But software is* different because it's in total crisis, have been for maybe two-three decades now, in any case so long that most software developers who are notoriously young people have never seen a simple and working software system in their life. Maybe in 1970's software development paradigms were not any better but problems were limited by the software size and complexity limitations coming from the hardware and languages available.
*) describing the situation as it is now, not what I would want it to be; in my opinion, software should be evaluated against the same standards as hardware, buildings, bridges, and so on, depending on where it is used.
Now, no one seems to have an idea how to build even remotely robust software, no one seems to have an idea how to build any software for less than say $100 million, especially if taxpayer is involved. No one even dares to discuss about responsibilities. There's a lot of software science telling us how to develop robust software but all of this is failing in practice.
Software simply doesn't need to work at all, no one bats an eye. For example here in the Great Finland we have this new healthcare IT system which reportedly by doctors prevents proper care completely, nearly or actually killing people left and right (for example, by completely blocking the availability of any patient information, including allergies to medicines, in emergency situations in ICU), but no one can do anything, and no one is and will be held responsible. It was so expensive it has to be used despite how many human lives it takes. So we can't even say "hey, let's free everybody from responsibility and just drop this software, revert back to the old working system, and forget about this ever happened"; no, we have to continue making human sacrifices on the altar of modern software paradigms.
Hardware on the other hand, not so much. The situation is so much better we can afford to have this discussion about responsibilities, remove the few bad apples. Besides, no one buys hardware that is completely non-functional but for some strange reason, with software this doesn't seem to matter; I can't think of anything else but it's because people are so desperate with software, anything at all goes.
And yes, I again agree with mr. Animal that starting holding software developers responsible for the worst offends (like killing people directly or indirectly) would help solve the software crisis. The task is enormous but it has to start somewhere. On the technical side I applaud mr. Animal's efforts regarding C standard library improvements/replacements to get rid of all the built-in traps in current specifications and implementations.