It wouldn't be only the speed increase, but the fact that we'd be
modernizing FreeDOS as a whole.

I think of it this way: What would Microsoft have done had they not gone
exclusively to Windows? I am doubtless they would've migrated MS-DOS to a
32-bit platform years ago. If we were to do such a move, we would not only
make the OS as fast as possible, but also open up roads to modern
development tools and allow the running of apps made by a whole host of
compilers - because not that many (in comparison) support a 16-bit target.
We would make FreeDOS more attractive to programmers because their compiler
would already have the means of generating appropriate code, they would
have no 640 KB (or 1 MB or 1.5 MB...) RAM ceiling to contend with and they
wouldn't have to worry about switching back and forth between protected
mode and real mode to run their programs.

The speed advantages which pure 32-bit would have over 16-bit + DPMI would
not only be from eliminating the delays from entering and exiting protected
mode and the associated memory copies, but also the more subtle details
like variable limit checking and the elimination of the overhead incurred
from having to work with the clumsy segment:offset addressing method.

I have no tasks which I do right now which demand more speed, but my line
of thought is that - if we're trying to take DOS where Microsoft didn't -
then going the 32-bit route would be the way to go.

On the other hand, there's the reality that, while they in all likelihood
Microsoft would've gone to 32-bits... they never did. Their DOS was a
16-bit product and all the software built for it expects it to be that way.
Keeping that tradition intact will do nothing to harm our project, and - as
Mike said - there are plenty of other things we can do to modernize and
improve the project as a whole, especially the userspace. And there's the
naming issue: anything that takes us into the 32-bit world technically is
no longer FreeDOS as we know it.

On Thu, Jan 1, 2015 at 10:44 PM, Dave Pratt <davidpr...@aol.com> wrote:

> Mercury,
>
> It looks like your idea in terms of a benefit of a 32 bit FreeDOS is this:
>
> > FreeDOS would suddenly be the most blazing fast DOS ever conceived.
>
> Fair ?
>
> Why do you think that pure 32 bit will be significantly faster than the
> current model of 16 bit plus DPMI ? (I suppose there is some buffer copy
> that takes place in a DOS extender for each INT 21 call ?  Is that all we
> are trying to get rid of ? )
>
> Why do you think yourself and other users need a faster DOS?  Are there
> tasks you're doing now that are very slow ?
>
> Personally new hardware is so fast compared to what we had in the 1990s
> that I don't see a benefit for most users on any increases in speed, I'm
> curious to hear your experience.
>
> Are there other benefits you see to the 32 bit DOS?
>
> Thanks,
>
> Dave
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming! The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net
> _______________________________________________
> Freedos-devel mailing list
> Freedos-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/freedos-devel
>
>
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel

Reply via email to