Saturday, March 29, 2008

Open Source Overview from OSBC

The slides from the presentation by the NorthBridge Capital Partners from OSBC has some interesting information. The most interesting finding is that the subscription business model is likely to become much more important in the future. This finding certainly reflects my experience. Most of my open source clients are adopting a subscription model because of its attractiveness to potential customers. They also stated that SAAS (among cloud computing, software appliances and virtual infrastructure) is likely to have the greatest effect on software delivery and business models. If SAAS continues to be important it may drive the adoption of the Affero GPL (with its focus on "network use") instead of the GPLv3 (just a reminder that the Affero GPL is simply the GPLv3 with a network use provision added).

The survey confirms the Gartner report that open source will continue to expand rapidly. According to the survey, more than half (55 percent) of the respondants believe that open source software will increase to 25% to 50% percent of software used as compared with proprietary software. The respondants believed that over the next five years open source software would have the greatest effect in the Web publishing and content management system (CMS) markets, but the least effect in security tools.

Saturday, March 22, 2008

Open Source Growing at an Exponential Rate

Although many of those of us in the open source community believe that open source is growing rapidly, a new study validates these conclusions: the study released March 14 by SAP Research concludes that open source software is growing at an exponential rate in "additions to open source projects, the total project size (measured in source lines of code), the number of new open source projects, and the total number of open source projects". They also noted that the total amount of source code and the total number of projects double about every 14 months. The report is based on information from the service.

The consequences of such growth are dramatic. As we noted at the Open Source Think Tank, open source software is becoming standard and the comparison for users is now between functionality and price rather than "closed" and "open" source software. In fact, this report suggests that "open source" may cease to be a different type of software, it will simply become another way (and probably the dominant way) of developing and distributing software.

Wednesday, March 19, 2008

What Questions Should the Open Source Community Ask Brad Smith, General Counsel, of Microsoft

Stephen Walli brought to my attention that the open source community will have the chance to ask questions of Brad Smith, the General Counsel of Microsoft, who will be giving the keynote presentation at OSBC next week. Stephen has some interesting thoughts about how the open source community should take advantage of this opportunity.

He also mentioned that a website has been set up to suggest questions to the panel. I encourage you to participate!

Advice for Startups: Working with Large Corporations as Partners and Investors

I recently participated in a webinar with Bob Ackerman of Allegis Capital, Mark Klopp of BellMason Group (formerly head of Eastman Chemical's venture capital program) and Claudia Fan Munce of IBM (Managing Director of their venture program). I was included because of my experience in working with startups as well as corporate investors (I run DLA Piper's Corporate Venture practice). The webinar addressed the issues of dealing with corporations as partners and as investors.

Large corporations are the major customers and the most common exit strategy for venture-backed startups. The webinar was a very valuable summary of how to deal with corporations as partners and potential investors. For example, Bob Ackerman described how IronPort was able to obtain a dominant position in the French market by partnering with Societe Generale.

As I noted in the discussion, startups need to be careful about being too reliant on large corporations for critical functions: large corporations may decide that a market is no longer attractive and leave them and the startup may be left in a very difficult position. For example, we worked with one startup who was making lead/acid batteries and relied upon a corporate partner to manufacture the batteries. Lead/acid batteries are difficult to manufacture and require careful attention to environmental issues. They are not popular neighbors, so the corporate partner was very important to the startup. Working with our client, we negotiated a six month notice period prior to termination of the manufacturing agreement. About a year later, the corporate partner unexpectedly declared that they were leaving the manufacturing business and told the startup that they were shutting down the factory in thirty days. We reminded them of the six month notice period and were able to find an alternative manufacturer.

Startups also need to understand the different expectations of large corporations: large corporations value predictability and planning. Startups, on the other hand, change their products and strategies. This mismatch of business cultures and expectations needs to be recognized and managed.

And corporate investors are very different from traditional venture capital investors. Corporate investors are frequently very concerned about the potential sale of their portfolio companies to "competitors", a term which can vary over time. Startups and traditional venture capital investors are focused on economic returns. Consequently, startups need to be careful that their corporate investors do not control the potential sale of the startup such as by owning a majority of a series of preferred stock which has a protective provision providing approval rights over the sale of the company.

If you are interested in learning more you can get more information on obtaining a CD of the webinar at:

Thursday, March 13, 2008

Thinking Legal for Open Source Success: Trademarks & Licenses

I have been reading the dueling posts between Matthew Aslett, Tim Bowden and Ian Skerrett about the relative value of licenses and trademarks to the success of an open source company. are particularly interesting to me because I have been putting together my material for my presentation at OSBC: Implementing Your Open Source Business Strategy through Your Legal Strategy.

I think that any such discussion needs to first clarify the business model of the company: a dual licensing model has very different legal strategy from a services based model. You also need to consider the context of the business: the nature of the software (applications v. infrastructure), competitors, ownership of intellectual property rights and channels. For example, Tim discusses the difficulty of Linux distributors who are competing with the same code base. Yet Linux companies have a very significant problem: they do not control the intellectual property rights in the code base. These rights are distributed among the thousands of contributors. This situation is quite different from a company such as MySQL, Zimbra or SugarCRM who own the intellectual property rights in their code base. MySQL, Zimbra and SugarCRM can offer their products under a commercial license as well as the open source license (the "dual license" model) and even change the licenses (SugarCRM shifted from the Sugar Public License to GPLv3). Consequently, some of the discussions are comparing apples to oranges because new companies have much more flexibility in developing a legal strategy than companies working with an existing project.

If you are starting a company and will control the intellectual property rights in the code base, you will be able to use licenses, trade secrets, trademarks and patents as part of your legal strategy. However, if you are using an existing project licensed under the GPL, you will be limited in your license choice and you will not be able to use trade secrets and your patents will be of limited value because they may be licensed under GPLv2 (they are licensed under GPLv3 if you have modified the code).

License choice can be very important: I was speaking with a senior executive of JBoss and he was convinced that the choice of LGPL significantly limited their success because it is such a difficult license to understand (I personally view the LGPL as harder to interpret than the GPL). However, I don't think that the selection of GPL is a silver bullet for a company. Zimbra was using the Zimbra Public License, a version of the Mozilla Public License. Sleepycat was using its own special license (a form of BSD).

If your business model requires that your software integrate with third party software, the GPL is a poor choice. Many third party software vendors will be reluctant to integrate their software with the code if they need to license their software under the GPL. In that case, BSD or the MPL (a weak copyleft license) will be a better choice.

However, trademarks are very important in the open source software industry. In fact, they may be the major asset of a company that is using a common code base such as Linux. Trademarks take on a special importance because open source software is frequently distributed outside of traditional channels.

Unfortunately, the question of the importance of license vs. trademark cannot be answered in general. Like many questions in business (and the law) context is critical. If you are interested in more discussion about this question, join me at OSBC!

Affero General Public License Approved by OSI

The Affero General Public License ("AGPL") has been approved by the OSI. The AGPL is an important option for companies whose software can be used to provide a service. It includes a "network use" provision which requires that the source code be made available to "users" of the service. For a more information about the AGPL, you can read my earlier post.

Thanks to Funambol for submitting the license to the OSI and following up. OSI approval is important for adoption in many large organizations which use OSI approval as a basis to approve the use of software distributed under such a license. I am particularly pleased that OSI adopted AGPL because Fabrizio Capobianco at Funambol and I have a bet about whether GPLv3 or AGPL will be the dominant license in five years. He can no longer complain that it is not a level playing field! So consider AGPL, but remember I have money riding on your adoption of GPLv3.