Software Patents

My friend Tim Lee makes the case against software patents in The New York Times. An old Bill Gates memo makes for a good framing device:

Microsoft sang a very different tune in 1991. In a memo to his senior executives, Bill Gates wrote, “If people had understood how patents would be granted when most of today’s ideas were invented, and had taken out patents, the industry would be at a complete standstill today.” Mr. Gates worried that “some large company will patent some obvious thing” and use the patent to “take as much of our profits as they want.”

Mr. Gates wrote his 1991 memo shortly after the courts began allowing patents on software in the 1980s. At the time Microsoft was a growing company challenging entrenched incumbents like I.B.M. and Novell. It had only eight patents to its name. Recognizing the threat to his company, Mr. Gates initiated an aggressive patenting program. Today Microsoft holds more than 6,000 patents.

It’s not surprising that Microsoft — now an entrenched incumbent — has had a change of heart. But Mr. Gates was right in 1991: patents are bad for the software industry. Nothing illustrates that better than the conflict between Verizon and Vonage.

Read the whole thing. I guess I'm not thrilled with the word choice around "bad for the software industry." Patents are bad for the development of new software. If you define "software industry" as "incumbent for profit software firms" it may be good for the "industry." The thing to keep in mind with any sort of IP protection is that strong IP creates, on the one hand, an incentive for innovation but at the same time it also creates a barrier to innovation. In the case of software patents, the balance tips overwhelmingly in the direction of creating barriers — indeed, the main incentive it creates is merely for the innovative production of patents rather than of actual products.

Comments

Heavy state protection of the status quo is why civilizations like China (100 AD to 1900 AD) withered and declined. If the fruits of innovation are confiscated by unproductive mandarins, then no one will bother to innovate.

Posted by Don Williams | June 9, 2007 11:20 AM

The thing to keep in mind with any sort of IP protection is that strong IP creates, on the one hand, an incentive for innovation but at the same time it also creates a barrier to innovation.

It's not so much as a barrier to innovation as a barrier to (market) entry.

Arguably, patents spur innovation as competitors attempt to design their products around them. But a lot of this new innovation is in one sense wasted, because it is focused on avoiding another's IP not creating a product that is "improved" from the perspective of the consumer.

Posted by von | June 9, 2007 11:40 AM

Correction: Change above reference to Chinese history from "100 AD-1900 AD" to "1200 AD -1900AD". Things went to hell mainly after the duck-fornicating filthy Manchus took over.

Posted by Don Williams | June 9, 2007 12:09 PM

Frankly, I agree that software patents---at least in anything remotely like their current form---make absolutely no sense and are fairly destructive to innovation and economic progress.

I say this as someone who has filed three major software patent applications so far this year...

Posted by RKU | June 9, 2007 12:22 PM

First of all, there is no such thing as "IP". There are trademarks, trade secrets, copyrights, and patents. All very different, with differing levels of innovation vs. incentive balance.

Second, the software industry developed for 50 years (1945-1985, roughly) on an open, shared-knowledge basis. Many of the "new techniques" that corporations such as Microsoft are trying to patent have either been known since the 1950s or were developed using other techiques and tools that were developed on an open-knowledge basis. For these corporations to suddenly decide that they can lock up this knowledge using patents (in effect, stealing the workd of every software developer since 1945) is not only corrosive of the entire process of knowledge development, it is sickening and immoral.

Cranky

Posted by Cranky Observer | June 9, 2007 12:38 PM

I loathe software patents.

To be fair though, if one should be able to patent genes, cultivars, mechanical contrivances and so forth, then one should be able to patent software contrivances as well.

Consider the internal combustion engine. It received several patents. Over the years it gives way to software: software-based fuel injection replaced the carburetor in the 80's. Now software is set to supplant the camshaft as well.

I don't see how you can consider contrivances such as the carburetor of the camshaft worthy of patents but not their software equivalents.

Better to eliminate patents altogether.

Posted by tiny tim | June 9, 2007 12:42 PM

Mr. Gates worried that “some large company will patent some obvious thing” and use the patent to “take as much of our profits as they want.”

Earlier this year, AT&T/Lucent dusted off an old MP3 patent and won a $1.53 billion dollar verdict
against Microsoft, the largest patent infringement award ever.

The infringing code is the backup MP3 encoder on Windows media player, but damages were calculated from the retail sale price of every computer sold with Windows on it That adds up fast.

Posted by Trevor | June 9, 2007 12:42 PM

> I don't see how you can consider
> contrivances such as the carburetor
> of the camshaft worthy of patents but
> not their software equivalents.

