As members of the open source industry, we know that the industry is fundamentally global. However, being global doesn’t mean being the “same”; in fact, this intellectual understanding was brought home to me very directly when I attended the first Open Source Think Tank in 2008 in Paris. We had a member of the French Parliament discuss their adoption of open source software. The sophistication about the value of open source software and its advantages was considerably higher than in the United States at that time. The market for open source software is different in Europe and companies need to understand those differences to effectively sell into Europe. And Europe has a very different perspective about the future of open source software.
A great way to learn about these differences is by attending the Third Annual European Open Source Think Tank sponsored by Olliance Group and DLA Piper. The European Open Source Think Tank will be held on September 28 & 29 in Paris, France (for more information, see www.thinktankeu.olliancegroup.com). In addition, we are once again proud to partner with the Open World Forum, Europe’s leading conference shaping the future of the digital world, on September 30th and August 1st (www.openworldforum.org). As I noted in my earlier post, I am doing a significant amount of work with open source in the cloud and I am particular eager to attend the panel on open source in the cloud which has a great group of panelists. http://www.openworldforum.org/attend/agenda/open-cloud-conference.
The Open Source Think Tank has become the event where leading global industry experts gather and work collaboratively to chart the future of the commercial open source software industry and shape the evolution of cloud computing. Just as we did in the Napa Open Source Think Tank this spring, we will be moving to an all real-world business case format. We will use selected case studies and focus on the growing commercial maturity and complexity of open source and the evolution of cloud computing and SaaS. Just a reminder: the Think Tank is not a traditional conference, all attendees are expected to contribute and actively participate in the brainstorming and workshop format.
This event consistently sells out so you should register early. If you have not received an invitation, please complete the Request Invitation form at the event website. We are sorry that we may not be able to accept everyone who applies; the event is limited to senior executives who have significant open source experience.
The recent announcement about the “Openstack” product from Rackspace emphasizes the increasing importance of the cloud to the open source industry. http://blogs.the451group.com/opensource/2010/07/19/openstack-from-rackspace-nasa-shows-power-of-open-source-in-clouds/.
For the last year, I have seen an increasing number of open source issues relating to cloud computing (for a general review of cloud contracts, see my presentation as part of a panel for PLI http://www.pli.edu/product/mp3_detail.asp?id=103540). Many of these legal issues are new; it reminds me of the legal situation in 2004 when many fundamental legal issues in open source licenses, such as the scope of the GPL and the remedies for breach of an open source license, were unsettled. The use of open source software in the cloud raises new and difficult legal issues such as when is use in the cloud a “distribution” (which triggers obligations to make source code available in the GPL family of licenses) and what is an “aggregate” in the cloud. We had a very lively discussion about these issues at the Stanford Law School E-Commerce conference (most of these questions deal with the GPL family of licenses so the use of the Apache license by OpenStack project will avoid some of these questions).
The obligation to provide source code under GPLv2 is based on the “distribution” of object code of the GPL Licensed Code. The term distribution does not have a simple or obvious meaning; “distribution is defined in Section 106(3) of the Copyright Act of 1976 (“1976 Act”) as the right “to distribute copies or phonorecords of the copyrighted work to the public by sale or other transfer of ownership, or by rental, lease, or lending.” The term “public” is not separately defined in the 1976 Act but is defined by reference to “publication”. Publication is defined as follows: “Publication is the distribution of copies or phonorecords of a work to the public by sale or other transfer of ownership, or by rental, lease, or lending. The offering to distribute copies or phonorecords to a group of persons for purposes of further distribution, public performance, or public display, constitutes publication. A public performance or display of a work does not of itself constitute publication.” Publication is a complicated concept which derives from the Copyright Act of 1909 (“1909 Act”). Under the 1909 Act, publication without copyright notice had very significant legal consequences: it could result in the loss of all rights under the 1909 Act, placing the work in the public domain.
