that analysis without looking at the specification in great detail and looking at the prosecution history. So in other words, I don’t know how you can do any of this in the more difficult cases at least without a thorough claim construction.
LEVY: Well, all right. So let me see if I can say back to you what I think your point is because I think I get it. So I think what you’re saying is that, look, maybe these simple cases you can do in the pleadings but for some of these more complicated ones really it’s just over simplification to try and do it without, at a minimum, some claim construction and possibly some genuine fact finding about what the state of the art was, what was known, what was truly conventional.
QUINN: So let me say one more thing, and throw one other key thought out there because I want you to talk about all of this if you can. I also fundamentally don’t understand how in so many of these cases, and you see the particularly at the PTAB where they will reverse an examiner on 102 and 103 and say, no, examiner, this is novel and nonobvious. But you know it’s patent ineligible because it’s an abstract idea. That just doesn’t compute for me. How could it be abstract to the point of infirmity if you did a 102 and 103 analysis?
LEVY: Yeah, but that’s not quite what they’re saying so I have to disagree with you there.
QUINN: Okay, go ahead.
LEVY: So when they reverse a 102 or 103 what they’re saying is the art that you have cited does not teach or anticipate or disclose this claim. So they’re not saying that it is novel and is non-obvious; they’re saying this art doesn’t work. That’s it. I mean, when they reverse it. So it is completely possible for art to not — for some particular cited pieces of art to not disclose the claim but the claim still be patent ineligible.
QUINN: Yeah, I still have a problem.
LEVY: Now, you have to look at particular cases to see whether the decision is wrong or right but I mean, that’s what they’re saying and I don’t have a problem with that as a general matter. Particular cases might be different but I think it’s perfectly possible.
CLICK HERE to CONTINUE READING… We will pick up our conversation with discussion about whether it is possible for a patent claim that has been evaluated under 102 and 103 to be so infirm that it is properly characterized as abstract. We also discuss whether innovation leads to patents, or the prospect of being able to obtain a patent incentivizes innovation.