On 26 Nov 2005 at 20:36, Simon Troup wrote: > >This strikes me as arbitrary. Virtual PC does not take up any > >significant processor cycles, unless you are running a Win program, > >so the only practical objection would be A) the cost or B) the disk > >space. Cross-grading Fin from Mac to Win is no more costly than any > >Fin upgrade. They come on the same disc. Virtual PC does have an > >associated cost (c. $200 last time I checked), most of which is the > >Win license. I can accept cost as an objection, but not disc space, > >and not the nebulous objection quoted above. > > I don't really think I need to explain why configuring an OS within an > emulator in another OS is unattractive. I've used Virtual PC with Win > 98 SE, Win XP and Win 2000 - it's buggy and has a tremendous > performance hit, things don't work at anything like native windows > speed.
But you're missing the major point that Johannes brought up here (in the form of a question): would VPC really be an *emulator* any longer, since there'd no longer be a non-compatible motherboard underneath OS X? Much of the overhead in VPC is going to be in a layer that the whole of VPC sits on top of that translates Windows hardware calls into something that OS X, sitting under VPC, can handle. This won't always be a matter of just a straight translation -- sometimes there won't be any corresponding analogous access method, and instead something will have to handled in some other way. The level of complexity could be huge. But if there's no longer a foreign hardware platform, VPC could then talk directly to OS X's hardware abstraction layer (I can't think of a scenario where it would work to have VPC talk to the hardware directly through its own HAL, since that could lead to contention between the two OS's), which could possibly improve the performance enormously. Now, it could be that there's not much difference there, since the OS X HAL would remain the same, but it's *possible* that it could work better. > It also absolutely requires you to learn and manage another OS, > another whole font library etc. etc. - when you are your own IT > department there's a lot to be said for keeping it simple. The scenario that Robert Patterson described, where VPC is just a virtual machine in which your FinWin runs (just as Mac Classic is a virtual machine in which an older program can run), would be a scenario where the emulator itself takes away all those issues for you, providing a layer that gives your Windows application access to the OS X resources as though they were native to Windows. And Finale really has virtually no administrative overhead from the standpoint of the OS -- Finale is well-engineered in being extremely self-contained so that it doesn't get all tied up in the OS and require configuration changes. The only issue would be configuring your MIDI, and the hope would be that VPC would automatically detect OS X's MIDI support and give you access by default, the same way Windows running natively detects and auto-configures MIDI hardware. And keep in mind that the simply virtual machine model is the way Windows emulators already work when running on top of Linux, and, indeed, distributions like Lindows (or whatever it's called now) are built entirely in this way, with the idea of giving you Windows with painless Linux underneath. > Where does this idea comes from that VPC functions only as a window > server? (and I use the term in a unix sense) - it doesn't, and it > doesn't make an argument to pretend that one day it might, as I doubt > it ever will. You mean like an x-terminal? I don't think anybody at all is suggesting anything of the sort. Robert Patterson's suggestion was that each Windows app could get its own virtual machine, instead of, I believe, the current scenario where there's a single virtual machine, and all the Windows apps run within that. Chances are good that VPC running on PPC would be prohibitively resource hungry in that scenario, whereas MacIntel might give enough performance improvement to make it possible to run them separately. An alternative would be to have a single VM running the underyling emulator, and then have child VMs that provide the GUI and communicate with the underlying VPC VM that provices the low-level Windows layer. There are any number of possibilities here, and the speculation is all based on the fact that we don't know which direction things will go or how things will shake out. We don't know if MakeMusic will take 3 years to provide a native MacIntel version of Finale -- their history with supporting OS X is not promising in that regard. And if that's the case, given the possibilities for how VPC may be able to be improved, it opens up all the options that Johannes has speculated about. Of course, you *are* correct that it's all speculation. But you don't know that your view of the path history will take is any more likely than that forecast by others as possible. -- David W. Fenton http://www.bway.net/~dfenton David Fenton Associates http://www.bway.net/~dfassoc _______________________________________________ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
