I’m at the Business Success Using Open Source Software panel at the ABA Annual Meeting in San Francisco. The participants are:
- Katherine Spelman (moderator, Cobalt LLP)
- Pam Samuelson (Boalt Hall)
- David Kappos (IBM)
- Karen Copenhaver (Choate Hall & Stewart; fmr. GC, Black Duck Software)
- David Marr (Sun)
As with all of my liveblogging, these are my impressions of the proceedings and are not for attribution to the participants, and certainly not to me.
OSS has come a long way in the 20 years it’s been on the scene. It started as a set of programming tools put out by the FSF; we lawyers eventually came to understand and like the GPL, turning copyright law into a tool of inclusion rather than exclusion.
OSS has now grown in an amazing way; there are more than 150,000 OSS projects, dozens of license agreements, Wikipedia, world-class programming tools, Firefox, Linux, and a myriad of business models that make OSS as a development paradigm not only profitable but preferred.
We can’t find a software start-up in the Valley that isn’t using OSS. If you don’t think your clients are using OSS, you have a surprise coming.
And now governments are permitting and even mandating the presence of OSS in their bidding processes.
This brings us to some questions: How does business get profitably conducted around free software? What are the roots of OSS? Can it apply to other fields? Why haven’t other fields adopted open development models? How does the OSS model that works so well for copyrights work for patents? What are the limits of the OSS model, and what’s coming in the future?
A number of business models have emerged around OSS:
- Support and service. This is the Red Hat model. Software is distributed under an OSS license, with services offered for a fee.
- Building block model. This is the TiVo model. Companies are using OSS as a component of their business — embedded Linux, etc.
- Add-on solutions model. This is the IBM model. Using OSS (typically Linux) and offering solutions that build around that and integrate it into larger situations. This is an extremely successful business and scales well.
- Proprietary extension model. This involves developing proprietary software that links to OSS in conformance with the GPL, then donating that proprietary software to the community as its competitive advantage fades. IBm does this with WebSphere.
Most of those business models leverage shared infrastructure. This is all about avoiding spending money on unproductive effort and spending it close to customers’ needs.
Can this work outside of software? Yes. IBM instituted an organization called power.org, where we put a giant amount of architecture information about the power architecture (which is a microprocessor architecture) into the public domain. It’s working: there’s a vibrant community.
What about pharma? An organization called CAMBIA in Australia licenses bio innovations under open principles.
And it works for information, too: Wikipedia and, indeed, the GPL3 process itself. Open development will work wherever there is commonality of problems, opportunity for profit or reputational gain from contributions, and call for open standards for interoperability.
It can work for patents as well as copyrights. IBM donated 500 software patents for free use in OSS. We wanted to make patents less threateing to OSS — and even a tool to promote OSS. Many other companies, including OBM, have announced more pledges since then. Then there was the formation of OIN. Google just joined as a licensee. This is rationalizing the discoursein the OSS/patent area.
Also, Peer-to-Patent is bringing the two areas together. It’s allowing experts from all over the world to submit and comment on prior art. It’s apparent that this is going to be the wave of the future — “the USPTO meets the 21st century.”
So what about efforts to capture the millions of lines of OSS as prior art? We’ve been working on that — there’s lots still to do, but the idea is to categorize the code by patent classes so that they can easily be used by patent authorities.
OSS is on its way to turning patents to a positive force for OSS.
OSS is going to continue to flourish and grow in areas other than software. We had a CodeJam to help solve major world problems, and taking the IP stress out of the system allows people to collaborate better.
OSS is here to stay; it will grow; it will spread to other problems. It is showing that, like copyrights, patents can not only coexist with OSS but fuel its growth. The OSS model with thrive alongside proprietary development, and will create plenty of legal issues along the way.
I’ll be talking about OSS and competition. I got interested in this when I taught a class with Mitch Kapor. I’ve also written a paper on it.
- Proprietary software and OSS are in a deadly war, and one will strangle the other. (This is hyperbole. Eben Moglen writes about “triumph” over proprietary software; Craig Mundie writes similarly harshly about OSS. Pay no attention.)
- FOSS is hostile to IP. (This is also false. Some people who like OSS are also hostile to IP, but nothing inherent in OSS is outside the standard IP narrative.) The specifics of the license don’t matter that much to the developers. It’s a social constitution, a social norm. Not only will you get sued if you violate the license, but you’ll be shunned — and that’s sometimes worse.
I agree with David that FOSS and proprietary software can coexist in the same products, and will increasingly do so. Also, there are some sectors where one or the other will just prevail: avionics software will probably always be mostly proprietary.
But some places, there is head-to-head competition. IE v. Firefox. Linux v. various server OSes. You get more innovation in such situations, and that’s great. Firefox is driving innovation in IE, and vice versa. MySQL has put price pressure on Oracle. And so on.
OSS has significant development advantages: spreading out common work, a supportive community, and it can be (though it is not necessarily) cheaper. It’s certainly easier to modify internally, by its nature.
There are some risks. One is, “who do I call if it breaks?” For this reason, support services can be sold. Forking can lead to incompatability. Patent risks are still real. Linux may now be in a really good place, but not all projects have companies throwing patent licenses at them. And it’s not easy to transition a proprietary product into a successful OSS project. (Netscape had to be rebuilt substantially to make it into an OSS product.)
Overall, OSS can be less risky: less investment than proprietary development, and somewhat less chance for financial reward.
I’ll be talking about the GPL. The area that can be most difficult for us as lawyers — copyleft. Ask them:
- What GPL code is being used?
- What form are the bits in — binary or source?
- What has been done with the bits?
Then you get three hard questions:
- When is the source obligation triggered?
- What must we release?
- How must we release it?
[He has very good slides that explain the answers to these questions, including an amazing slide with a spectrum of situations drawing the line when copyleft attaches. I'll try to get a copy and post them.]
Design that Matters is a company that started a number of years ago. You could graduate from an engineering school without having anything you designed get manufactured. These kids got this idea that people should put their designs up against manufactured products — and that they could put together a portfolio of products that NGOs around the world needed and use them as student design projects. Then the students could put their designs back into the portfolio, if they wished.
The first product was an IV clamp. (They were quite non-intuitive.) The second was a projector for schools. (Turns out school is at night, and there’s no light for books.) The third is a baby incubator. (Existing incubators required consistent electricity and frequent changes to the filters.)
Building a world where it’s easy (but not mandatory) to share — with an open source design model — was the goal. It’s a wonderful story. We’re five years into it, and we’ve never had an IP issue. The sharing rate is very high. The students realize the joy of being permitted to share.
When I think about OSS, I think about making things possible, like Wikipedia, that could not have been imagined before.
The idea I wanted you to get was that world-leading companies are embracing this model as a way to find efficiency — not just as a legal curiosity or a project for hairy geeks. None of us is alone, and OSS is a very sophisticated acknowledgement of our interdependence.
It has restored competition to the software industry, and the compeititon is perfect, due to the ability to fork any project. There’s more and more infrastructure that none of us wants to build alone. It is providing, and is going to provide, a platform for innovation.
[Then some very interesting questions and answers, which went too fast for me to write down.]