[lug] Linux can use Windows drivers shock

[Chris Riddoch]

Chan Kar Heng <karheng at softhome.net> writes:
> hopefully interesting...
> got the link from someone...
> http://www.theinquirer.net/?article=12207

Yeah, it's interesting. I'm skeptical of its real value to Linux
folks, though.  From the article:

> This, in an ironic twist of "embrace and extend" would allow Linux
> to claim compatibility with "off the shelf Windows LAN drivers",
> while encouraging hardware vendors that prefer to deliver better
> (native) Linux drivers to continue writing their own.

I don't see how this encourages native driver development *at all*.  I
do see, on the other hand, how it encourages hardware vendors to
program to the status quo, Windows drivers, and continue to take the
attitude that if the Linux folks can't get it to work with the
compatibility layer, well, it'll still be unsupported.

It would be great if hardware vendors would actually develop and
support Linux drivers, and even better if they would offer their
hardware specs and free software drivers.  Failing that, the next best
thing is to have hardware specs available, but no drivers.  The worst
situation we can be in is to have neither hardware specs *nor*
drivers, and have to resort to reverse engineering (by those in
countries where it's legal to do so) in order to get anything useful
at all.

I think OpenOffice is great, specifically because we now have a good
codebase for development and can tweak it if necessary.  As Free
Enough Software, it doesn't leave us at anyone's mercy.  I think Wine
is also great, because it helps people migrate, and provides a stopgap
measure for those decreasing categories of software there isn't a Free
equivalent of -- but I do consider it stopgap.

The idea of running hardware drivers through emulation layers seems
even more stopgap.  This is where my ideological half and my pragmatic
half can't quite agree: it's good in the sense that people can use the
layer to make use of hardware they otherwise can't, but it's bad in
that it can provide an excuse to hardware vendors to not release specs
on their hardware.  Without specs, we are denied the *opportunity* to
make a better Free driver, and we are restricted to only what the
hardware vendor is willing to offer us.  It could take substantially
longer for Free drivers to be written for Linux if a driver layer is
available.  How strong is that argument?  Depends.  It's an inference
based on a hypothetical situation, take it for what it is.

Either way, don't go out and buy any old hardware just because its
driver will now be supported by this compatibility layer.  We need to
demonstrate our ideals with our checkbooks and buy hardware from
vendors that help the development of Free drivers with (at the very
least) published specifications.

It's also worth pointing out that a fair amount of the instability of
Windows has been reported to be due to crashy third-party hardware
drivers.  I can't vouch for the fact myself, but if it rings true at
all, we're better off without them anyway.  We want more eyes on the
crash-prone parts of the system.

-- 
Christopher.Riddoch at colorado.edu
 - epistemological humility -