More on Comes v. Microsoft
Some more stuff has been put up at Groklaw in the Comes v. Microsoft case. Specifically, an expert report regarding (alleged) undocumented API’s in Windows, and the transcript of that expert’s deposition. I’d definitely suggest checking out one or both if you have the time and the geeky mindset to enjoy them.
I had originally intended to just skim through both, but I ended up reading both word for word, including the 96-page transcript. (The report is quite large, but there’s only 16 pages of actual reading. The rest is all references.) I must admit that, having read through the transcript, it certainly seems that MS’s lawyers have brushed up on their geek terms. While there were a few statements that made me chuckle, there were no blatant displays of igorance.
A few things that are of particular interest here:
- The line drawn between the OS and the applications, blurry at best, is constantly shifting to meet Microsoft’s needs. (The defense lawyer at one point calls it a “bright line”. Yeah… right…)
- The age of some of the code being reviewed. While most of the review focuses on Windows XP and Vista, a majority of the deposition seems to be centered around old code reviews of Windows 95/98 from previous trials.
- Some references to excellent reading in the course of details of things the expert witness had read in preparation for the deposition.
The line being drawn between applications and the OS (1) is, in my opinion, ridiculous. And it would seem that Andrew Schulman (the expert) agrees with me on that count. (I’d link to a specific page here, but the transcript is a PDF.) Microsoft appears to be trying to say, based on interpretation of a previous ruling, that anything distributed as part of a separate application (such as, say, Internet Explorer) is not a part of the Windows OS. At face value that doesn’t sound all that bad. If IE7 includes a particular file, it is part of IE7, not a part of Windows, right?
Well, not exactly. Lets say that IE7 includes an ntoskrnl.exe file. (I know, not a DLL, but just bear with me here.) Now if IE7 provides an ntoskrnl.exe file that is not the same as the one shipped with Windows, does that suddenly make it a part of IE7 and not Windows? Well, of course not! It’s the kernel! However, given the interpretation of the previous ruling, this seems to be exactly what MS is insinuating. With stuff like this going on it’s no wonder that there have been no substantial changes in MS’s behavior since the ‘98 trial.
The other interesting bit (3) is a reference to this paper, presented by Ken Thompson in 1995. If you’re not comfortable reading program source code you probably won’t be able to make heads or tails of this, but it’s an excellent article. It spells out a way to perpetually propagate an exploit of your choice through a system. By compromising a single program one time (in this case the C compiler) it is possible to make all future program builds “infected” even when compiled from perfectly clean source. I’d definitely suggest giving it a read if you’re anywhere near calling yourself a geek.
