Is this 'problem' real? Or is it just something coders have programmed themselves to believe?
There is a difference, I think.
From my experience, in the case of musicians, because of the creative and instinctual nature, and the very real underlying rules of rhythm and music theory, you start with that basis and move on from it. It can be a painful process with independent personalities and such- but there is very real common ground.
With development, that foundation has to be laid. A bad design and bad architecture can kill an entire project from the beginning... but it spends a lot of time and energy in dying. It's not for a lack of wanting to, and I think there is nothing bad in trying. But there really needs to be a guiding force or somethings laid out from the beginning that are not likely to be.
Look at the coding jams and such- they're more gimmicky than anything, producing things, but none of them being really lucrative or even finished. Sometimes they are the building blocks of something else to come along later. But more often, they are forgotten, swept aside, and never reach fruition.
For programming projects, more than music, there are roles for a reason, as much as I hate the structure. You need a champion/project owner that has the vision that people submit to... and everything is measured against that vision- and you have his development side- the architect/designer. Then the project managers that keep it rolling along. I always like to separate testing from anything else... but some disagree on that point too. Then user acceptance... which goes back to the project owner. As much as I hate it... it works. And without it, I've seen commercial projects fail terribly.