To date, between 50 ad 60 licenses have been issued within the open source community. And concern is growing over the legal problems that could pop up as interactions between the various licenses become more frequent. In this exclusive interview, SearchEnterpriseLinux.com site expert Bruce Perens gives his take on licenses in the open source community, and responds to comments made by the Free Software Foundation's general counsel Eben...
Moglen at the LinuxWorld Summit in New York last week.
Is there a belief today that there are far too many open source licenses?
Bruce Perens: Yes. In general a company can do all it wants with open source while using only three licenses: BSD, LGPL, and GPL. Each of those licenses fulfills a different business purpose on a range between no restrictions at all (BSD) and mandatory sharing (GPL).
How does having too many licenses hurt businesses trying to adopt or develop open source technology?
Perens: The combinatorial problem: understanding all of the possible combinations of 60 licenses and their consequences would be very difficult. Open source is supposed to promote sharing of code. Having lots of incompatible licenses deters sharing of code.
At the LinuxWorld Summitt, Free Software Foundation general counsel Eben Moglen preached an end to the commercialization of software, and predicted that through organizations like SourceForge, the future of software would be a free one. What's your take on that?
Perens: The Free Software Foundation would like to see the end of proprietary software, but a lot of us who are identified as open source promoters would be content to see free and proprietary software able to compete fairly on a level playing field. Currently, the field is tilted against us through legislation that only takes the proprietary software paradigm into consideration, and by the anti-competitive efforts of Microsoft and its ilk - for example, the use of proprietary file formats and intercommunication protocols to lock out competition. Proprietary software companies should differentiate themselves on features rather than use enforced incompatibility to protect their markets.
There are a few products that work best as proprietary software. TurboTax is one: it's not written for love, it has very deep information on tax law that must also be very timely because tax law changes from year to year, it has a high liability load. But most software is fine as open source. There's no need for anyone to buy a proprietary office suite or web browser or operating system.
How are companies going to make money?
Perens: I have a 24-page paper on this at Perens.com called The Emerging Economic Paradigm of Open Source. A short explanation might be that for most companies, software is a business enabler rather than the product. Consider Amazon, for example. They sell books, but their business is entirely enabled by software and would not work without software. For Amazon, software is a cost-center and books are the profit center. Amazon has "recommendation" software that differentiates their business from other companies such as Barnes & Noble. It recommends other books based on the book that a customer shows interest in. Amazon must keep that recommendation software proprietary for as long as it continues to differentiate their business. But Amazon uses Apache and Linux, which do not differentiate their business.
Open source is a way that companies like Amazon can distribute the cost and risk of software development for non-differentiating, cost-center software. By sharing work on Apache and Linux with other companies, Amazon can only save money and increase their profits. So, the economic lesson is that open source can help companies distribute the cost and risk of the development of non-differentiating, cost-center software, and can help them move more of their software budget into the development of business differentiators. By doing so, they make more money.
With Moglen's examples, is there a chance of another SCO-type battle?
Given what has happened to SCO, I think businesses would be well advised not to start a fight with us. But software patents can potentially be worse than SCO, because their legal case would be more solid. We have done a lot of work to prepare for such attacks and continue to work on that.