I am stuck with Windows IE 8.0 for as long as I am stuck with
Vectorvest. I've tried it under Crossover Office but the browser crashes a hell of a lot. Vectorvest is a very unique bit of software but is written by simpletons in the form of some odd visual basic browser plugin and ASP on the server side. This software is incredibly old and because Microsoft maintains a reasonably stable ABI, Vectorvest never need to change it. Microsoft's long term dominance seems to be less about embrace, extend, extinguish these days are more about: comfort developers, maintain support until the end of time itself, and allow them to never change anything.
When you already can avoid keeping up with Windows releases as much as possible, would you even bother with the effort of porting to another platform? From a business perspective it means taking on an expense of keeping up to date with things (or apt and version changes will simply outgrow the product that uses ooold libs) that the same people won't need to do with Windows.
Not all Windows programs have this luxury. Browsers tend to need to keep up with Windows (independent taskbar thumbnails per tab, on 7 for example). Opera has to keep up with Qt because for a long time it used Qt 1.3 and just about everything even the usually ancient debian lenny uses Qt-2.0. This made it slower to release 10.3 on Linux. This here is the problem, the downstream tards at just about every distro never consider the benefit of keeping older libs, especially when there is no conflict in having this backwards compatibility any more.
I think Microsoft will sit on Windows 7 for a while and maintain compatibility with it in anything they release in a future. Some of the games and software of today in a decade could keep people using Windows as the dominant OS for a long time to come.