I often ponder the importance of software patents in the relation to building a strong software company. I'm in the camp that believes that software patents are nearly useless in the practical sense and yet they are still somehow valuable.
Software startups try to get patents in order to show that they have a meritorious idea. Just putting in a provisional patent application is enough to show an investor that you are at least thinking about IP. Part of the patent application process is finding out if there is an existing patent already covering your idea, which can sometimes put an abrupt end to your entire idea if there are already a few infringing patents. The research portion is valuable in that you can truthfully answer the inevitable question of whether there is an existing patent and go into your venture with additional knowledge of potential competitors and technology owners.
One misconception about patents is that some people believe that they provide critical protection against a competitor trying to develop a similar technology. Patent litigation is almost always painful and wasteful. Most VCs, and their blogs, tell you that patents are just part of your company's mosaic of competitive defense. Superior product, customer service, sales and marketing are usually more effective at beating a competitor than patent litigation. Sales and marketing are especially important as it has been shown in the past that the superior product does not always win.
A company may use its patent portfolio as a sort of "mutually assured destruction" in case of litigation. Say you want to go after IBM for infringing on one of your patents. It may be advisable to be sure that you are not infringing on one of IBM's patents too. It would be very problematic if you went after IBM only to have IBM come after you for a completely different patent violation.
It can be shockingly expensive to get a patent. You can do it all yourself, and some people do, but most people need an attorney to do the application and convert you description of how your mechanism works into a legally defendable patent. Attorneys are always expensive compared to most other costs but consider that all the large companies have several attorneys on staff and they are writing patents too. You probably want an experienced eye and a person who understands the lingo to ensure that your patent is accepted and not invalidated should litigation ever become necessary.
A patent may go through several iterations before being accepted by the USPTO. Patent applications do get rejected to get further clarifications and that means more lawyer time. Hiring an attorney who has done many software patents may be expensive from an hourly perspective, but a quality patent may not get ever be rejected or need clarification through the review process and cost you less in the long run.
One founder told me he spent around $50K for a single patent. I took this as an exaggeration, but that is a big number. Other attorneys may be less expensive and just as good. I just hope for $50k that he invented something really interesting.
An interesting note is that you can hire an attorney located anywhere. You may want to hire a local attorney but there is no actual requirement to do so. You could be located in Vermont and your patent attorney can be in Texas.
A second component of cost are the international patents. Yes, you also have to submit applications to foreign nations. Every country has a different fee structure for patents and it can be prohibitively expensive to get them all.
One thing to note is that your patent is describing how something works, not why. I can't speak for you, but I find it annoying to only describe how something works. Knowing how something works is obviously important but knowing why it works is what give the reader the ability to improvise and evolve the technology. It's sort of like being told about how C++ class hierarchies work without being told why you should use them. A smart person could probably figure it out, or rather, many smart people have figured it out, but it took a lot of effort to gain that knowledge when it could have been more efficiently conveyed.
An oddity about patents is that they are supposed to protect something novel. I often think that most software patents are not really brilliant. Yes, they represent hard work and that is of merit alone. I remember a patent given the static instrumentation of machine code called "Method for inserting new machine instructions into preexisting machine code to monitor preexisting machine access to memory" by Reed Hastings. Reed came up with an innovative way to monitor memory access in an application by statically adding additional instructions for all memory allocation, deallocation and memory accessing code. This patent makes it possible to track when your application reads or writes beyond the boundaries of a given memory allocation. Its a cool patent and its one that most people wouldn't have figured out without a lot of work and real thought. I worked for a competitor and we had an equally interesting patent that dealt with doing the instrumentation at run-time instead. One thing interesting to note: Reed is the founder of Netflix. Only an engineer could optimize the movie rental process by putting in a queue and automating the entire process so I no longer had to interact with video store employees.
A lot of software patents are not this brilliant. To me, almost all email and instant messaging patents for mobile devices are dumb. Mobile devices like cell phones and iPods are arguably more powerful than the early 8-bit 8088 processors that had network access. You can go back further than that if you consider modem and smart terminal protocols. It should be obvious to pretty much anyone that if you can get a smart terminal to transfer data that you could also make a modern cell phone, with more memory and a vastly superior network connection, do the same or more work.
Any developer can spend 6 months and develop a new protocol and a system for handling email or chat too. While there may be some hard work, its still only email on a pretty powerful platform. There aren't many ways to make email technology remarkably better than it already is. However, many of these patents are for marginally new technologies that can use email. A cell phone with email is not novel or inventive because, in my opinion, it is an obvious extension of a networked device and has been done previously on far more primitive devices. Its clearly not one of those things that takes months of research to see if it can be done, its something that can obviously be done. Yet we still have patents for email on mobile devices. I'm sure there are dozens of more egregious examples of an obvious technology, or application of technology, that was patented.
The patent "troll" has become the enemy of the software developer in this regard. The patent troll starts a company and buys up as many patents as they can and then looks for violators and sues them for licensing fees or blocks the product entirely. It really doesn't matter if we agree with the practice or not, the US legal system allows it.
