PE>> The next message contains what I conceive as the way the public
 PE>> domain dos replacement should contain.  I thought that was reasonably
 PE>> straightforward until I found out that real mode was real mode and
 PE>> protected mode was protected mode, and never the twain shall meet.
 PE>> Can someone tell me whether it is technically feasible.

 RS> Well, its certainly damned difficult and you have to ask what the point
 RS> of it all is with something as primitive and brain dead as DOS.

It's popular, and for someone who just wants to do a bit of WP, it would
seem reasonable.  Actually, someone suggested that I should see what
commercial applications there are for a basic version of DOS in
"rugged environments" (or something like that).  Like PC's that
are used to control a softdrink machine.

 RS> For starters you have to completely redo all the bios code, in all roms
 RS> on all cards which you well ever use. Thats one area which brought OS2
 RS> unstuck pretty severely and is a massive problem alone.

Why do I need to redo all the bios code?  I was planning on using it lock
stock and barrel.

 RS> Taken at the ultimate its not possible to just wave a magic wand and
 RS> allow any old DOS apps to run in protected mode on your flash protected
 RS> mode version of DOS anyway. The best you can do is do it the way OS2
 RS> does it, completely replace everything that the DOS app thinks is under
 RS> it with new stuff, and the virtual machine for the DOS app to run in.
 RS> But thats essentially just OS2 with the fancy stuff like multitasking
 RS> and VM and the OS2 API discarded. A massive operation.

I'm certainly not trying to create a massive operation.  What I have in
mind is a clean 32-bit OS, for both application programs and the operating
system to see, but still make it use the existing BIOS.  If I have to
translate addresses internally, or if I have to read sectors into low
memory, and then coyp them into high memory, it doesn't worry me a bit.

I'm also only planning to produce only as much functionality as is provided
by MSDOS.SYS and IO.SYS, ie about 80k worth of 8086 assembler.  Do you
think that replacing 80k of assembler in C is a massive operation?  BFN.

Paul

--- GoldED/2 2.42.G0614