What I am doing here is clean-room reverse engineering. Essentially it involves taking existing binary modules and sources of information and writing up what they do in a way that isnt violating the copyright on the origonal (IANAL but I dont think what I posted violates the copyright on the windows binaries). Then it involves someone else who hasnt seen the origonal copyrighted source using the resulting information to re-implement whatever the information describes.
As long as the person who origonally did the reverse engineering never contributes anything to the resulting code, it is (AFAIK, IAMAL) perfectly legal (after all, this is exactly what was done when the first clone PC bios was written, one group disassembled the origonal IBM BIOS and wrote up documents on how it works and how to talk to it. Then another group not connected to the people who did the origonal reverse engineering took the documentation and re-implemented the BIOS from just the documentation without looking at the orgininal IBM implementation)
I never had any plans to actually write code for this (since doing so is out of my skill range) and as long as I never do so in the future (thus "tainting" the code I write as being a potential Derived Work of the microsoft code I looked at), I dont see any way that the results of this investigation could be considered a derived work of any microsoft code.
Although if I have a mis-understanding of this situation, please point out what I am doing that would make code written by someone else using the information in my posts a Derived Work of microsofts origonal binaries.