Legally and practically speaking, because obviousness and prior art are almost impossible for the Patent Office to determine. The PO has a hard enough time keeping mechanical engineers (and MEs with law degress) on staff, but that is at least mitigated by obviousness being fairly easy for a person of mechanical bent to determine.

To judge software for obviousness and prior art would require the Patent Office to keep on staff people who have the following set of qualifications: better-than-competent computer programmers; knowledge of the history of software stretching back to 1885; experience and creativity to judge obviousness; preferably legal training as well. Such people are in very short supply and are typically paid in the $250k-$1M range in Silicon Valley. History shows that the Patent Office has /never/ been able to attract or maintain any significant number of people on staff, and as a result has issued some unbelievable whoppers in terms of prior art alone.

Morally and ethically speaking, that "new" software was undoubtedly created using knowledge and techniques developed during the era of open software knowledge (1945-1985 roughly). In fact there is a very good chance that the embedded software you described was developed using the GNU C compiler (GCC). Read the history of GNU and the GCC and then explain to me the resulting product is 100% independent of the open work that came before it.

Cranky

Posted by Cranky Observer | June 9, 2007 1:19 PM

It' worth mentioning that legal scholarship has been on this case for a while, and courts are starting to take notice. The recent Supreme Curt decision in KSR v. Teleflex (pdf link to opinion) makes it much harder to patent "obvious" combinations of pre-existing elements, or at least harder for those patents to pass muster if challenged in court. I'd expect the ruling to have a disproportionately large effect on the software industry, where it has been too easy to patent fairly straightforward improvents on existing methods. For example, a lot of people hold patents for "well-known process X, but on the internet!" that probably aren't worth a whole lot anymore.

Posted by Trevor | June 9, 2007 2:05 PM

Actually, I'm more bothered by software copyright laws. Under the current law, you can write a program, copyright the object (i.e. compiled for computer) code, which is gobbledygook that nobody can understand, while keeping the source code hidden away as a trade secret. How is progress even remotely served by this approach? All I can say is, thanks for the DMCA, Mr. Clinton!

Posted by Adam D | June 9, 2007 2:09 PM

While patents can clearly be used in ways that inhibit innovation, it's important to remember that they do serve an important purpose. Without IP protection, innovative startups with unique technology could do nothing to prevent the behemoths (MSFT, IBM, etc.) from simply duplicating the concept (with far more resources). Ultimately, without patents, we would have far fewer entrepreneurial ventures and far more domination by the big guys. Patents can at times stifle innovation, but on balance, I think they have produced more than they have taken away.

Posted by MP | June 9, 2007 3:34 PM

While patents can clearly be used in ways that inhibit innovation, it's important to remember that they do serve an important purpose. Without IP protection, innovative startups with unique technology could do nothing to prevent the behemoths (MSFT, IBM, etc.) from simply duplicating the concept (with far more resources). Ultimately, without patents, we would have far fewer entrepreneurial ventures and far more domination by the big guys. Patents can at times stifle innovation, but on balance, I think they have produced more than they have taken away.

Posted by MP | June 9, 2007 3:35 PM

The thing that gets forgotten in this discussion is that IP protection, including software patents, is one of the most important factors VCs look at when deciding whether or not to finance entrepreneurs. Don't kid yourselves, the real barrier to entry for start-ups is financing.

The "problem" of software patents largely arose after the dot com crash when companies tuned to litigating their IP portfolios to provide some return to their investors. But how is that a bad thing? The fact that Microsoft gets slapped for a couple hundred million now and then is a small price for society to pay to make it easier for entrepreneurs in the software industry to get financed.

The primary effect of the Supreme Court's recent decisions regarding patent law (AT&T, Ebay, KSR) is to decrease the value of software patent portfolios. And that will have an adverse effect on the willingness of investors to bet on software start-ups.

Posted by cj | June 9, 2007 3:57 PM

"I'm not thrilled with the word choice around "bad for the software industry."

Thanks for making this point. I'm always frustrated when people talk about what is "good for industry," or talk about "balancing" the needs of consumers with the needs of business. The needs of businesses as such should get exactly zero weight.

On the other hand, I'm usually annoyed by people that use the idea of "protecting consumers" to promote policies that I believe ultimately will hurt consumers, by making businesses spend less energy on innovation and meeting user needs, and more energy on complying with complicated regulations and avoiding legal problems.

Posted by ed | June 9, 2007 4:05 PM

> The primary effect of the Supreme Court's
> recent decisions regarding patent law (AT&T,
> Ebay, KSR) is to decrease the value of software
> patent portfolios. And that will have an adverse
> effect on the willingness of investors to bet on
> software start-ups.

