I'm a bit confused, what about that publication provides ANY additional information to the thread?
Or are you insinuating I need a crash course in "computers"? Kernighan is a legend and a good writer but I would be surprised if the average IEEE journal reader didn't already know everything he wrote about...
In my intro to operating systems class, Stanford used "NACHOS" (not another completely heuristic operating system). I loved that course, but it looks like they have switched to "Pintos" more recently.
This (in my opinion) shows there is of course no perfect model, and I'll give the benefit of the doubt that a good teacher will always search out what they think will help students the most...
I have no idea why you think using the kernel from the most popular OS in the world as a study aid is a bad thing. Personally, the last few set-top boxes I have worked on have been Linux, and I think that whole industry is pretty much dominated by the Linux kernel. I haven't done any serious Windows development in years. But for many new grads a solid understanding of the Windows kernel would be invaluable to their future jobs. In the end, at the kernel level most of the fundamental design principles are pretty much the same - what *I* am amazed at is how many candidates I interview don't even know the basics of virtual memory, disk I/O, process scheduling, multithreading/sychronization, etc.
Once again, I'm trying not to pass judgement... but read your post and decide if it added anything useful. The only concrete phrase in the whole thing was "direct observation and investigation", which yes, is provable, and no, you haven't shown that the MS program does not include.