Oooh. Turns out the FAT file system is patented, and Microsoft has a commercially reasonable licensing plan for it. Of course, given the way the patent system is structured, this kind of thing is absolutely inevitable.
As I've written before, I think Paul Krugman got it wrong when he argued that Microsoft's monopoly position is not much of a threat. On the other hand, kudos to him for publicizing the rotten state of the voting machine industry. Those in the know have been aware of serious problems for some time, but the media so far has been doing a good job of keeping the unwashed masses blissfully unaware. Love Krugman or hate him, it'll be harder to keep this issue quiet now.
A followup to my recent post calling into question Argyll's multidimensional interpolation algorithm. I had some email back and forth with Graeme, and it's now clear that the multidimensional interpolation itself is quite good. The "wiggliness" in my profiles seems to have been caused by an overzealous attempt to optimize the per-channel curves. With those disabled, I'm now getting some really, really good profiles.
I still think that my idea of a diffusion equation based interpolation algorithm might have merit, but considering the good results I'm seeing now from Argyll, it pretty much goes on the back burner along with two or three dozen others. It might become interesting if reasonably priced handheld spectrocolorimeters with a USB port ever make it to market - typical professional profiles are made from 3000 or more patches, but it might be possible to get good results from a much smaller sample, if done right.
More proof links
Claus Dahl sent in a link to yet another proof system/language. In the meantime, I've been corresponding with Norm Megill of Metamath fame, Freek Wiedijk, and Michael Meyling of Hilbert II. There's something addictive about formal proofs. How else to explain why there are so many projects to work on them?