Some people would argue that some engineers are just patent trolls too because they don't actually build anything with the patent. There are engineers that own interesting patents on things like membrane keyboards and collect licensing fees but they don't actually build the device. It would seem that the difference is only that the troll bought the patent rather than inventing the technology. Clearly, being the inventor is seen as respectable and probably rightly so.
A little known fact about litigation is that trial attorneys turn down cases if they don't believe it to be winnable. We often times think that trial lawyers are there for the money only and that they are happy to take any case. That's not true at all. They take cases that they think they can win or settle and at least break even. Remember that attorneys have staff, office space and a variety of other expenses that they pay just like the rest of us. Trying any case is expensive for them so there is a lot of reason to only go after cases you can win. Granted, there are some attorneys that will try anything with the hopes of making a name for themselves and thats their choice.
The problems of the patent system are often blamed on big companies like IBM and Microsoft for pushing out such huge quantities of patents, on companies that put request and get obvious patents, on patent trolls for demanding licensing fees for things they don't use but own and the attorneys that go after perceived infringers. I, like many other people, feel that the patent system needs to be reformed such that obvious patents can't make it through the system. I believe that real innovation that leads to significant competitive advantage and needs to be protected. Unfortunately, it is often difficult to quantify the difference between obvious and brilliant because a person may see all patents in a narrow area as being obvious.
An engineer specializing in compiler design may see all or many compiler related patents as obvious. A couple of patents a year may be seen as truly brilliant by her. However, her peers may disagree and think that different patents are more meritorious. Is it shocking then that the bar for getting a patent is low? If an original kind of screw (see Ball Screw Device) can be patented then why not a system for transferring email to a cellphone?
One could argue that perhaps the ball screw patent is in fact a very specific device and not particularly broad. This is coming from possibly the least handy engineer that ever walked the earth, though I did successfully calibrate the laser sight on my wife's miter saw. I have to ask myself whether I am interpreting the patent as truly innovative as it is possible or even likely that a bright mechanical engineer will say that its obvious. So how do we reform the patent system when creativity is being judged by people who are actually subjectively determining if a patent application is actually valid?
One camp argues that we should voluntarily stop submitting software patents. The problem is that many companies are not going to stop making patents. They need it as a part of the aforementioned legal defense strategy too. Should you be the first one to stop and to hobble yourself?
Having no patents in your company's portfolio may be fine depending on what you are doing. However, an angel or a VC investor is not going to complain if you have one or more patents or provisional patents. Having patents is still a good thing to have in your pocket. Its sort of like having a bachelor vs. an associates degree. Having paying customers of course puts you in the masters degree category.
I just mentioned the provisional patent and that's a handy tool in figuring out what you want to do concerning a possible patent. Take a look at this link http://www.uspto.gov/web/offices/pac/doc/general/index.html#prov to get some of the details. The idea is that you can put in an application that doesn't go through the review process but it gets time stamped. Here's an important line from the link: "Provisional application provides the means to establish an early effective filing date in a patent application and permits the term "Patent Pending" to be applied in connection with the invention." Provisional patents start the clock on your patent and marks the date that it was originally submitted. Being the first patent for an invention is important. There's some fellow out there that invented the telephone before Alexander Graham Bell, but his patent application got filed just a little later than Bell's.
Provisional patents don't have to be complete either. About.com has this link that explains what can be missing . The big value is that you can describe your mechanism quickly and get it into the system before someone else does and then you have a year to put together the complete patent application or to drop the invention entirely.
You may decide to give up on the invention for a number of reasons. One is that you find out that someone has done exactly the same thing already in some publicly accessible form such as a book, blog, magazine, OSS project or another existing patent. The publication creates the notion of prior art and invalidates your patent. No need to throw good money away on something that is not going to get approved or can be easily invalidated. Another reason is that you may not have the money to get the patent. Its hard to tell what will happen in the course of a year. You may get a lot of customers for your product or big investments or you may go bankrupt. No need to put in the cash for getting a patent if your company will sink before the patent gets approved.
You may want to get an attorney nonetheless to check with her about concerns of overlapping patents and how you can avoid a patent issue entirely. Avoiding an infringement may be as easy as changing a couple of simple operations. It could be more complex than a few simple changes too. The attorney is going to ask you if you think the patent infringes, and you probably do if you are calling, but she can explain more about the criteria for patent infringements. The difficult thing to comprehend is that you are the technology expert and the attorney is the legal expert. You can tell if the patent describes the technology you invented and she can tell you what the existing patent is claiming specifically. The patents claims may not match up to your idea.
The problem with any patent application is that it is expensive and it is questionable how much value the patent actually brings to your organization. They certainly aren't the shield that you hoped they are as the expenses of litigation are high and can sometimes cause more harm than good. It appears that patents just get relegated to an entry in a list of check-boxes, which is fine if that patent check-box was the last reservation that an investor or a customer had.
My own opinion on reform is that they probably should be legislated away or the pool of reviewers needs to be greatly expanded and improved. However, I have never actually met a patent reviewer for software patents. What if they already are a bright set of individuals well versed in software development and algorithms? What if the proposed public peer review scheme really works and patent quality shoots up as well as review time dropping significantly? Wouldn't software patents be good then?
0 comments:
Post a Comment