> >
> >My feeling is that if this would work at all, it would require a significant
> amount of effort. I have to ask myself why would many people want this? It
> seems easier and more reliable to just run win95 on another box joined to your
> linux box by ethernet. Once you've connected the machines via ethernet, you can
> share most everything anyway, so why bother trying to get them both to run on
> one machine?
Well, suppose you wanted to run one of those many cool CD-ROM based
games that are available for Windows...
Seriously, I would find it very convenient to be able to hot key into
and out of a linux/Win95 environment. The latter certainly does suck
(pre-empting a whole lot of flames, I hope:-) but my kids do like to
run doom (in spite of my installing xdoom under linux) and a whole
range of educational and entertainment software and my wife, alas,
uses Power Point and Word to make presentations. Right now I have to
reboot a dual boot machine to facilitate different uses, which means
that I lose NFS mounts (the rebooted machine is a sometimes NFS
server) and a whole bunch of other stuff on my home LAN.
I recognize that it would be awesomely difficult if not "impossible"
to make linux run WinDoze as a symmetric device sharing OS (hell, I
can't get Win95 to configure correctly when it has the whole damn
system to play with, let alone when it's trying to grab devices
through some sort of virtual interface). I do, however, hold out some
wee bit of hope for a Windows API emulator, e.g. -- Wine or the
equivalent, capable of running generic games and software under the
LINUX environment -- no real devices need apply.
Even this, of course, is awesomely difficult, since Microsoft isn't
exactly jumping to provide API support to emulator writers that enable
other OS's to run Windows software without paying them (Microsoft)
money.
I'd like to propose a different software item. It is very clearly
difficult to write a good emulator of a proprietary package, partly
because you are always reverse engineering the parts the owner want to
keep hidden, and partly because as soon as you succeed they know how
to make a trivial change that will cause your efforts to fail. It is,
perhaps, easier to write a software converter. I'd like to see a
"compiler" written that reads in a windows application, converts its
code more or less permanently to linux/X format with suitable widget
library and device replacements, and writes it out to disk. Call it,
say, execonvert. Run "execonvert doom doomlinux" and voila! A linux
runnable version of the actual distribution windows doom.
Advantages to conversion as opposed to emulation: You could gradually
tune the conversion process to get a good matchup at the windowing
level, and then tune it for performance. If the product worked, it
would enable ALL the makers of Win95 software to release a linux
version TOMORROW, which would really put linux in direct competion
with Win95 (a thing sorely needed). It would actually be the first
step toward liberating the entire web from "OS-iry". Make the
conversion process an automated prerun step at the OS level and one OS
is suddenly able to run software written for many.
If emulation is possible, binary conversion is possible (and far
easier). The only real question is whether or not it could really be
completely automated, or whether it would need to be an interactive
program (choose where to put the menus or their format, finagle its
I/O/ streams, handle format conversions to printers, etc.).
rgb
Robert G. Brown http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567 Fax: 919-660-2525 email:rgb@phy.duke.edu