SimonC>According to VA some 3,000 people licensed VA (this was reported here in the past) - when Drobe conducted a poll of users a relatively small number of those 3,000 responded. Do I really have to join the dots here ? *If* VA kept some people using RISC OS for a while it only did so temporarily and I strongly suspect these are now fully paid up Windows only users.
Emulation makes it a more risky proposition for hardware developers to produce *native* (that is ARM based non-windows dependant machines) primarily because the manufacters can't easily gauge how many likely purchasers they're going to get and may be less likely to risk it when faced by a compeditor who has much lower development costs - and hence one who has a lesser risk.
As to assertions that Emulators are *faster* than native - for the most part that is groundless, where VA *has* a speed advantage (disk access) this is largely due to Windows caching writes in RAM. And I don't know about you I'd much rather have my data safely tucked away on my HD rather than floating in some volatile fashion in RAM (even if the former is slower).
As to the dubious "faster cheaper" argument for PC's emulating RISC OS first off [and my appologies for reshashing this] PC's have higher running costs (power, AV updates, maintainance time), faster is arguable (and if you're a "fan" of VA chances are any comments I make you'll ignore anyway), finally emulation is just that - an *approximation* of real hardware. This means you can't be *absolutely* certain that it's performance will exactly match with real RISC OS/ARM hardware in any event.
Besides I'd argue it the other way, if there were no RISC OS native hardware I'd see no real point in emulating it under Windows (or Linux for that matter). Once the only way of running RISC OS is under emulation then it's arguably no more a "live" platform than an emulated Sinclair Spectrum..... and Linux and Windows based apps do enough to be useful without resorting to the "fiction" of running a RISC OS app under emulation.