[Tfug] 2 weeks of Hackintosh fun..

[Bexley Hall]

--- On Mon, 11/10/08, John Gruenenfelder <johng at as.arizona.edu> wrote:

> > "The Linux Driver Myth."
> >
> >      Linux supports more different types of devices than any other
> >      operating system ever has in the history of computing.
> 
> And yet there is still some truth to it...
> 
> Essentially, it all comes down to very new hardware versus
> all other hardware.  The problem is that retail stores (and many
> online stores, for that matter) sell *new* items.

Exactly!  I wrote a driver for my 9-track with Pertec interface.
Wanna know how many *other* people have asked for this *wonderful*
driver to date??  (Hint:  how many 9-track transports have you
encountered in the past 1, 5, 10, 20 years??  :< )

> 1) I wanted to buy a webcam.  Being diligent, I researched
> available models to
> find one that suited my needs and had good reviews.  And,
> certainly, it had to
> be usable in Linux.  I found a great giant list of webcams
> supported and thought "Great!".  Then I went to buy one.

So, you are now paying the "Linux tax" -- having to research
the device and the availability of support *before* heading
out to the store.  I'm sure most cameras advertise "Windows
compatible" right on the side of the box!  And, maybe even
contain a CD so you can get up and running right away.

Unfortunately, you don't see that "Linux compatible" proclamation
on many products...  :<  (and, if it did, it would be followed
by a 5 digit kernel revision, etc.)

> 2) I needed a PCI wifi card for a machine.  Decent
> selection at the store.
> Bought one that was listed as working.  Changed chipset and
> this time there
> was no driver available (except ndiswrapper hackery). 
> Returned it for card
> #2.  Same thing.  Returned it for card #3.  Finally,
> sucess.
>  
> I should make *very* clear that this is NOT the fault of
> Linux, but rather manufacturers who change chipsets at the
> drop of a hat, don't change model numbers, and put insufficient
> information on both boxes and their own websites.

Yes, but I suspect all of those were still supported by their
Windows drivers!  Vendors have a vested interest in their
markets.  They have real money on the line.  If they increase
their costs too much (e.g., supporting a developer who is
maintaining a driver), then they lose their competitive edge
(especially true in this market which is all *commodity*
oriented with very little value added, etc.).  So, they show
what they think the "expected value" of supporting an OS in
very direct ways.  There is no control over Linux kernel
releases and distro support that appears "stable" from outside
that community.  MS has a 3 year cycle.  So, a developer "knows"
he has three years of useful life in his product.  OTOH, if
the Linux kernel "turns left" next week, the developer could find
himself scrambling to come up with *another* driver.  And, annoying
his customers (which translates to phone calls, emails, etc.)
because they don't know why their device isn't working as expected.

Note that I am not saying the Linux camp goes about making changes
willy nilly without thinking about backwards compatibility.  (and
we already know MS doesn't care *too* much about this -- but, they
also tend to follow a *real* release schedule that you can anticipate
for years to come).  I think for a vendor to adopt Linux, they would
have to take control of their own distro -- and then you end up
with everyone picking their own "standard" (so *nothing* works).

Remember, the vendor cares nothing about *who's* OS "succeeds".
They just want to *know* which cart to hitch their horse to...

--don