(This was meant to be a brief reply to this post, but it ballooned into a post all its own.)
Open source automotive software makes sense on paper. I am all for it, even though it would obviate my employment as a test engineer for an automotive supplier. So let me tell you why it won’t happen.
- Automotive manufacturers do not understand open source. They only see the liability, not the potential. If anything happens in the field requiring a recall, they are on the hook anyway. So they assume they should control the whole process and develop everything themselves from scratch. I’m not entirely sure some of them understand software itself, as evidenced by John Deere presuming ownership over the software in their tractors.
- Who would test this open source software (besides the manufacturer)? With web software, you can spin up a remote server on a cheap laptop and you are off and running, in relatively little time, at low cost. With automotive software, the platform costs are at least 5 figures, not to mention the specialized tools to upgrade the software of the various components. It is a much larger barrier to entry for third-party testing.
Again, I agree in principle that it would be a good idea. But it will never happen unless lawmakers force it to, and they do not have a great sense of open source software either. Of course, it will be interesting to see if traditional software companies (Google and Apple) will bring this mentality whey they enter the space in a few years time.