But meanwhile, it's also interesting to read Andrew Tridgell's recent post about his efforts at upgrading the operating system (Ubuntu) on one of his machines. It's amazing that he succeeded at all; I can't think of many engineers that I know of who would have been able to muster up the tools and techniques necessary to break through and resolve this problem. But the reason I thought it was fascinating to read both these articles on the same day was Tridge's description of how the recent work on speeding up Ubuntu boot times by parallelizing the startup work had made it so hard to diagnose the boot failures:
To debug startup problems you need to be able to watch the startup process in action, to see what is waiting. This is much harder these days with the new upstart init system now used in Ubuntu, as startup is much more parallel than it used to be. Adding some echo lines to init scripts used to be a useful technique, but it is much harder to get anything sensible out of that when using upstart.
If you want to learn more about Upstart, here's a good place to start. And then you can find the project's home page for more information.
Really, though, there's no getting around it: parallelism is indeed simultaneously both crucial for future development work, and the source of tremendous complexity and intricacy when things go wrong. Two sides of the same coin.