He seemed to say pretty clearly up front that Open Source was an assumed good. What I saw as his main point was that a particular reason Open Source is good has been largely overlooked. Namely, that new programmers are growning up and being taught with no foundation in "the classics."

I have believed for a while that you can give someone a dictionary, thesaurus and grammar text, and they won't be a great writer. You can even have them read whatever cannon of classics you choose, and the average person might come to appreciate good writing. But very few will go on to become good writers themselves. That is still a special talent.

But, take away all this background, and the odds of anyone's becoming a great writer are almost nil. How is it, then, that we expect programmers to advance on the achievements of their predecessors if they aren't allowed to see their work?

I'm not saying WindowsXX is great code. But even bad code can teach us ... what not to do. If we can't see the code, we just keep making the same mistakes over and over again. With the code, we can learn what to do, what not to do, and how to spot the difference. If every generation has to re-learn the same lessons, progress slows to a crawl.