> I don't want to help bad programs run; I want to stop them from running.

This attitude is the OPPOSITE of what Linus Torvalds advocates and may be the primary reason why Linux is more successful than OpenBSD.

I should add that successful software may be valuable to some but less to others.

This roughly matches my experience with OpenBSD. OpenBSD and any software that runs on it is absolutely rock solid.

However, it's not necessarily easy to install any of the mediocrely-written, but possibly useful, software that is abundantly available for Linux.

This may or may not make sense depending on what you are doing.

Right, http://en.wikipedia.org/wiki/Worse_is_better

I used to have a SparcStation running NetBSD that I used in a way similar to what this paper proposes - testing my software in an environment where some of my (implicit) assumptions were no longer true (a non-Linux system running on a big-endian CPU that caused a bus error on non-aligned memory accesses) to uncover bugs.

So, I like the idea.

When I read the title I originally thought it's about Perl 5, hostile development environments and communities. But it's only about about hostile hackers.