Except that there is very little evidence that any world-changing or even very good software came out of the dotcom boom. Most of it has been thrown away by now. That which remains was primarily based on Apache, PHP, and Linux(tm) which are Open Source(tm).

Prior to the dotcom boom, the invasion of venture capitalists, and the hunger for the "flip" software entreprenurs had to do it the old-fashioned way: carefully build a good product, develop a market for it, and grow the company step-by-step as did the founders of ASK, JDEdwards, and WordPerfect. Oracle and Google too for that matter.

And they protected their work as Adam D notes above with a combination of copyright on the source code (the copyright on the object code didn't mean much) and trade secrets. But the founders of WordPerfect never sued anyone for trying to build compatibility with the WordPerfect file format into other software. Whereas Microsoft is currently attempting to use patents to lock down the Word 2007 file format hard, to the point that in combination with the included Digital Restrictions Management and the DMCA you will no longer be the owner of documents you type in Word. Again it is a bit unclear to me how this helps society.

Cranky

Posted by Cranky Observer | June 9, 2007 4:25 PM

Here is my take on the op-ed which is posted at http://www.againstmonopoly.org/:

Timothy Lee writes "A Patent Lie" attacking software patents link here. The article makes several good points, ending up by arguing that copyright provides better and less expensive protection, inducing greater competition and lower costs to the consumer.

But why have any protection, other than the basic secrecy which allows the software writer to avoid publishing the code. If the idea behind the software, like Amazon's one-click ordering software, is so obvious, why protect it and prevent anyone else from producing software which serves the same purpose?

Why not reduce the wages of monopoly to an absolute minimum?

Posted by John Bennett | June 9, 2007 4:33 PM

Cranky, you're kidding, right?

The boom was about a lot more than just pets.com. There is no way that we would have had the build out of high-speed internet, online music, social networking sites, some of the more sophisticated web-development tools, or even the growth in Linux as a Windows alternative, without the dot com boom.

A lot of stupid crap came out of it too (and some of it even got patented), but you can't seriously contend that computer technology would be where it is today without what happened in the 90's.

BTW largely financed by VCs who investigated the underlying IP portfolio of these companies.

Posted by cj | June 9, 2007 4:58 PM

cj,
Having been involved with software and what is now known as the Internet since the 1970s, and having read widely in the history thereof, I will stand by my statement.

Cranky

Posted by Cranky Observer | June 9, 2007 6:37 PM

It's certainly true, cj, that the dotcom boom and the growth of the internet pushed hardware spending further then it would have gone in those years otherwise, and in that way the boom probably pushed computer hardware advancement further than it would have gone. However, 1) I'm not convinced hardware advanced all that much faster during those years than in previous, and 2) the biggest factor pushing hardware beyond its normal rate of advancement in those years was probably the appearance of AMD as a serious competitor to Intel.

Now if you're talking about software, I just don't see how patent or even copyright protections helped dotcoms advance the art. We've gotten much better at putting together web pages with database backends that scale pretty well, and we've gotten slowly better at working around the limitations of web browsers (while also slowly correcting those limitations), but this is mostly scut and kludge work, hardly the stuff of inventions and breakthroughs. The best I can think of is that Google and the other search providers have advanced the art of large-scale search, but they mostly protect their search products with secrecy and natural barriers to entry (e.g. it's not a trivial thing to duplicate Google's data farms), not with patent or copyright.

Posted by brian | June 9, 2007 7:30 PM

Re "However, 1) I'm not convinced hardware advanced all that much faster during those years [1990s] than in previous, and 2) the biggest factor pushing hardware beyond its normal rate of advancement in those years was probably the appearance of AMD as a serious competitor to Intel."
--------
I disagree. Much of the push behind the dotcom stuff was HARDWARE developments. Specifically, erebium-doped fiber optics allowed propagation of laser beams for thousands of miles without the need to convert the light back into electrical pulses every 100 miles or so in order to use slow electronic equipment to amplify the signal, convert it back into laser light etc. Very useful for transoceanic communications.

Plus the ability to put several laser beams of different colors (frequencies) on the same fiber.
A big deal when you want to send images or movies vice small text files.

It takes roughly 4-5 years for something in the lab to break out into commercial deployments but George Gilder, for example, was beating the drums over the Telecosm back around 1995.

Posted by Don Williams | June 9, 2007 8:19 PM

The interesting case to look at is not MS and MP3s but Qualcomm vs Broadcom, resulting in many new generation cell phones now unable to be sold in the US.

http://arstechnica.com/news.ars/post/20070607-itc-to-bar-import-of-new-handsets-in-patent-dustup.html

This seems like the sort of thing that could generate real political heat. The fact that it couldn't have happened to a nicer company in the patent field (hah!) just makes it all that much more interesting.

