Subject: WinSand
From: Dean Beeler <canadacow@softhome.net>
Date: Wed, 02 Jun 1999 22:37:12 -0500

I was suggesting the whole "fake" a wrapper for the WinModem drivers
long before it was popular but since I failed to get anyone's attention
back then, I decided to take a little action on my own.  In the case of
PCTel's WinModem's, they have 3 VxD's that link to the COMM.VxD present
in the USER32.VxD library.  All that is needed it a simulation of the
COMM.VxD and Window's VxD loading services.  Rudimentary .INF
interpretting could be implemented to allow users to install their
WinModems (now LinModems) as easily as they did under Windoze 95.  Once
the VxD connects to our simulated COMM.VxD, our service would represent
itself as one of the serial character devices in Linux, /dev/cua2 or
something and away we go.  There are several advantages to this
approach:
1. It could potentially hit several birds with one stone.  If done
correctly, we could
immediately support several WinModems instead of just one at a time as
we hacked
out specs.
2. When they upgrade their WinModem drivers, they can also upgrade their
LinModem drivers.
3. The same source code to do this could potentially get other Windows
only devices to work under Linux.

Disadvantages of this are:
1. Misses the goal of the original LinModem.org's intention to turn a
$1,500 multi-tasking, multi-threading OS into a $50 answering machine.

Ok... enough of trying to convince people I have a good idea.  If we
just keep talking we'll just end up with a lot of conversation and no
LinModems.  Presently, I am canabalizing source code from the most
current version of Wine to implement VxD services under Linux.  I don't
know who said that Wine can't run Direct-X programs, but I run the
Windows 95 Quake II full screen in Linux with it with no problems.
Perhaps you should download and compile the new source.  It does take
quite a bit to compile on ones own system (160MB for me).

Dean