Journey in absurd: open-source Windows

Today’s interesting reading is certainly Stormy Peter’s post about hypothetically open-sourcing Windows, while I agree that the conclusion is that Windows is unlikely to get open sourced any time soon, I don’t sincerely agree on other points.

First of all, I get the impression that she’s suggesting that the only reason Linux exists is to be a Free replacement for WIndows, which is certainly not the case; even if Windows were open-source by nature, I’m sure we’d have Linux, and FreeBSD, and NetBSD, and OpenBSD, and so on so forth. The reason for this is that the whole architecture behind the system is different, and is designed to work for different use-cases. Maybe we wouldn’t have the Linux desktop as we know it by now, but I’m not sure of that either. Maybe the only project that would then not have been created, or that could be then absorbed back into Windows, would be ReactOS.

Then there is another problem: confusing Free Software and Open Source. Even if Microsoft open-sourced Windows, adopting the same code would likely not be possible even for projects like Wine and ReactOS that would be able to use it as it is, because the license might well be incompatible with the rest of them.

And by the way, most of the question could probably be answered by looking at how Apple open sourced big chunks of its operating system . While there is probably no point in even trying to get GNU/Darwin to work, the fact that Apple releases code for most of its basic operating system does provide useful insights for stuff like filesystem hacking and even SCSI MMC commands hacking, even just being able to read its sources. It also provides access to the actual software which for instance give you access to the fsck command for HFS+ volumes on Linux (I should update it by the way).

Or if you prefer, at how Sun created OpenSolaris, although one has to argue that in the latter case there is much more similarity with Linux and the rest of *BSD systems that it says very little about how a similar situation with Windows would turn out to be. And in both cases, people still pay for Solaris and Mac OS X.

In general, I think that if Microsoft were to open-source even just bits of its kernel and basic drivers, the main advantages would again come out of filesystem support (relatively, since the filesystems of FreeBSD, Sun Solaris, NetBSD and OpenBSD are really not that well supported by Linux already), and probably some ACPI support that might be lacking in Linux for now. It would be nice, though, if stuff like WMI would then be understandable.

But since we know already that open-sourcing Windows is something that is likely to happen in conjunction with Duke Nukem Forever release, all this is absolutely absurd and should not be thought too much about.