The problem as that these directives and standards are completely beside the point. They do manage a lot of security-relevant stuff, but in the end what’s hacked isn’t the documentation the standards require but the actual code running on the machine.

And there’s pretty little that the standard can do against actual honest bugs and mistakes in the code.

Was it Singapore that required you to own/rent a dedicated parking space for your car before you can apply for a license?

I got it in 6:05. Nice game.

That’s OOP: https://xkcd.com/927/

If I understand OOP correctly, the NT compatibility promise is the idea that not only the kernel but also the userspace is standardized, backwards compatible and stable.

OOP really doesn’t seem to understand how Linux works under the hood and what the issue is.

The main problem is not that e.g. installed libraries are in different paths under different distros, but that installing libraries is different between different distros, and that different versions are available.

A library package might be called libXYZ-dev in one distro and libXYZ-devel in another, and in the third distro it’s not even available or available in the wrong versions. That’s the real problem, and OOPs approach of splitting package distribution from the stability of the user space is completely contrary to their proposed solution.

They just assumed two contradicting goals would somehow make a coherent whole.

That author just saw Windows and didn’t understand why that system works, saw Linux and didn’t understand why there’s fragmentation problems there and thought that applying the Windows model to Linux somehow magically works.

Edit: I just noticed that the OP and the author of the blog post have the same user name and thus likely are the same person.