Software Engineering discussion

I don't think that code gets any more complex than the synchronization and priority code in an OS. As the authors discussed, the testing and debug (especially post mortem) are nightmarish. It would have been interesting to see how Linux approaches this Solaris function/problem.

Ugly code that can't be made any more beautiful equals beautiful code?

I like to think about mutex problems, because they are like fun puzzles. I like the problem they are working on; it's very puzzle-like. Although, debugging real world synchronization bugs is miserable.

I would guess there are some pattern application in this chapter, but I didn't exactly identify which patterns were used or which patterns may work better.