Message-Id:	<199703210323.WAA12124@keeper.albany.net>
Comments:	Authenticated sender is <jamesl AT mail DOT albany DOT net>
From:	"James Lefavour" <jamesl AT mail DOT albany DOT net>
To:	opendos AT mail DOT tacoma DOT net
Date:	Thu, 20 Mar 1997 22:20:19 -0500
MIME-Version:	1.0
Subject:	Re: [opendos] Wishlist v2.0
Reply-to:	jamesl AT albany DOT net
In-reply-to:	<199703210151.UAA10572@adams.berk.net>

> > Precisely.  That way even users of speech synthesis software and the 
> > like would have access to the full power of the OS, and I think that 
> > everyone would like this solution (anyone who disagrees with this 
> > statement, feel free to speak up). Of course, a 32-bit OS wouldn't 
> > work on a 286 or earlier, but we would still continue to support the 
> > 286 market with the 16-bit OD.
> 
> I think that if it is possible (nothing is impossible somethings are just
> more difficult)
> a gui should developed that supports 16 + 32 bit machines. It seems to me
> that when developers
> develop ( i don't claim to know the depth and breadth of this) a program,
> they seem to "punish" , so to speak those that use slower machines. No like
> any other good program out there, a gui should be written with MODES , 286
> modes, 386+ modes etc..
> 
> how is this done? i don't know . but it would be nice.
> the spirit behind linux was that old hardware needn't be cast into the
> trash, ( i said the L word again <G>)
> but could be used with this operating system, Mind you the effort was only
> halfhearteed, as
> they left out older machines...
As I understand it, writing code that would have these "modes" would 
unnecessarily "bloat" the finished product.  While the source could 
be set up to compile conditionally one way or the other, not all 
users will be working from the source level (one must assume that 
most pc-users will want nothing to do with source, once people really 
start going after it).

Therefore, to keep the executable smaller, one version would be 
compiled for 386 or whatever, and one would be compiled for 
286 machines et c. Since 286 machines can't use 386 code, it would be 
a waste of disk space to have a program on a 286 user's computer that 
includes all the 386 instructions. It can be done, at startup the 
program can determine the cpu type, etc. but the question is, just 
because we can do a thing, does it necessarily follow that we must?

That is why I believe that there will eventually be two versions, 
even if they are compiled from the same source tree.

Jim

jamesl AT albany DOT net - http://www.albany.net/~jamesl/

- Raw text -