On 8 Jun 2003 at 21:17, Philip M. Aker wrote: > On Saturday, Jun 7, 2003, at 13:23 America/Vancouver, David W. Fenton > wrote: > > >> Multitasking as an issue is only important to Windows developers > >> because DOS (and I think Windows before W95) never had anything like > >> the (old) MacOS event loop concept in the first place. > > > Eh? > > DOS was single-tasking, so is not a relevant comparison to any version > > of Mac OS. > > But you just noted to Jari: "Uh, DOS underlies all the Win9x kernels > (Win95, Win98, WinME)..."
Yes, and you're in Windows, not DOS. Finale has certain features common to both the Windows and Mac versions. The fact that the OS underlying Finale on a Mac is not the same as the one on Windows does not change the fact that Finale has its own internal workings that are independent of the OS. That the Win9x kernels had DOS underneath, at some level, does not mean that the the restrictions of DOS apply to Win9x. If it did, Win9x wouldn't be usable at all. Really, most of this is about how often the CPU has to switch into REAL mode. The 32-bit file I/O system means that real mode is not used much, and DOS is bypassed for a lot of things that it was designed to do. Mostly DOS is just the OS loader (though it's not quite that simple). And DOS windows are launched as applications within Windows (i.e., child processes of Windows itself). This was true in Win3.x, as well. What was not true in Win3.x was that any programs got their own protected memory space. In Win9x, Win32 programs run in protected memory with full pre- emptive multitasking and multithreading. All Win16 programs (including DOS) run in a single virtual machine, so a crash in one Win16 program is likely to take down all Win16 programs. Since the Win9x kernel depends on certain Win16 subsystems for direct hardward access that run in the common Win16 virtual machine, this can take down the whole OS. This is not the case in NT-based versions of Windows, of course, as each 16-bit process gets its own virtual machine, and hardware access is never direct, but always through the Hardware Abstraction Layer. > So I'll stand corrected on when exactly DOS died but have to ask why > not? DOS was the Microsoft system at the time. At *what* time? You haven't specified. Cooperative multi-tasking was native to Windows before 1990 (how long before I can't say exactly, as I first started using Windows with version 3.0 in 1990). The limitations of DOS, while relevant to the overall stability of Windows (and to some parts of memory management) certainly did not limit the ability of Windows to multi-task. > > Win3.x was used cooperative multi-tasking, just like Mac OS, and used > > a messaging queue, as any multi-tasking operating environment must. > > > Win95 introduced a combination of cooperative multi-tasking and > > preemptive multi-tasking, depending on whether the software was fully > > 32-bit or depended on 16-bit components. > > > To claim that the "event loop" was an advantage that made > > multi-tasking unnecessary is rather ridiculous. > > Firstly, I claimed that "Multitasking as an issue is only important to > Windows developers...". Why would that be? Multitasking is something that all GUIs provide. > Secondly, I've previously explained to Jari that the keyword in my > phrase was "concept". The whole Macintosh concept was first realized in > 1984. I'm not sure of the date, but certainly before the end of 1986, > Apple released AUX (Apple Unix); which is multitasking by definition > and also supported a concurrently running MacOS that had all GUI > features of that period. So the statement: > > > Microsoft got it right long before Apple, by introducing decent > > multi-tasking by Windows 3.0 > > is very misleading because obviously Apple had a multitasking system at > their disposal years before Microsoft but for one reason or another > decided it was not advantageous to use it at the time. Perhaps the > Amiga fiasco scared them off a bit. I saw Windows 3.0 many years ago. > It was pathetic. OK, I was trying to make sense of what you said, which to me was nonsensical. You said Windows programmers had to deal with multi- tasking and Mac programmers didn't. I took that to me that your remarks about the Mac messaging loop were not counted by you as multitasking. I didn't understand what you could possibly mean by that, but took you at what I understood to be your word. > Finally, I have to question whether either you or Jari truly understand > that the (old) MacOS had it's way of arriving at solutions that MS has > to use multitasking for. There you go -- I cannot make any logical sense out of what you've just said. First you say Mac has multitasking. Then you say it doesn't. I thought Macs multitasked (cooperatively) from the release of the very first Mac in 1984 (or whenever it was). They allowed you to run more than one application at a time, right? So, they were multitasking. Now, *how* they did it is another issue, and one where it's pretty clear they did it differently from what MS implemented in Windows. But the MS approach was better in the long run, as it worked for a lot longer, and laid the groundwork for real multitasking much sooner than the Mac got it. > . . . If I could use your previous example from the > repertoire (and BTW don't you mean K. 333?): Yes. I always second-guess myself on the K. numbers for that group of sonatas. > > The classic example of this is the Classical era notation of the > > apoggiatura. Take Mozart's K. 332 Sonata in Bb, which begins with the > > falling figure, 16th-note appoggiatura, 8th note, 16th, 16th. It is > > played as 4 16ths, and the late 19th century Mozart edition published > > by B&H transcribed it as that. > > > But the result is that it obscres the musical significance of that > > first note. The original notation makes quite clear that it is a > > non-harmonic tone, that the one notated as the 8th note is the > > harmonic tone. It points out to the reader that the passage starts > > with a harmonic dissonance. > > To attempt an analogy, I think that you and Jari have been deceived by > only being exposed to and only analyzing the B&H version of the Mac > score. And furthermore, don't understand that AppleEvents give the > Mac--and hence its event loop--an extra dimension that makes it > transcendent to Microsoft's notion. And prior to MacOS X, transcendent > to any unix as well. Then why did Apple have to trash their whole OS and use a different one to implement real multitasking and virtual memory? That is, if it was so great, why did it prevent Apple from creating a modern, stable, scalable OS out of it? -- David W. Fenton http://www.bway.net/~dfenton David Fenton Associates http://www.bway.net/~dfassoc _______________________________________________ Finale mailing list [EMAIL PROTECTED] http://mail.shsu.edu/mailman/listinfo/finale
