Conversely, I can think of a lot of crappy programmers who have not ever contributed to an open source project or even put together their own website.
I understand that you find that belief "objectionable" and "pervasive". Have you considered the possibility that it might be correct? Because you have failed to present any arguments against it.
There's a pretty strong empirical argument for it, just from the observed populations, but there are also theoretical reasons to believe it:
• People who only program at work, and on closed-source projects, only ever get feedback on their programming within the context of work, from the other people who work at the same company, probably on the same project. Most projects and many companies are staffed entirely with bad programmers, so it's very likely that you'll never get a good programmer to look at your code that way, so you'll never get mentorship; you'll have to learn everything about programming by trial and error, which would take many lifetimes. At many companies, it's even worse than that — there's no code review.
• People who practice a skill 80 hours a week improve much faster than people who only practice it 40 hours a week. Many people don't even have the opportunity to program for anything close to 40 hours a week at work.
• People who practice something because they must, rather than because they love it, will never improve beyond the minimal level of competency demanded of them, because that takes further effort.
Now, there's a limit to how much you can productively practice a skill, and maybe you have a job where Rob Pike vets your checkins and that will suck up as much effort as you can manage to throw at it. But there was probably a time when you didn't, and there'll probably be another time when you don't.
1: This isn't really a relevant point. If all of the other people in their office are amazing programmers then maybe they are getting fabulous feedback. My complaint really doesn't address this and it's an entirely unknown since you don't have any information about anyone's particular work environment. That said, I'll agree that at a place without code reviews you don't even have a shot at getting good feedback.
2. Now this is definitely a valid complaint. That said I wonder how many different technologies people who work on some side projects work with on average in a month. Perhaps this problem is balanced by the folks who only program at work having more depth in their chosen technology but less breadth overall. Again, valid complaint... but it's likely more complicated than you've made it out to be.
3. This is just not true... People have all sorts of reasons they choose to master a particular skill. You make it sound as if someone can't put their heart into learning their trade simply because they have a belief that it's their job to perform at their very best and not just at a "minimally competent" level.
Maybe the opinion exists because it is true.. but your bullet points don't provide compelling arguments and they reek of the same elitism I was originally talking about.
With regard to #1, your chances of working your way up to getting code review comments from Linus Torvalds or Rob Pike is a lot better if you're contributing to (other people's!) open-source projects than in almost any non-OSS work environment.
I don't think people have this attitude because of these theoretical reasons, though. I think they believe it for empirical reasons — because they don't know any first-class programmers who only code at work and don't work on open-source software. Do you? I was just trying to explain the evidence I've observed, not trying to convince you from first principles.
Though it occurs to me that participating in OSS would allow you to work with far more developers than you would otherwise.Though I'd argue that it isn't really as direct contact as the workplace.