lwn.net just featured an article the sustainability of open source, which seems to be a bit a topic in various places since a while. I’ve made a keynote at Siemens Linux Community Event 2018 last year which lends itself to a different take on all this:

The slides for those who don’t like videos.

This talk was mostly aimed at managers of engineering teams and projects with fairly little experience in shipping open source, and much less experience in shipping open source through upstream cross vendor projects like the kernel. It goes through all the usual failings and missteps and explains why an upstream first strategy is the right one, but with a twist: Instead of technical reasons, it’s all based on economical considerations of why open source is succeeding. Fundamentally it’s not about the better software, or the cheaper prize, or that the software freedoms are a good thing worth supporting.

Instead open source is eating the world because it enables a much more competitive software market. And all the best practices around open development are just to enable that highly competitive market. Instead of arguing that open source has open development and strongly favours public discussions because that results in better collaboration and better software we put on the economic lens, and private discussions become insider trading and collusions. And that’s just not considered cool in a competitive market. Similar arguments can be made with everything else going on in open source projects.

Circling back to the list of articles at the top I think it’s worth looking at the sustainability of open source as an economic issue of an extremely competitive market, in other words, as a market failure: Occasionally the result is that no one gets paid, the customers only receive a sub-par product with all costs externalized - costs like keeping up with security issues. And like with other market failures, a solution needs to be externally imposed through regulations, taxation and transfers to internalize all the costs again into the product’s prize. Frankly no idea how that would look like in practice though.

Anyway, just a thought, but good enough a reason to finally publish the recording and slides of my talk, which covers this just in passing in an offhand remark.

Update: Fix slides link.