From: et472@FreeNet.Carleton.CA   
      
   "R Flowers" (rflowers@Tinbowlinggreen.com) writes:   
   > "Mike Y"  wrote in message   
   > news:1ruXi.168$512.113@newsfe05.lga...   
   >> The Coco could run Flex and OS-9!  Could a VIC-20 or a C-64 ever   
   >> even DREAM of running stuff like that?   
   >>   
   > This is a question out of pure curiosity, I am not trying to be obtuse. What   
   > was special about the CoCo that would allow it to run a multi-tasking OS vs.   
   > the C64? Was there something intrinsic in the hardware (like some kind of   
   > memory management stuff), or was it all just software?   
   >   
   The 6809 CPU was a more advanced design than the 6502 used in the   
   C64.  At the very least, it was designed for re-entrant code.  It   
   also did not have the limited stack, 256 bytes or was it 128? that   
   the 6502 was limited to, and indeed had two stack pointers.  It had   
   a much more extensive set of address modes.  And while I don't think   
   it has much relevance to mutli-tasking, you could write programs   
   that were fully relocateable (ie unlike the 6502 were all the jmp's   
   were absolute and the branches were relative (and limited to plus or   
   minus 128 bytes), the 6809 had a full set of branch instructions, both   
   8 bit and 16 bit).   
      
   Motorola designed the 6809 to be better able to deal with high level   
   languages.  That famous series in Byte about 1979 covered the philosophy   
   of the design well.   
      
   So Motorola went to Microware, that had designed some advanced monitors   
   and operating systems for the 6800 and had them design a high level   
   language, which was BASIC09. Microware on its own initiative came   
   up with OS-9, multi-tasking and multi-user.   
      
   The initial version of OS-9 did require paging hardware, so you could   
   shift memory in and out of the address space and decide where it would   
   sit.   
      
   The second version did not, and that's what ran on the CoCo.  All   
   it needed was regular interrupts to split up the time. Of course, the   
   CoCo was limited in its use of the OS.  With it's software based   
   serial port, that used up a lot of time.  I remember printing things   
   in the background while trying to type in another, and it slowed   
   things down so much that you could watch the cursor moving all over   
   the screen, something that would normally be invisible because it   
   happens so fast.  And of course, with a 1MHz clock and limited serial   
   ports, the ability to have multiple users was beyond most.   
      
   The CoCo III was more like the "mainframe" 6809 systems that were   
   intended for using OS-9.  It certainly seemed designed for OS-9   
   than for anything else.  It did include the required paging hardware   
   so you could move 8K segments of ram into the address space, and   
   designate where it was, and of course had the registers so there   
   was user and system space.  Better interrupt control also helped.   
      
     Michael   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)