A common goal of the promoters of Linux and free software in general is getting people to switch from Windows to Linux.
Of course you can do it the hard way and simply make them use Linux, but this way may be a bit like pushing a kid into the deep end of the pool to teach him to swim.
A more subtle approach is slowly replacing the software they use by free software, Firefox instead of Internet Explorer, Thunderbird instead of Outlook, OpenOffice instead of MS Office, ...
This approach will slowly build confidence in free software and may, in the end, even make the user make the switch to Linux, as he already knows the software and thus doesn't have to take such a big step.
The problem only is that many developers of free software already use Linux, which can make it hard to develop for Windows. And free software developed in Windows may not receive the attention free software developed in Linux may get, thus the quality is likely to suffer.
So, to get to the point: Today I was browsing around, reading about the features of the upcoming Fedora 11 and noticed the feature Windows cross compiler, which of course revolves about MinGW, a toolchain running in Linux, compiling for Windows.
Now let me post a little quote from that page (to be found in the paragraph "Benefit to Fedora"):
Secondly — and this is the long term goal — it should mean that Windows users can have a greater choice of higher quality open source applications available to them. (Developers who are primarily Linux users will actually be able to produce daily builds for Windows and respond to bug reports from Windows users, so the quality will generally improve). Windows users can use these applications instead of the usual choice of closed, proprietary, expensive apps, giving them at least some freedom. Once they are used to the high quality free applications available, it will be much easier for them to switch to a fully open platform.
This pretty much corresponds to what I just said, as it pretty much is the reason for this post.
So, to conclude: Using MinGW it may be possible to develop software in Linux that can also run on Windows and, as it might likely be cross-platform, receive more attention than a program that is purely Windows, thus be of higher quality and in the end possibly contribute to users switching away from Windows.
All that said I think it is an important step that the Fedora-guys work in this direction, but it is equally important that users of Linux stop shunning everybody who uses Windows and maybe try to offer the software they create also to those users they always keep telling they should switch to Linux because then the sun will shine so much brighter and because it helps fighting global warming...