Jim Moore’s blog: Innovation, Strategy, Public Policy

Students, inventors, innovators of all kinds learn by studying how previous innovations were done.  What the patent system does is assure that this information–how things work–is “open sourced” for all to learn from. In order to get a patent you must disclose how your idea works “so that someone skilled in the ordinary art” of your field can duplicate it.  That is, you have to give up your trade secrets.  This requirement, by the way, is why many companies do not file software patents–they don’t want to share their ideas.

When a software patent application is filed, 18 months later it is made public. Typically this is 2 years before the application will even be considered by a patent examiner.  This means that ideas are in the public sphere, the commons, for all to learn from well before any property rights have even begun to be considered for the inventor.

Most patents are rejected by the patent office.  Months ensue, and typically if any patent is attained it is for a much narrower scope than originally asked for by the inventor.  Meanwhile the key ideas have now been in the public domain for typically about 5 years–all the time fueling innovation.

Patents drive innovation because they spur inventors to try to work around and improve existing technology.  Unpatented ideas often become the basis for popularization only because they are free, and thus stifle innovation.

Consider an analogous case from another field: Why do you think the King James Bible is the most published?  Is it the most current or accurate or scholarly?  No.  Hardly.  The King James is long out of copyright. Publishers make high profits from publishing and promoting it.

Why is Apache the widest used web server?  Is it the best possible web server? No.  Hardly. It is the cheapest.

Same with Linux.  What is Linux, but a clone of Unix.  A commodity. Why is it popular? Because it is best? No. Because it is cheapest.

Have Apache and Linux spurred innnovation in operating systems? No, they have dampened innovation by making it nearly impossible for any company to make money in operating systems, and thus afford R&D investment.  Sun was the company innovating in the Unix tradition. What has happened to Sun?  It has been wiped out by Linux-based competitors, and become a notably unsuccessful and non-innovative open source company.  When is the last time you looked to Sun for leading edge innovation?  In the late 80s, before the commoditization of Unix/Linux.

Ironically, the best spur to innovation in operating systems is Microsoft Windows. Why?  Because customers are willing to pay to use it, and thus prospective competitors know they might be able to sell an innovative contender.  Who has taken advantage of the Microsoft Windows opportunity?  Apple, by investing R&D in a highly innovative advanced version of the Unix/Mach lineage, building on user interface ideas from Xerox PARC innovation.  Who has the best operating system for small computers?  Apple.

Thomas Jefferson invented the American patent system as a way to provide incentives to inventors to open up their trade secrets.  The quid pro quo was that inventors who would share their ideas got a limited-time property right to their inventions.  This has worked brilliantly in the US for 200 years.

Don’t believe the anti-patent PR. Note what companies are behind it:  IBM, with control of 45,000 patents, including 37000 directly owned in its name, and 3000 patents filed newly each year.  And Intel, whose CFO coined the term “patent troll.”  The celebrity lawyers who argue against patents, such as Larry Lessig, work for centers funded by IBM and other large patent holders.  These lawyers are not, notably technologists.  By contrast, look to real, pioneering technologists, such as Bill Joy or Danny Hillis. They are, wisely I think, supporters of patents.  Hmmm.  Think about it.

1. What scares big companies is disruptive innovaton carried out by individuals and small firms. There are no scale effects in innovation. Bach, Mozart were individuals. Larry Ellison of Oracle famously says that software technology innovation can no longer be accomplished in large companies. This is why Oracle acquires companies right and left. Microsoft spends billions of dollars a year, and only comes up with incremental improvements.

The same is broadly true in other fields. Most biotech innovation happens in universities or small labs. Even Apple, the famous innovator, is highly dependent on small outside design firms, and a few individuals carefully nurtured in the firm, for both look-and-feel and technology innovation.

The US as a whole needs innovation.

The US needs to nurture innovative individuals.

2. The following analysis shows that strong patents are pro-innovation. I have excerpted a small bit, but please click to the full article. From Science magazine.

The author points out that patents of building blocks promoted innovation across history and across industries.

3. This puts to lie the arguments supporting a bill that weakens patent law, currently being pushed by Patrick Leahey of Vermont.

Efforts to weaken patent law come almost entirely from two groups. First, the lobbyists for very large, traditional firms. In information technology, it is the established “gorillas” like IBM, Microsoft, Google, and Cisco. Second, from policy wonks who are funded by the big companies.

These big gorilla techs named above are today’s version of the big auto companies, big plastics companies and big pharma companie, each of which tried to quash patents as they tightened control over their industries.

Big companies dominate their industries by economies of scale, by monopolies and oligopolies on distribution, and on stopping new entrants by “killing the snake while it is small” (or buying the snake when it is mid-sized).

In autos, they succeeded. In plastics and pharma they did not. In materials and in biotech there are vibrant communities of independent think tanks and innovation laboratories that make their living licensing patented inventions to larger firms or directly to the public.

The information technology industry looks more like the early auto company days than it does the early biotech days. This is because small companies in information technology have not figured out the value of their intellectual property, and how to protect it.

Big companies in information technology depend on squashing small companies, buying intellectual property on the cheap from small companies, and buying pioneering companies on the cheap. They can easily do this when the small companies either don’t protect their intellectual property (a problem in many small software companies) or open source their technology in unwise ways. Big companies are now trying to weaken the patent system before the small companies wake up to its value.

Who is taking the money from these big companies to do their bidding? Senator Patrick Leahey of Vermont–for whom John Chambers, CEO of Cisco, held a big fundraiser in Vermont. Why did Chambers, who lives in Silicon Valley, come to Vermont to do a fundraiser. You connect the dots.

Leahey is now aggressively pushing changes to the patent system that make it harder for small companies to succeed, and easier for big companies to triumph.

