UML Merged Into 2.5.34 , VxD's. Do we *need* to use that?

[P. Christeas]

Raul Dias wrote:
> Ilja Kamps <ikarus at ikarus.ath.cx> wrote:
>
> <...>
>
> >and the idea of a wine kernel module is still
> >insane imho, I mean there wouldn't be any real advantagem, unless you go
> >the way ReactOS took.
>
> AFAIK, one area which would improve with a kernel module is speed increase
> in threads that insanely uses mutex.
>
> []'s
> Raul Dias

I do have a philosophical question about that point.
Should every part of Win32 be emulated in Wine?

Such kind of technology tries to enable hw drivers to run under Linux, right? 
Until it reaches a mature state (which seems to me long to come) it will have 
many bugs, which may cause the kernel to die (bad!). Nevertheless, assuming 
that VxD's do work, they may be bad for the future of Linux and the platforms 
Wine runs on. 
This way, for example, HP may say that they will not develop a driver for 
their product, since the Win32 one should run under Wine. They will also 
claim that it's Wine community's fault that the driver is buggy.
One guy suggested that Wine should never reach 1.0, because it will help 
software companies never abandon Win32 for a cross-platform solution. I don't 
fully agree, but drivers may be an issue.
IMHO Wine should only enable existing apps (the ones users heavily depend on) 
to run under other OS's and help users switch to those OS's.