@Martin_Guenther, I appreciate your concern and understand your perspective. I’d like to add some context to the discussion.
Open Robotics would not exist today without the support of the US Department of Defense (DOD). Back at Willow Garage around 2010-2011 we often talked about creating a separate organization to specifically focus on open source development of ROS & Gazebo. What was missing was a way to fund such an organization. A contract from DARPA to develop simulation software for the forthcoming DARPA Robotics Challenge (DRC) allowed us to leave Willow and start OSRF in 2012. We were helped by a generous donation from Willow to cover our initial operating expenses, but it was the DRC contract that made it possible to build a team and create a sustainable organization around it. The DRC program alone supported ~10 people working on Gazebo and related projects for more than 3 years.
When Willow was winding down in early 2013, we were able to hire the core ROS team and thereby take responsibility for ROS itself only because of a grant that we had through the National Robotics Initiative. The grant was 1/3 funded by the US Army, with the rest funded equally by NASA and the National Science Foundation (NSF). Taken together that funding supported ~4 people for 3 years to maintain ROS and to start work on ROS 2.
It’s impossible to know for sure, but it’s reasonable to wonder whether ROS and/or Gazebo would have become orphaned projects had we not received that DOD funding at that time. Perhaps the community or another organization would have picked them up. But in my opinion, probably not.
Since that time we’ve worked (and continue to work) on a number of DOD-funded programs from various agencies. We’ve never tried to hide that support; you can find 4 DOD agency logos on our sponsors page. We’ve also never tried to hide the work that we do for those agencies. Quite the opposite: in all these programs the code and documentation that we develop is 100% open source. To my mind, that’s the best possible use of public money that’s paying for R&D: build technology that can be used by everyone, everywhere.
In addition, the fact that our government funding, DOD and otherwise, is all channeled into open source products has an important consequence: we cannot work on anything that is classified or subject to export control. As you might imagine, this constraint rules out pretty much anything related to weapons or other offensive technology. We go further, judging each potential project based on the apparent intent of the people running the program and the likely application of the technology. Over the years we have declined many funding opportunities because we found them to disagree with our personal and organizational principles.
Of course it would be naive in the extreme to imagine that technology that we develop and release won’t ever be used in a manner that we disagree with. I’m quite sure that somewhere out there are weaponized robots running ROS and being simulated with Gazebo. We can’t stop that from happening and still make our software open source; clause 6 of the Open Source Definition says, “The license must not restrict anyone from making use of the program in a specific field of endeavor.” But we can and do choose not to build those robots or simulations ourselves.
Today, thanks in large part to the maturation of Gazebo and the growing promise of ROS 2, Open Robotics enjoys significant support from non-government sources. In 2013 government funding, primarily from DOD, made up more than 85% of our funding. In 2017 that figure was less than 25%. In the future could we move completely away from government funding? Perhaps, but it’s worth considering all the implications: while we’re very grateful for the support of our industry sponsors, they generally (and understandably) are asking us for near-term improvements and releases that will support their product or service offerings. Those priorities are vital to keep us grounded in the needs of today, but it’s primarily government R&D programs that give us the time and flexibility to take on bigger, riskier, more speculative projects. And in the US, those programs are primarily funded by DOD. I’d prefer to see agencies like NSF have a larger budget, but I’m not in a position to make that happen.
It’s worth noting that this debate is an old one. For an eloquent argument on the other side, see Ben Kuipers at Michigan. Ben doesn’t just make the argument, he lives it, which I greatly respect.