Leahey has, to put it bluntly, sold out.

He probably figured no one would notice. This is typically how Senators agree to sell out. They pick an issue that is important to a group of wealthy companies, and try to slip changes in unnoticed.

Shame on your Senator Leahey.

BIOTECHNOLOGY: The Patenting of DNA — Doll 280 (5364): 689 — Science
..in the USPTO’s view, new areas of technology do not create the need for a whole new specialized patent law. In many ways, the arguments currently being used for DNA sequence technology resemble those voiced 30 to 40 years ago when polymer chemistry was an emerging technology. At that time, people argued that if broad generic claims were granted on the building blocks of basic polymers, it would devastate the industry. In fact, no such disaster occurred. For example, the issuing in 1965 of a basic patent broadly claiming a vulcanizable copolymer of aliphatic mono-olefins and unsaturated bridged-ring hydrocarbons (3) did not preclude the later issuing of patents to different inventors for several copolymers of this type (4). These patents represent early examples of ethylene-propylene-diene monomer (EPDM) rubbers, which are highly weather- and ozone-resistant, stable to thermal aging, and have good electrical insulating properties. These EPDM rubbers have been commercially important as components in tires, weather stripping, radiator hoses, wire insulation, impact modifiers, and roofing.

EPDM copolymers were assembled from three basic building blocks that could be combined in many different ways and, as such, generic and specific claims to these copolymers are analogous to claims that may be issued to DNA inventions. Just as the issuing of broad product claims at the early stages of this technology did not deter development of other new vulcanizable copolymers, the issuing of relatively broad claims in genomic technology should not deter inventions in genomics. Two relevant examples of this in the field of biotechnology are the polymerase chain reaction (PCR) and the human immunodeficiency virus (HIV) protease, which were patented and then widely licensed to permit the biotech industry to continue to grow and benefit from these inventions.

The same patentability analysis is conducted for every patent application, regardless of whether the application is for a computer chip, a mechanical apparatus, a pharmaceutical, or a piece of DNA. In every field of technology–whether emerging, complex, or competitive–all the conditions for patentability (such as statutory subject matter utility, enablement, written description, novelty, and non-obviousness) must be met before a claim is allowed (5).

Economic research on patent licensing by industry demonstrates a maturity curve. Disputes over intellectual property are comparatively rare in industries where intellectual property ownership has long been accepted, such as semiconductors, electronics, telecommunications and biotechnology. Intellectual property licensing is well-understood and actively pursued among companes. While there is inevitable grumbling about prices, generally-acceptable licensing terms have been established in these industries. Most important, there is a vibrant sub-sector of each of these industries where “invention companies” thrive as part of the ecosystem. I have a loft near MIT on Brookline Street, and all around me are biotech idea companies living happy, comfortable lives tackling hard intellectual problems, developing inventions and intellectual property, and licensing the resulting technology solutions to larger industry players who provide downstream development, testing, marketing and distribution.

By contrast in software intellectual property practices are strikingly immature. This can be hard for outsiders to understand, as there is very little real difference between software inventions and analogous inventions in semiconductors, electronics, telecom and biotech.

Why is the intellectual property ethos so immature in the software industry? Mostly because software as a distinct field was so abstract as to be almost invisible until the 1980s. In the academic engineering world, software started out in applied mathematics and computer science as a field came relatively recently. In business, software was seen as a necessary part of hardware, not as a distinct product or service offering. Indeed, up until the 1980s software was typically given away by hardware companies in order to sell their (big) iron. It took an anti-trust settlement with IBM regarding the IBM 360 mainframe to force IBM to unbundle software from hardware. In the minicomputer days, Sun Microsystems was one of the first companies to see software as the basis for a distinct business–an idea that was considered radical at the time. The layered architecture of the personal computer world finally enabled software companies to form a distinct, free-standing industry.

Software patents became recognized as a distinct, official domain of invention in the United States in the mid-1980s. Telecom patents by contrast went back to the beginning of the century.

In the long history of intellectual property, software in now coming of age.

In the software industry companies large and small continue to underemphasize intellectual property licensing. The rationalized ethos of cross-licensing among companies is barely developed. Instead, firms large and small continue to come to blows on the few propery conflicts that rise to high stakes. A myriad of opportunies for licensing and collaboration on intellectual property development are ignored, and–importantly–infringement is a daily business practice.

Slowly, however, the industry is growing up. IBM sets the right pattern in some ways. IBM has long offered to license pretty much any of its 34,000 owned and 45,000 owned-or-controlled patents for 1% of another company’s “relevant revenue” (i.e. revenue directly attributable to employing the patented intellectual property). IBM typically would offer its entire portfolio for 5% of revenues. For many years, other firms found this deal attractive.

Today IBM will license most of its software intellectual property portfolio for 1% of relevant revenues in total. The thinking seems to be that most firms can expect to be licensing software intellectual property from several others. Thus 5% is probably what a firm can reasonably budget for all of its relevant licenses, so IBM figures 1%, or 1 of 5 points, is about what the market will bear for its contributions. This seems a reasonable assumption to me. In other industries, this percentage is about the norm.

We are also seeing the beginnings of a distinction in the software industry between invention and distribution. IBM, interestingly, is becoming more of a distribution firm (a systems integrator), relying on others’ ideas–while continuing its own lab, as well. Google, on the other hand, appears to be investing heavily in ideas–while developing its distribution platform.

What I think will be a most interesting next development is the rise of smaller, entrepreneurial invention firms in software. I look forward to Brookline Street in Cambridge being lined with small software engineering firms solving the hardest problems on the planet–and partnering with larger firms to take the resulting inventions to the world market.