What would a computing environment look like that simply didn't support proprietary code on a technical level?
@joepie91 ive thought about that a lot before, it allowed for way better security, debugging (:P) and perhaps trusting trust since there only needs to be one blob at boot and that's it
@joepie91 I mean, technically speaking stuff like WebOS/FirefoxOS and such could be candidates, if it wasn't for HTML/JS obfuscation I guess?
I think only being able to run opensource code is a good start, though hard to avoid obfuscation, and making sure the licenses are correct are another level.. perhaps you could have it only run code from a repository where it requires human verification and integration to set up? Something like the F-droid store for android etc?
@joepie91 Though it is a rather dated and niche model, Lisp machines and later computational environments inspired by them (such as GNU Emacs) come to mind.
One way or another, I think you need to eliminate machine code as a means of software distribution (though things like JIT and caching may still play a role as an implementation detail).
That said, from an environmental standpoint, we cannot have every computer in the world running Gentoo. The resources expended on reproducible tasks like software builds needs to be managed by somebody (not just one organization with a monopoly though).
@joepie91 Could you potentially co-opt some of the existing mechanisms of locking the hardware down? Like the kernel will only run signed code, and somehow *waves hands* limit availability of signatures to code that has verifiable source code availability?
Not sure how that works in practice though...