That would be interesting, since there are well-established solutions for emulating the XO [laptop.org] in a virtual image (mostly for development purposes). These could probably be adapted to run this modified Windows XP. I imagine that a trimmed-down XP running in a virtual machine would be very useful. It would run quickly and could thus easily fill the gap of running a few Windows apps on an otherwise FLOSS machine.

No doubt Microsoft would try to create license terms to prohibit such usage, but without cooperation from the hardware designers in the OLPC project, I'm not sure they will have any technical ability to lock-out their Windows XP version from being run in virtual machines.

The first is the driving force behind all of Microsoft's actions (and, in fact, almost everyone's): money.

Absolutely. But I think both of your ideas are off the mark (though you start to get it a bit later). The goal, here, probably isn't to make money selling Windows to XO users. In fact, I'll bet dollars to donuts their plan is to give away their port for free. No, the goal is to get people familiarized with Windows products. Remember, the developing world today will be the markets of the future for MS. Having an entire generation of children exposed to Windows could be a very good thing for Microsoft when those economies begin to mature.

More likely because a key factor in the Windows+Office+IE monopoly is its ubiquity. Remember: it's only a "de facto standard" or "industry standard" if, indeed, pretty much everyone and their grandma is inofficially accepting it as a standard, and when your program/format/whatever is subconsciously synonimous with the whole category.

The way it works is like this: (very nearly) everyone uses product X (where X can be Excel, Word, whatever) with its proprietary format Y. At home, at work, etc. The effects are, in no particular order:

1. That it's taken for granted that almost everyone already knows how to work with X, but you might need to train them to use the competitors' equivalent. This is a very big factor when corporations decide to standardize on something. And, at least subconsciously, it's then a factor in what people use at home. You've already used or seen X used at work, so there's no point in wasting your time learning something else instead.

2. Because of 1, knowing how to use program X suddenly is a "skill" you might need at work. You know chances are overwhelming that, unless you're a linux admin or such, the PC at your next job will have X installed on it, and you'll be expected to know how to use it. It might even be an explicit requirement in the job ad. (Remember: training them is expensive, so you might as well hire those who already have the skills you want.)

So the maths already becomes screwed up. If product X costs, say, 500$, it already paid for itself with interest if having that skill saves you even a month of looking for a new job. Or if it lets you move to a job that pays as little as 50$ a month more than your current one, it paid for itself in 10 months flat. "But some other equivalent is free" just lost a lot of appeal in that context.

3. Because "everyone" has program X, thus they "all" can (and do) use its proprietary format Y. (See the recent linked story about even most OOo users saving in .doc or .xls format.) So it becomes the de-facto format of communication, and everyone is supposed to be able to read and write it flawlessly. If you're the guy who can't read format Y, you're as much an oddball as if you were the local luddite without a phone.

And especially for a company, "we don't do Y files" is a big no-no. It doesn't take more than one contract lost with a big customer because you told him you don't want to install Word, to make a bigger loss than buying a retail copy of it for every computer.

This is somewhat easier to get around, since nowadays OOo does a decent job of reading _most_ office files. But, still, the more it gets taken for granted, the more you're expected to be a 100% flawless emulation, down to the 65536 bug. And it gets pointed as a show-stopper if one guy's spreadsheet uses some obscure old function or macro that you don't emulate 100% accurately.

4. Even more importantly, well, you can't have a monopoly on interchangeable separate pieces. That kind of a market can be attacked one product at a time. You want every product to depend on every other product. You want people to say, "yeah, Linux is nice, but does it run the latest version of Word?" and the like.

But to cut this long rant shorter, again, it all boils down to ubiquity. It boils down to the next manager doing any purchasing thinking "naah, _everyone_ knows how to use Windows and Word, but we'd need to retrain everyone if we installed Linux and OOo."

And in that aspect, raising a whole bloody generation of Indians and Chinese on Linux and OOo, is probably something that scares the seven shades of shit out of MS. It's the kind of thing that could lead to "nah, if we're offshoring there and/or importing workers from there anyway, we'd Linux and OOo are for free and we'd need to retrain them for Windows and MS Office." Or worse yet, to realizing, "hmm, everyone there uses ODF, don't they? I guess it would cost us more to force them to accept .xls files." It's that kind of things tha

If they can get Vista running on the XO, I'll eat my shoe.

sure it'll run Vista... you just have to duct-tape this $1000 laptop to it