When I met Bob Sutor at a meeting in Brussels in 2005, I told him that I had a hard time understanding the meaning of open source. Open source licenses come with various restrictions that don't fit well with commercial interests. His answer was "It depends." Wow, it depends. How can it fit?
The title of my presentation at the Open Source Business Conference last week (SAP: Friend Or Foe Of Open Source?) refers to Glyn Moody's blog post in the Linux Journal last year, which is just an indication that either SAP actually still has a hard time with open source (like me back in 2005) or our approach is simply not well-known.
Well, that was actually the reason why SAP sponsored the conference and why I gave an overview and reasoning about our engagement with open source. I believe that SAP actually has a sophisticated, value-driven approach towards open source. Also, we are on our way to significantly increase our engagement: we will use more open source in our products and we will contribute more to open source projects. The reasoning is outlined in the presentation: increased development productivity, continued innovation, protection of investments and reuse of skill sets.
When in the past we asked ourselves "why use open source?" we will from now on ask "why not use open source?" (provided that there is a mature and well-adopted open source technology). Pretty significant change compared to how we managed open source in the past (actually as a risk).
As for the future of open source in general, I believe that there are two trends that will have at least some impact. Firstly, more and more software will be made available as a service. This questions the meaning of the source code (and its availability) in such offerings since one would rather create add-ons to the SaaS offering and would not modify the core SaaS modules. Secondly, more design decisions will happen on a higher abstraction layer of the software stack. Model-driven development turns to become a trend where collaboration happens on the design of business processes and not on the design of the source code since the source code will often be generated out of the process model. These two trends don't mean that open source will become irrelevant, but only that more (open) innovation will happen in other areas (add-ons instead of core platform modules) and that the object of innovation more often will not be the source code but a process model, a service definition, an implementation best practice, etc. The principles we know from the open source community today will need to be applied to these new contexts.