Posted by Maynard Handley | June 9, 2007 11:01 PM

Six points. The author, Lee, is clueless, but that isn’t a point.

Point 1: Most patents do not cover broad concepts. They address specific problems and offer incremental improvements. Claims of such patents are usually narrow. That makes them relatively easy to get.

Point 2: Lee’s article, and most posts above, confuse patent prosecution and patent litigation. During patent prosecution, a patent examiner usually doesn’t have the resources to uncover the prior art. The patent office is under funded, and examiners are overworked and under trained. They haven’t sufficient time to do a thorough search of the prior art. Yet a decision to grant or deny a patent must be based on the information that was uncovered.

Litigation is entirely different. During litigation, legions of lawyers and engineers are sent out to uncover prior art. In addition to finding prior art, an infringer hires experts to give their opinions about obviousness. If an infringer can’t find prior art that proves obviousness, the reason is clear.

Point 3: IMHO, the biggest problem with patents is the way they’re enforced. Patentees attempt to expand the meaning of claim language in ways that were not contemplated by the inventors and patent examiners. That, and injunctive relief should not be granted to a company that does not have a product, or against a company that cannot change its product to avoid infringement.

Point 4: Don’t shed a tear for Qualcomm. They built and (ruthlessly) maintained an industry on patents. But now, cell phones are functioning more like multimedia computers, so their grip on the industry has been crumbling.

Point 5: Microsoft’s stance against patents is understandable. They don’t need patents. They already have monopoly-like power. Patents give other parties better footing against Microsoft.

Point 6: Patent protection was available even before the Supreme Court’s Chakrabarty decision in 1980, yet the software industry is still thriving. The US has the strongest industry in the world, and there is no sign of it slowing. In the 90’s, the biggest impediment to progress in operating systems and office suites was not patents. It was Microsoft. With their monopoly-like power, they killed off superior products by their competitors. With patents, their competitors might have had a chance.

Posted by Klaatu | June 10, 2007 1:11 AM

And yet, have you once heard the presidential candidates sound off on IP law?

It's interesting what becomes a major facet of our civil discourse and what does not.

Posted by Anthony Damiani | June 10, 2007 1:20 AM

Klaatu | June 10, 2007 1:11 AM is largely correct, but I'll add a few observations.

Regarding his/her Point 2, it is my impression that the Patent Office actually has gotten much better at accessing prior art than they had, say, 20 years ago. Particularly in the "software" area (more on that later). One reason is better access via Internet search engines, and the fact that more companies are patenting inventions embodied in software. In the 1960s and 1970s, inventions embodied in software were not patented, and information regarding them were rarely widely published. Now they are, which is why the Patent Office's access to prior art has increased markedly.

Regarding Point 3, injunctive relief is not required, but it strains credulity to believe that a patent holder who does not have a commercial interest in the invention described in the patent would not be willing to license the patent. Note that I said "commercial interest." It may be that the patent holder holds the patent and may be unwilling to license the patent because it has competing products that would not be covered by the patent.

Regarding Point 5, I have no inside information regarding Microsoft, but it is my impression that they went on a patent acquisition frenzy as a result of two things. One, MS was approached by IBM to license IBM's patents, and had virtually nothing to cross-license (cross-licensing might have reduced the royalties that IBM would have requested). Two, Remember the Stacker case? MS made use of Stacker's patented technology, and had no patents to counter-sue with. MS was found liable of patent infringement, and bought the company at a rather hefty price to avoid having to remove the Stacker technology from its products.

Regarding Point 6, IIRC Chakrabarty was not a software issue, it was, instead, a genetic engineering issue.

Regarding the "more later," two points. One, a patent is a grant, for a limited time, for the inventor (or his company) to induce the disclosure of the details of an invention. A patent is not awarded except in exchange for the disclosure.

Two, there is no such thing as a "software patent." An invention embodied in an encoding on a disk may just as well have been embodied in one or several chips (hardware). The "software patent" idea is a misnomer, pure and simple.

Posted by raj | June 10, 2007 12:05 PM

Chakrabarty was a supreme court decision that simply said everything may be patented. It wasn't until the federal circuit's 1995 decision in Beauregard that the flood gates opened for software patents. Before Beauregard, patent attorneys needed to be a little sneaky about drafting patent claims directed at software (part of a process, or as an algorithm executed by a computer).

