Still, I think you are vastly underestimating "the artist-who-dabbles-in-code".
I'm sure you can find a few in the world who can do a lot. And a lot more who can't. There's no point in discussing an undefined person.
Also, don't under-estimate the power of aesthetics in logical situations, it does wonders for solving problems you haven't thought of. It is a means of problem avoidance.
I love to make things more elegant, but not at all costs, and if I can't find an elegant way to add the functionality in a timely manner I'll go the less elegant route. Better hacky than not there at all.
Have thousands of people work on something for free. Give it away for free, and see if does not have any sort of disruptive effect on a market where competitive products are expensive. Although Torvalds is well of, Gates is much better off.
Torvalds doesn't own the product, so IBM's or Red Hat's bank account would be a better measure. And Microsoft is a bit absurd of an example for you to be using of the elegant way of doing things -- MS is infamous for its hacky approach and bugs, it's probably better used as an illustration of how what looks like a mess (see OOXML specs) can succeed in the market.
Anyhow, linux being free of charge is irrelivant to the matter -- especially since businesses pay many hundreds of dollars for RHEL and SLED, and as Microsoft likes to point out the TCO for linux is higher since the techies cost more. Most of the time it's used instead of Windows or OS X because it's judged better for the purpose, not because of price. Windows has come pre-installed on my computers, thus free, but that never stops me from deleting it in favor of a more stable, robust and flexible OS.
My point about unhealthy obsession with clean code is best illustrated as Theo De Raadt syndrome. Says Theo on why Linux sucks and OpenBSD rules: "It's terrible. Everyone is using it, and they don't realize how bad it is. And the Linux people will just stick with it and add to it rather than stepping back and saying, 'This is garbage and we should fix it.' [...] Linux has never been about quality. There are so many parts of the system that are just these cheap little hacks, and it happens to run."
I'm sure Theo's kernel is prettier, but that doesn't make it better when it comes at the cost of the robustness of the product. Hence you see the market largely ignoring OpenBSD (since you have this notion that being $0 is linux's trick, keep in mind that OpenBSD is also free). I can respect Theo's approach, or at least I could if he weren't always a jerk about it, but no matter how much he talks about how linux is a heap of junk the plain and simple reality of the matter is that linux has proven itself in practice and held up to continued development, which is what really matters.
In the same vein, from
this article:
"You know what I found? Right in the kernel, in the heart of the operating system, I found a developer's comment that said, 'Does this belong here?' What kind of confidence does that inspire? Right then I knew it was time to switch."
Developer code comments are not the sanest way to judge a product.
If I may I'd like to disagree with this. I find that even on my "one-person" projects, a good SVN or CVS setup goes a long way. Not only does it provide for a way to leave a project for months on end without losing the structure and logs, but it makes management of it simpler.
I think any "serious" project needs to start with a revision control system.
I've looked into CVS and subversion and they don't seem to offer me much. I have my own release management scripts which serve my needs -- automating the creation and upload of a final release, parking a copy of the version in an archive where my comparison tools make use of the history in various automated ways, automatically updating my websites to reflect the new releases and so on.
Maybe I'll give them another look sometime if I run into a problem with my system, but I judge my needs by what's working for me. Of course my situation is rather unique since I've spent my entire professional life working on what's essentially one project (outside of a few little one-offs on the side). My point is simply that there are unique circumstances out there.