To: Gerald R. Lampton who wrote (20225 ) 6/25/1998 12:55:00 AM From: Charles Hughes Respond to of 24154
>>>Obviously, I've done that, and obviously, it gives me the DOS user interface. The fact that Windows 95 is able to emulate DOS does not answer the question of whether it in fact is DOS plus a GUI integrated together. <<< Sorry, I made too many assumptions. If you know programming and especially, if you know DOS and Windows, you know that DOS 7 (the DOS that comes with Win95) needing DOS drivers to run, for instance, the CDROM drive, and Windows turning those off and using separate Windows drivers, a fact you you might have noted during the transfer of control to DOS, is very significant. It means that these two systems have different underpinnings to a great extent. If you can run an old DOS game in the DOS boot mode, it is significant because those games need direct access to the hardware, something Win95 and processor protected mode would not allow if they were running. Actually, there was a long debate at the time as to whether they would ship these products together or separately, it was in all the trade pubs at the time. They could have done either. What we have now is something like a dual-boot system with 2 OS, plus an environment within windows that can emulate DOS for most purposes, probably using a lot of the code from DOS to accomplish that. winNT, on the other hand, appears to only emulate DOS, and not all DOS programs will run because of that. <<<Windows 95 is integrated in the sense that the two functionalities -- DOS and graphical interface -- do not exist separately: the code that is required to produce one also produces the other.<<< I really don't see how this can be anything but an error of the court. In fact you have to program for these systems using completely different tools. They appear to have been written using different tools. The code for one is 16 bit, the code for the other 32 bit. they use different processor modes. They have different APIs. They have incompatible memory modes. It is true that many DOS calls can be reached from Win95. It is also true that you can run many Posix and OS2 calls from Windows NT, and that you can reach many Windows and DOS calls when running Unix if you use Soft Windows, and that under mainframe systems like XA you can pull in various other operating systems in shells. One OS can host another, perhaps even without the 2nd OS having to be modified. Another example is Novell servers, where the underlying Unix system is made to look like DOS, or runs DOS on top of it. You can run NT on an Alpha or PPC chip (usetabe), via emulation software that imitates the 386 instruction set. I have a program written by a friend of mine that lets me run my old CPM 8080 programs on my P333. (Those old Kaypro programs sure look impressively fast at 100 times original speed!) In fact it will do this running in the DOS emulator of NT, an emu inside an emu. The point of all this is that the apparently fluid transfer of control from one os to another has nothing at all to do with whether they are the same code, or even whether they were written for the same CPU chip. The whole thing is being faked. This faking has been pretty easy in the past, because all OS did pretty much the same things, and so you could map things from one environment to another without many problems. We were used to the idea that some OS had windows (generically speaking) and others didn't, that some had threads and others didn't, and there were a few other mapping issues, the techniques for mapping which had also been taken care of over time. This is one of the reasons that programmer folk were so bemused by the idea of IE being part of the OS. We are real used to OS being pretty similar things at the OS level. That and the fact that doing it was so obviously aimed at wiping out Netscape, as Reginald has gleefully pointed out on more than one occassion, rather than being based on any need or market demand. However, this integration/emulation of OS variants has really not been the source of many complaints of anticompetitive behaviour, as each one is distinct, cross-licensed, and so forth. So what does it matter that the judges seemingly had no clue what they were looking at in regards to this particular part of their opinion. I don't see meat there to reverse the decision, do you? Chaz