In <[EMAIL PROTECTED]>, on 04/30/98
at 11:00 PM, Steve Curry <[EMAIL PROTECTED]> said:
>>How about a Win32-Linux Project? There is already a Win32-OS2 Project and
>>it is producing a PE to LX converter (pe2lx.exe) that converts Win32
>>programs into native OS/2 programs...no recompile necessary...and
>>performance is essentially equivalent. This is rather easily possible
>>with OS/2 because it supports (incompletely) Open32 which is based on
>>Win32. IBM left out some functions when they incorporated Open32 into
>>OS/2 but those holes are being filled by a small group of independent
>>programmers (http://www.os2ss.com/win32-os2.html). Could something along
>>this line work for Linux? Perhaps WINE could be molded into this kind of
>>project instead...
>SURVEY SAYS!!!! Good idea! ; )
I certainly understand the desire for developers to make native Linux
solutions that are on par or better than software available on
Winblows...the same feeling/idea sits in the OS/2 community (at this
stage, it is far more likely that Linux would succeed at this goal rather
than OS/2, at least until OS/2 is folded into a JavaOS...). The problem
is getting started. If you want a lot of people who would not consider
Linux before to actually take it seriously, they will want to run the
software they use today (doesn't have to be M$ software but some other
ISV). That is where a Win32-OS/2 or Win32-Linux Project comes in. It is
certainly far easier, and more likely to succeed, to convert non-M$
software since ISVs don't, as a rule, change the operating system with the
installation of their software (new/replacement DLLs, for example). M$ is
the culprit of constantly altering the os with every install of M$
software. In this vein, it is considered unlikely in the near term that
such programs as Office97 will be successfully converted by the Win32-OS2
Project converter. You CAN convert all the DLLs and .exe's to native
OS/2-Open32 format, certainly, but it wont run because M$ has made their
software load into the >512MB address space and the present version of
OS/2 doesn't support this (there is hope that this limit will be
eliminated soon in a fixpack). A problem like this, however, should not
affect Linux (am I right?).
The idea behind the Win32-OS2 Project is to make the operating system more
attractive by making it possible to run "your" favorite Windblows software
on a better os/different os. Once people are using it in more numbers, a
critical mass of users may drive the development of native solutions
rather than merely Win32 versions. There is a danger in this, however,
that ISVs could simply say, "Why spend time and money developing OS/2 (or
Linux) native apps when all we need to do is make a Win32 version and let
the enduser convert it as necessary?"
The native version runs a bit faster and has more certain full
functionality.
The advantage to the Win32-OS/2 Project is that M$ really cannot alter
Win32 all that much. To do so would break not only the vast bulk of their
own software, but the software used by millions of people from other
developers. This would only serve to REALLY anger a LOT of people who
would suddenly be REQUIRED to buy new software to replace their perfectly
fine software. And one other thing...it looks possible that the DOJ may
require that M$ fully release their API for ALL developers with full
documentation. There is at least one case before them in complaint that
M$ alters their API and uses hidden, undocumented features to hurt their
competitors and make their own software "seem" better/faster/more stable
(if stable can be used in the same breath with M$). If that happens, then
they could do nothing at all that wouldn't be immediately available to any
developer. A conversion project group would also have access to all the
api functions and be able to keep up to date for proper conversion of ALL
win32 software.
As it is, the Win32-OS2 Project team only needs to add the occasional DLL
to add/provide some functionality or new feature that is left out of
Open32. This is a far cry from having to rework from scratch every time
M$ screws up their api.
There are advantages and disadvantages to a conversion project, but it can
certainly provide an "in" to getting more butts into the Linux seats.
patrick
--
This message brought to you by OS/2 Warp 4
-----------------------------------------------------------
[EMAIL PROTECTED]
-----------------------------------------------------------
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
