You are currently browsing the archives for May, 2007
During his OSBC keynote, Eben Moglen said that “Science must be free. Knowledge must be free. Let freedom ring.” This parallels the argument he’s made in the past that patent software is akin to patenting math.
This is compelling rhetoric, and demonstrates Moglen’s abilities as a lawyer. But rhetoric is a double-edged sword, in that it can be put in the service of logic and truth — JFK or Martin Luther King come to mind — or in opposition to logic and truth. I believe Moglen’s rhetoric falls in the latter category.
There are two main flaws in Moglen’s arguments. First, he confuses knowledge with invention. My undergrad degree is in Physics, so I know first-hand the huge difference between Physics and Engineering. Our purposes were different, our training was different, our culture was different than the Engineering students down the hill at UCLA. The goal of the physicist is to further knowledge without any thought to its immediate usefulness. It’s rare that any new discovery in Physics has any commercial application within the twenty-year period of patent protection. Engineering’s goal, on the other hand, is to make things that solve an immediate problem, often a commercial one. They couldn’t be more different.
The same difference is found between theoretical and applied mathematics. When Calculus was invented by Isaac Newton, it had no commercial application. On the other hand, figuring out how to perform real-time Fourier transforms to compute 3D distance readings from sound echoes hardly advanced the state of mathematical knowledge, but it did make sonar possible.
Invention, and knowledge, two entirely different things that Moglen lumps together. Knowledge is free, always has been, and that didn’t change with software patents. Invention, a commercial activity, has historically been protected for a short period of time. I don’t know if Moglen deliberately confuses these two categories, or just doesn’t understand the difference, but either way, it’s a shame.
As a result of his confusion of knowledge and invention, Moglen implies that true invention must be instantiated in a physical invention, since any innovation instantiated digitally, according to Moglen, is not invention but knowledge. He is selling software short. Is invention any less deserving of protection because it results in software rather than a physical contraption? Software engineers should take offense at that implication. They are no less inventors, or innovators, than someone building a new hybrid engine or a nano drug delivery device. His artificial conflation of invention and knowledge leads to a necessary short-changing of digital invention.
What’s a shame is that Moglen doesn’t need to use these rhetorical excesses to make his case. His view is, at bottom, an ethical argument based on â€œthe way things ought to beâ€ rather than a utilitarian argument. But patent law isn’t meant to instantiate any moral construct; it is solely a pragmatic balancing act between the need for society to incentivize invention while allowing invention to be utilized broadly by society. Property fundamentalists on both the right and left argue for or against patents from moral principles, but the law does not.
Ultimately, the question regarding software patents is a utilitarian one: do they benefit the common good? This is where software patents fail miserably. The big software vendors, my employer included, file for software patents in a bizarre process of patent proliferation to make sure that their competitors infringe on as many patents as they infringe of theirs. Everyone is infringing someone else’s patents. Patents impose a financial burden on innovation rather than a financial incentive to innovate. They don’t work.
This is why software patents are a bad idea: in a strictly utilitarian analysis, they are not only unnecessary, but counter-productive.
It’s a shame that Moglen, Stallman and the FSF argue against software patents on moral grounds using inapt analogies like “knowledge = invention” and â€œsoftware = mathâ€ when the utilitarian arguments are much more compelling and easily embraced. But the FSF is an ideological movement, and not a practical one. As a result, they hurt their cause as much as help it.
I spent the last couple days at the Open Source Business Conference, along with Gary Ardito. Unlike Gary and Matt Asay, I am not much of a live blogger, so here’s a recap of some of my thoughts a day after the fact.
I was struck by the stark contrast between the free software and the commercial open source crowds. While many of the keynotes advocated a free software ideology (and I don’t mean that word in a negative sense), the enterprise IT, open source start-up and VC speakers were very pragmatic. While Eben Moglen argued for the virtue of the GPL, SugarCRM was unapologetic about their use of their SugarCRM Public License. While Linux and Gnu have been developed by broad communities, the commercial open source companies have very little code developed by anyone not on their payroll. For them, open source isn’t a movement, or an ideology, or a moral and ethical imperative — it’s a strategy to build successful software businesses.
And it works.
Larry Augustin shared some software company P&L numbers from the golden years of the 1990s, the not-so-golden years of the late 90s to now, and the hopefully-new-golden-years of commercial open source. During the late 90s, sales and marketing expenses for software companies shot up, coming at the expense of R&D. The market tightened, and buyers could make vendors go through lengthy and expensive sales processes with RFIs, RFPs, POCs etc. The cost of generating leads by buying lists and cold calling became too expensive. The push model no longer was financially viable.
The advantage of open source isn’t so much in the development end, although there’s some benefit there. According to Augustin, the benefit is in moving to a pull sales and marketing model. Vendors don’t have to explain their software to prospects, because they already have it installed, are already playing with it. Trade shows have been replaced by Google, with a much lower cost way to educate the market and generate leads. The software financial model has returned to balance.
Not that there’s no advantage on the development side, but it’s not necessily in developing the core product. Community involvement helps with enhancements around the product, such as localizations, interfaces, integration etc. Danese Cooper of Intel, and formerly of Sun, talked about community involvement with OpenOffice. Microsoft has localized MS Office to 13 languages, which isn’t enough to cover the world. When OpenOffice was open-sourced, Sun saw this localization gap. They made it easy for contributors to localize OO, and the community got excited. For the first time they could have an office suite in Estonian, Malay or Swahili. And now they do.
