On 25 May 2006 at 14:53, Robert Patterson wrote: > Leave it to David Fenton to argue with me when I'm agreeing with him.
I wasn't arguing. I was correcting what seemed to me to be misstatements of the history you were outlining. [] > > I believe that current Pentium chips > > use a form of emulation for real mode operations. > > But the chip runs them, which is what I said. Hence, there is no need > for a software emulation layer in the OS, as there was both for 68K on > PPC and now PPC on Intel in the Mac world. But this supports my point, that backward compatibility *could* be maintained if Apple had chosen to do so by getting Motorola to engineer the emulation into the chips they were using. Intel and Microsoft did a lot of affirmative things in terms of engineering and architecture to insure that backward compatibility was more easily implementable. Apple and Motorola chose not to, for whatever reason. This is a topic that's somewhat orthogonal to your points, but it's worth bringing up, I think, given that the difference is not just one of the resources Microsoft had available to throw at the problem, but of planning from the earliest stages of the conception of the major changes in hardware/software platform(s). [] > > You're conflating two different Windows kernels. > > No, I'm not. I know about the NT kernel. . .. That may be, but the comments you made applied only to the Win9x kernel and not to the NT kernel. > . . . I know about the DOS based > kernel. My statement conforms exactly with David's (unnecessary for > the topic) history lesson. By ca. 2000, "all platforms" were no longer > DOS-based. That is, there no longer was a DOS-based kernel. I can't > remember where WinME (DOS-based) fit into the dates, . . . It was released after Win2K, which came out in 1999, so I'm guessing it was released in 2000 or so. I told all my clients to avoid it entirely. Indeed, I had them skip Win98, too, and go to NT 4, and later Win2K (this was before WinXP came out, of course). > . . . and I wasn't > paying much attention to the details of "Home Edition", so it could > actually be a year or two after 2000. WinXP Home is built on the NT kernel. It is nothing more than a crippled version of the full WinXP Workstation, for which you have to pay extra, unfortunately. There is no justification for the existence of WinXP Home as Microsoft designed it, and no one should be buying it at all. Again, that's not a direct response to anything you wrote, just an appropriate detail. But it does show, I think, that your original comments were not clearly informed about the distinctions between the Win9x and NT kernels and which versions of Windows used which. And that was why I commented. > > I don't know why Apple could not figure out how to implement virtual > > memory management without tossing their original OS. > > That is what my original post explains, should David choose to read > it. The original MacOS required programs to access kernel memory, it > was vastly more complex than DOS, especially by the time they were > trying to implement virtual memory management, and Apple gave > themselves only a couple of years, as compared with Microsoft's 10-15 > (depending on how you count). Actually Apple attempted for many years > to do the full monty with the Copland project, but when that failed, > they were forced to do something qui ckly. Enter Unix and OSX. Sorry, but this is not a satisfactory explanation. Microsoft managed to create a DOS virtual machine that can run programs that try to manipulate memory directly by simply providing virtualizations of the actual hardware so the DOS program thinks it's writing to real memory. I don't see why that was not possible for Apple to have done the same thing, unless, of course, it's a matter of the Motorola chips simply not supporting that kind of virtual memory architecture. -- David W. Fenton http://dfenton.com David Fenton Associates http://dfenton.com/DFA/ _______________________________________________ Finale mailing list Finale@shsu.edu http://lists.shsu.edu/mailman/listinfo/finale