In any case, the point made above about the distinction between prosecution and litigation is a critical one. I can certainly attest that while litigating software patents there is no shortage of prior art, which google has made more accessible than ever, for defendants and their experts to opine on. Defendants are most certainly not at a disadvantage in patent litigation -- and now have an even stronger hand in light of recent supreme court and federal circuit decisions.

This is, imho, bad for the industry in that is gives the start-ups less access to capital, and less leverage to get their products into the established distribution channels -- Windows, established web portals, etc -- and makes start-ups less desirable as takeover targets. The hysterics over software patent "abuse" just ensures that the pools of cash the established players are sitting on will not circulate through the industry, and makes VCs more careful about distributing their capital. Bad outcomes both.

Posted by cj | June 10, 2007 1:31 PM

> This is, imho, bad for the industry in
> that is gives the start-ups less access to
> capital, and less leverage to get their products
> into the established distribution channels --
> Windows, established web portals, etc -- and makes
> start-ups less desirable as takeover targets.

Utter baloney. Let's take salesforce.com as an example (note that I have no idea if Salesforce.com files for patents). Is it a good product? You have to judge for yourself, but if you say "yes" I won't disagree. Is it innovative? In fact it is basically a reimplmentation of Act! on a web site. So Act! should file patents and be rewarded for innovation, right? Well, Act! is simply the best of hundreds of similar Apple II, MS-DOS, and (later) Apple Hypercard applications that sprung up about 6 months after the first PC address books were released. Clearly hundreds of people, at least, had the same idea at the same time and also had the resources to get something built and onto the market. Should the first one of those hundreds be allowed to file patents and put the rest out of business? But that is how the entire software industry has always worked, since 1960 at least - observe, share, build your own. Battle it out on quality, improvement, service, and customer value; don't try to shut down your competitor through the courts. Read up on the history of the ARC file vs ZIP. Or GIF vs JPEG.

Ah, but putting it on a _web site_ - that was a new idea! Except that multi-user remote connectivity access to database applications has been available since 1974 (that I know of; perhaps earlier). Did the first person to hang a database at the end of a phone line file a patent on that? If he had, would society be better off? Somehow I think not. Remote-hosted multi-user Act! systems have been available since the dial-up BBS days; should whoever was providing them be allowed to demand tribute from salesforce.com?

One of the greatest advantages of software is that it generally _doesn't require_ huge capital investment to get off the ground: 5 or 10 guys and some space over one of their parents' garages. As far as making start-ups "less desirable takeover targets", how about these startups try /building a profitable business/ the old-fashioned Bill Gates way rather than "flipping"? Since when did "flipping" become the highest economic goal?

In any case, Google acquires entities when they demonstrate that they have a good solid innovative product. I am sure Google doesn't mind if there are patents (they have to fight the 'patent library' game themselves), but that isn't their prime objective.

Cranky

Posted by Cranky Observer | June 10, 2007 6:12 PM

That's not really a correction, Matt.

If you say that the larger, less dynamic parts of the software industry are probably helped by software patents, then, unless you believe its good for an industry to be less dynamic and innovative, it does follow that software patents are bad for the software industry as a whole.

In the same way, GM is not as good for the automobile industry as Toyota is.

Posted by enigma_foundry | June 10, 2007 9:53 PM

Another wonderful use of software patents detailed by eWeek:

}} Security researchers, are you tired of handing your vulnerability discoveries over to your employer, as if that were what you're paid to do? Helping vendors securing their products—for free—so that their users won't be endangered by new vulnerabilities? Showing your hacking prowess off to your friends, groveling for security jobs or selling your raw discoveries to middlemen for a fraction—a pittance—of their real value?
.
Take heart, underappreciated, unremunerated vassals, for a new firm is offering to work with you on a vulnerability patch that they will then patent and go to court to defend. You'll split the profits with the firm, Intellectual Weapons, if they manage to sell the patch to the vendor. The firm may also try to patent any adaptations to an intrusion detection system or any other third-party software aimed at dealing with the vulnerability, so rest assured, there are many parties from which to potentially squeeze payoff.
.
Intellectual Weapons is offering to accept vulnerabilities you've discovered, as long as you haven't told anyone else, haven't discovered the vulnerability through illegal means or have any legal responsibility to tell a vendor about the vulnerability.
.
Also, the vulnerability has to be profitable—the product must be "highly valuable," according to the firm's site, "especially as a percentage of the vendor's revenue." The product can't be up for upcoming phaseout—after all, the system takes, on average, seven years to churn out a new patent. The vendor has to have deep pockets so it can pay damages, and your solution has to be simple enough to be explained to a jury. {{

Posted by Cranky Observer | June 11, 2007 10:02 AM