That's all true, but consider how many vulnerabilities we'd be seen if the programmers who wrote those applications also had to deal with pointers and memory management. You'd get the current problems plus all of the stuff that regularly snags systems programmers.
I agree and we can take it further. If we had to hand encode the instructions it would cause even more bugs. The more decisions you have to make as a programmer, the more chances there are of making errors. High level languages make some decisions for you, but people still end up making mistakes. Maybe the goal should be to increase the bar at which a programmer is considered competent.
I think you look at it a bit incorrectly. Those high level languages were designed not to avoid mistakes, but to push software complexity to a whole new level. You simply won't be able to get to that complexity with a lower level language, because people have limited ability to think about things and tend to fallback to solutions they can fit in their mind. Which is why you shouldn't expect high level languages to eliminate vulnerabilities, people still use their minds to the limits and make mistakes. Other approaches must be taken.
I think we agree. Its hard to encapsulate everything, so I just chose a smaller context for my argument. I was primarily thinking about resource allocation and such.
A HLL does indeed allow you to compose solutions to problems in terms of higher level abstractions, which I gather is your point.