So in spite of the moral and ethical dimensions to open source, for the majority of the OSBC attendees it’s all business. And that’s a good thing.
There were other highlights, such as a session on the MS-Novell agreement, but I’ll save those for other posts. It was two very well-spent days, so thanks to Matt and all the other organizers for a great conference.
Our new best friend Microsoft has raised the specter of their mythical 235 open-source-infringed patents, but of course they won’t divulge which patents these are. It kind of reminds me of the scene from the movie Field of Dreams where Kevin Costner is trying to get James Earl Jones to go to a baseball game with him:
I didn’t want to do it this way.
What the hell is that?
It’s a gun, what do you think it is?
It’s your finger.
No, it’s a gun.
Let me see it.
I’m not going to show you my gun.
Because, of course, there was no gun.
To be fair, the Microsoft situation is a little bit different. There certainly could be a gun with 235 bullets. The reason Microsoft won’t show the gun is that they have an economic disincentive to do so. If the open source community knows which patents are infringed, they can either challenge them or code around them, effectively taking away Microsoft’s weapon, and therefore MS’s ability to extort licensing payments from open source users.
What isn’t known is how many patents held by the Open Invention Network and its members Microsoft is infringing. This is one situation where the open source model is at a disadvantage: Microsoft can easily throw hundreds of lawyers at figuring out which patents open source may have infringed, while the open source community is not organized to easily do so for OIN patents that Microsoft infringes. What’s worse, open source means that Microsoft can pore through potentially offending code, while we can’t go through Microsoft’s code to document their infringements.
I have not been a fan of the Free Software Foundation, but here’s an area in which they could truly contribute, rather than writing yet another draft of GPLv3: set up an open commons project to identify patents that Microsoft may be infringing. This analysis would have to be done based on MS software’s behavior, since the code can’t be examined, but at least we would have a credible counter-claim to force Microsoft to tone down their extortion racket a bit.
And no, I don’t see this as contradictory to my generally favorable view of the Novell-Microsoft agreement. Novell has denied any implication that we acknowledge any infringement of Microsoft patents, and have been donating our patents to the OIN for the benefit of open source. This would be a complementary step to continue progress towards the ubiquity of open source and its inter-operability with Windows.
Many years ago, I read a book about an alternate universe where the laws of entropy are reversed. Instead of tools and artifacts deteriorating over time with use, they would improve with use. A hammer would become a better hammer the more it’s used. On the other hand, a hammer’s usefulness as a hammer would deteriorate if it was not used.
In this universe, if I wanted to make an ax, I’d get an ax-like rock, tie it to a stick, and start chopping wood. At first, my proto-ax wouldn’t be very effective, but with use it would gradually transform, getting sharper and sharper. The handle would get stronger, it’s attachment to the blade firmer, the weight distribution better balanced, all through â€œpracticingâ€ the ax. With enough use, I would have an extremely well-made ax, all through the backwards flow of entropy. But then if I were to leave my ax in the shed for a year and return, I would once again find a stick tied to a rock.
The novel explored the differences in behavior this reverse flow of entropy would cause. In our universe, lending our neighbor our lawn-mower is to do them a favor. In this backwards universe, if we want to do a neighbor a favor, we would borrow their lawn-mower, since we would be improving its quality by using it. If we had a table that we weren’t using, we wouldn’t store it away for future use but would instead find someone to use the table, since that would be the only way to maintain its utility as a table.
While I can’t remember the title or author of the book, I vividly recall its premise because of its complete reversal of the economics we take for granted (if anyone knows of this book, please leave a comment!). But of course this is fantasy, with no correlation to reality. Right?
Fast forward to the present. Last night I was reading The Success of Open Source by Steven Weber. A physical good is a â€œrivalâ€ good because if I use it, you can’t. Digital goods are â€œnon-rivalâ€, since my use of a music or video file in no way impairs your ability to use the same good as well, since an infinite number of copies can be made at zero cost. But Weber defines a new term, â€œanti-rivalâ€ goods (page 154), which increase in value the more they are used. Open source software are anti-rival goods, since not only does my use of a piece of software not prevent you from using it as well, but my use of the software increases the value you receive from your use of the same software. Partly this is due to the well-known network effect, where the more people using a network, the greater the utility of the network to each participant since they are able to communicate with more people. But it’s more than that. The more users that are using a piece of open source software, the more users that will identify and report bugs, and the more users that will contribute bug fixes, enhancements and extensions. In effect, the software improves with use. Anti-rival goods reverse the flow of entropy.
In the open source community, we help our friends by using their software, and distributing it to others. If we develop a piece of code that we don’t want to maintain or extend, we are incentivized to give it away so that others will, thereby preventing its gradual decay. It is eerily like that fantasy book I read in my younger days. Our intuition developed through using, buying and selling physical goods leads us entirely astray in the open source economy where giving is receiving and receiving giving. We don’t want to acquire things that have been â€œgentlyâ€ used, we want things that have been widely, harshly and unremittingly used. Things don’t wear out, they wear in, or up, as they improve with use. We don’t put things away to keep them new, we distribute them as widely as possible to keep them new.
The unrealistic, ridiculous fantasy universe I read about so many years ago and dismissed has become reality in the world of open source.
Update: Lance Gifford, with whom I share 50% of my genome, and who lent me the book to start with, solved the mystery. The Practice Effect by David Brin.