Ironically, cloud computing requires us to reach back to the “limited publication” doctrine from the 1909 Act to find legal guidance for how to interpret the scope of the distribution right. The doctrine of limited publication was created by the courts under the 1909 Act in order to avoid the forfeiture of rights under the 1909 Act through inadvertent transfer of a copy of the work without the required copyright notice. White v. Kimmel is the seminal case for the doctrine and defines a “limited distribution” as a distribution that “communicates the contents of a [manuscript] to a definitely selected group and for a limited purpose without the right of diffusion, distribution or sale.” White v. Kimmel, 193 F.2d 744, 746 (9th Cir. 1952). The doctrine of limited publication continues to be viable under the 1976 Act, but it has become much less important because the use of copyright notice is now optional and the failure to include it no longer puts the work into the public domain. The House Report of the 1976 Act defines distribution to the public as “it refers to persons under no explicit or implicit restriction with respect to disclosure of its contents.” Cloud computing may bring new life to the “limited publication” doctrine and open source lawyers may be reading a lot of 1909 Act cases (get ready, there are more than 50 of them that are relevant and thanks to our summer associate Eduardo Blanco from Northwestern University School of Law that helped me sort through them).
I have been reading the debate about “Open Core” which was stimulated by Jorg Janke post about Compiere. http://www.compieresource.com/2010/06/compiere-open-source-failed.html. The open source community owes Jorge Janke a huge debt of thanks for his frank discussion of what happened at Compiere. People are rarely eager to share the details of their failures. I think that the most important lesson from his posting is the critical importance of management that understands its market. The venture capital industry learned this lesson long ago. When making a decision about an investment, venture capitalists focus on the management team; they understand that the technology is important, but great execution with mediocre technology will win over poor execution with great technology. This rule has been established by decades (and thousands) of investments. Open source companies pose particular challenges for management because of the critical role of communities and their expectations in the success of the company. These issues are very different from traditional software companies. Compiere is a very telling example of the nature of those challenges.
This post has launched a discussion of the “open core” business model and whether it is true “open source” http://www.computerworlduk.com/community/blogs/index.cfm?entryid=3047&blogid=41. I have great respect for Simon Phipps and his contributions to the open source community, but I strongly disagree with his statements. I am very concerned that if he is successful, end users will have fewer software programs under open source licenses. This result arises because of the law of unintended consequences: the successful demonization of the open core model will result in fewer venture capital investments in companies using open source licenses.
In the interest of transparency, I work with over twenty open source companies, most of who were funded by venture capitalists and the vast majority of which use the “open core” model. These companies have provided significant value to end users through the software licensed under open source licenses. Simon states: “But to use the package effectively in production, a business probably won’t find the functions of the core package sufficient, even in the (usual) case of the core package being highly capable.” This statement is simply incorrect. I have sat through many Board meetings and, in fact, the conversion rate from “open source” to “commercial” licenses is generally less than 10% for these companies. Thus, more than nine out of ten end users find the functionality of the open source version satisfactory.
Simon says that open core does not provide software freedom for “end users”. Yet, nothing prevents the end users of the open source version to modify it and distribute it or otherwise exercise the rights under the license. In fact, Compiere demonstrates the fallacy of this position because it created two different forks. Simon complains about the lack of access to the “commercial extensions” of open core programs. However, as Marten Mickos notes, the effect on the end user of the employment of the Apache license is the same as the open core model: commercial extensions are not made available to the community. http://webmink.com/2010/06/24/links-for-2010-06-24/#comment-870. I agree with Matt Aslett that the open core model does not violate the Open Source Definition, either literally or in spirit. http://blogs.the451group.com/opensource/2010/07/02/open-core-is-not-a-crime/. (please note that this position is a personal one and does not reflect the view of the OSI which has not yet taken a position on this issue). Simon appears to be suggesting that only a “copyleft” approach in which all of the software must be available under an open source license to meet the Open Source Definition, which is simply incorrect (the Open Source Definition was a reaction to the limitations imposed by the copyleft approach).
I agree with Matt at one level that ultimately this debate will be decided by the market (i.e. end users). However, I don’t agree that it is futile. Most venture capitalists will not invest in companies that do not use the open core model, so if the open source community leaders are successful in demonizing the open core model, they will decrease the willingness of venture capitalists to invest in open source companies (just a reminder, that a recent book, Mastering the VC Game, recently noted that venture capitalists typically look at 300 companies for each company in which they invest). Although not all open source projects need venture capital support, venture capitalists have been a significant source of support for open source projects (as well as new software made available under open source licenses) and end users have been the beneficiaries of their investment. If the open core model is no longer considered open source, the biggest losers will be the end users; they will lose the opportunity to benefit from that investment and that is certainly not consistent with the goals of open source