[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: OT: Letter to TigerDirect



On Saturday 13 December 2003 03:57 am, Wayne Sitton wrote:
> I mean holly fril, If Gateway had a sticker that guaranteed that every
> piece of hardware on a specific Box was Linux compatible, no matter what
> os was shipped, they'd probably become better than they are currently. 
> But, why do they not get this?????  HP recently announced that they will
> start supporting Debian, but where on their site can I gat a laptop
> pre-installed with Debian?
> 
> I just want these companies to recognize that we buy too, and that we
> want to know if a particular product is Linux compatible.

This is a good idea, which I've had before. In fact, I liked it so much,
I thought maybe I should open a small shop that did precisely this.

Then you find out the problem:  When you slap a sticker on hardware
like that, you aren't just suggesting that it *probably* works with Linux,
you are *making a legal guarantee* that it works with Linux and 
*promising to support it* if it doesn't work on Linux.  That can be an
incredible burden!

Windows (and Sun and Macs) come in relatively few flavors.  Furthermore,
one flavor is always "recommended" and all others "deprecated", because
these companies want to force you to keep upgrading to their latest
and greatest OS.  So, as a hardware vendor, you're okay if you just
support that variety.  What's more, you've got a lot of backup -- you have
suppliers who *also* guarantee the hardware with that OS, and you've
got an OS tech-support number that you can dump people onto if you
can't figure it out, and a legal agreement that they're going to do
something for you (that's what those "designed for Microsoft Foo" stickers
are all about).

Now, if you are just doing this "in good faith" or making a "best effort",
you're asking for a lawsuit the first time somebody buys your stuff and
can't get the "backalley joe" Linux distro version 0.0.2 to work with it.
Either that, or you're just going to be giving people their money-back
an awful lot.

Now there *are* a *few* hardware vendors who do officially support
Linux (e.g. Netgear NICs -- which is why I buy from them. They not only
provide Linux drivers, they provide *source* for their Linux drivers in the
consumer-packaged box.  This is good practice!).

But if you had to build a retail computer business around only those 
suppliers, you'd never get anywhere.  It already sucks badly enough
being a hardware vendor -- hardly anything is more perishable than
the value of computer hardware.  You can *not* afford to hang on
to inventory.  So either you sell in volume, with extremely efficient JIT
supply, or you don't stay in business.  No time to hang around waiting
for a community driver to be written or fixed.

A couple of companies have actually tried to support Linux online.
You may be interested in http://www.elinux.com/ for example.  But,
even they don't really *guarantee* that their hardware works with
Linux (they are also quite biased towards the server rather than
the desktop and have limited selection -- which just means they've
followed their market).

And although, I'm pleased to see that e-Linux is still in business, I'm
not sure that all Linux users are exactly flocking there. So, as an
entrepreneur, it's hard to feel too motivated by this type of business
plan.  It's *not* obvious that the market is sitting there waiting for you.

This kind of dilemma is a direct consequence of Linux's development
model: Distributed open-source development and an indemnity from
responsibility (as in the Gnu disclaimer we put at the top of our
source files), makes it awkward for a company to underwrite our
successes in this way (whereas corporate-written proprietary software
comes with that guarantee, so the vendor can "pass the buck" if
things don't work as advertised).

My point?  If we're going to hang onto our software model (and
we should), then we're probably going to need a new distribution
model for the hardware, too.  I'm still not sure what that model
is, though I have a few ideas about pieces of it.  But it's a problem
to be solved.

Cheers,
Terry

--
Terry Hancock ( hancock at anansispaceworks.com )
Anansi Spaceworks  http://www.anansispaceworks.com



Reply to: