New or Updated in July 2018

Speculative study eventually must give way to the curious mixture of joy and trouble that comes with writing up some results. You can learn a lot from the research, but for any research that was really worth doing you learn an awful lot more from the discipline of writing it up. My recent research into what Windows really requires for signatures on kernel-mode drivers is especially problematic since I don’t even know where to fit the technical material into this website’s organisation. Meanwhile, a few pages that I’ve tried on results and implications near completion. Both will, inevitably, be revised when I do eventually get round to writing up the technical material that they depend on. Especially welcome to you may be that they each get shortened!


Kernel Mode

While working at that, and following the details through the half dozen releases that pass for Windows 10 service packs, I have at least ventured a first attempt at fitting these releases into existing annotations of structures and functions. The advantage of coming to this late is that I can adapt to how Microsoft itself seems to have settled on referring to the more significant of these updates now that all of them forever more are apparently to be version 10.0. I presently imagine the old major-dot-minor form of version number simply giving way to a year and month, as for 1511, 1607, 1703, 1709 and 1803, with “version 10.0” just assumed. Whether I can keep up with the frequency of these updates is a whole other question. I can’t keep up even as an ordinary Windows user: it’s far from obvious to me that Windows is a better product for this attitude of banging out the code.