Windows 9x is most definitely not just a Window Manager; its responsibilities comprise threading, context swtiching and tasking operations as well. It is an OS, just not a completely independent one (for the already stated reason of backwards compatibility). A Window Manager wouldn''t significantly change the nature of executable binaries, nor be able to run its own specific binaries.

Here''s something else: ever tried running a Windows executable in pure DOS? It says "This program requires Windows." The executable format is sufficiently different that the logic is summarily added to every Window exec to ensure DOS does not attempt to run it - just like DOS32. In fact, Windows can be likened to a DOS extender coupled with a Window Manager (GDI) and hardware device abstraction architecture (drivers, etc).

Someone mentioned VM/CMS and got a response about VMWare! (Similar idea though, but VM/CMS is major mainframe OS stuff, like IBM''s OS/390 - which, incidentally is also a VM-capable environment). A lot of misinformation and plain falsehood has been flying around this thread; to break it down, EvilCrap is not entirely clear about the distinction between DOS and BIOS, as someone correctly recognized. BIOS is the service that accepts values such as 0x33 (Mouse) along with values in the BX and DX registers, or so. DOS is a fairly minimal operating system that loads an application and then relinquishes complete control of the computer to the program.

To cut this short, you can write an OS to BIOS, just as you can write an OS that is a DOS executable (which is essentially what the aforementioned DOS extenders, such as DOS4GW.EXE, and Windows, WIN.EXE, were). But you appear to have quite a ways to go before you achieve this. Find out what it takes and how to acquire the necessary skills, and get to it.

Good luck.

---
Those who can do nothing criticize; those who can, critique.

WOW! Creating an OS?! Cool! A great job absolutely. but it sound something hard,don''t it?

Anyway,good luck!

NT, 2000, & XP still use the same interupts to write to disk as MS-DOS v1.00 did

And an error code of 2 still means ''file not found''

Magmai Kai Holmlor
- The disgruntled & disillusioned

i thought interrupts were linked to hardware interrrupts. So calling INT 0x13 would cause a interrrupt in the CPU.

Also, my understanding is that the code in the BIOS gets executed first which causes to read sector 0, track 0 of the hard disk, and the code there re-directs it to read io.sys and then command.com Windows, in my opinion, is not a true OS, since DOS needs to be loaded before Windows is loaded.

quote:

---

Original post by Dean Harding
But those are part of Windows as well. That''s like saying "you should be able to delete winword.exe after you upgrade to Word 2000, since winword.exe was part of Word 97"

---

command.com = DOS command line
win.com = Windows

See the difference?

command.com is just the program that executes the text-based shell. I''m not sure if that even contains the core DOS code, I think IO.SYS and MSDOS.SYS held that (by core DOS code, I mean the code for file system, memory allocation, etc.)

command.com is even present in my version of Windows (XP beta 2.)
When I run it, I don''t see command.com in my taskman. I do see NTVDM, the emulator which runs legacy 16-bit code. When you execute it on Win9x, it should run in its own virtual DOS machine ( I don''t have a 9x install, so I can''t check that.)

quote:

---

NT, 2000, & XP still use the same interupts to write to disk as MS-DOS v1.00 did

---

The NtWriteFile() API calls int 2e to switch to kernel mode. The kernel-mode side then issues several calls into the IO system. It can''t go through the BIOS, since BIOS is 16-bit, so it should (eventually) go to the disk controller driver.