I believe you (and every single sw developer in the world has read that article, besides), but I would argue about the "best". They certainly produce the best quality software in the world for NASA. Is it the most cost-effective per LOC? How fast can changes be implemented? These criteria are much more important to some clients (including my own) than for NASA.
This does not mean that all software should be developed in SEI level 1, but rapid-development practises have their place in industry. They are exceeedingly valuable in the real world.
Furthermore, the quality of that software is not due to whether the developers are middle-aged boring people from Houston rather than hyperactive goatee'd kids in SOMA, but because they observe a structured plan for producing and testing quality software.
> determine the "value" of these people, so they pay "what the market
> pays" and since in the last few years, all the money was free, changing
> jobs got you a raise and it dragged everyone up with it.
Pay is always "what the market pays". Changing jobs gets you a raise because your salary trails your accomplishments and new skills. Not to be snotty, but welcome to capitalism, enjoy your stay.
> Burnout in this field is no different from anywhere else: if you have
> false expectations, you're going to crash when you find out how wrong
> you are. In Kindergarten, I learned that if I run around in circles,
Au contraire, I believe it's because pressure is placed on getting individuals to deliver rather than on instituting appropriate, which is not to say highly structured, but appropriate process. The cult of the heroic coder is alive and well, and the individuals are always called to make up for the defects in the process.
Marco