From: Dick Taylor <dtaylor AT monmouth DOT com>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: 32bit/64bit DOS (was 32-bit DOS)
Date: Mon, 15 Sep 1997 15:37:33 -0400
Organization: EverGreenSys Inc
Lines: 35
Message-ID: <341D8E7D.1BB2@monmouth.com>
References: <199709142320 DOT JAA09381 AT solwarra DOT gbrmpa DOT gov DOT au> <5vic98$f6h$1 AT vnetnews DOT value DOT net> <5vjftp$avk AT top DOT mitre DOT org>
NNTP-Posting-Host: fh-ppp2.monmouth.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

Michael F Brenner wrote: (a thoughtful position on the reasons for
desiring a 32-bit DOS)

    After thinking a bit more, I'd like to support this idea
    with the following proposal. Why not generalize DJGPP in
    the following ways,
       a. provide a small newOS .h and .lib which can be invoked
          in any regular DJGPP DOS load module.
       b. create_newDOS() would invoke autoFAT and find an EMPTY 
          absolute disk sector and put FAT entries and boot code
          on the principal HD.
       c. test_newDOS() would cause a re-boot of the system
          with a hook to the newDOS loader, which would then
          re-activate the calling program
       d. the calling program from that point on would only have
          BIOS support and could perform whatever scheduling
          event, I/O or other tasks it requires.
       e. run_newOS() would function as a stub, invoking a
          complete, mature version of newOS whenever required
       f. examine_oldOS() would run under either the kernel
          or from a regular DOS DJGPP load module and would
          evaluate its current environment and save parameters
          needed for any newOS.
       g. examine_PC() would collect facts about the current
          physical environment and save them

       It would seem like there would be only several minor
       changes to an otherwise fairly complete OS support
       system as it stands. Any programmer can therefore 
       create whatever parts of a "complete" operating system
       they choose to, and in theory it should run on any PC
       platform. Not only 32-bit, but perhaps in 5 years 64-bit
       architectures as well.

       Dick Taylor