[nycbug-talk] OpenBSD and blobs
carton at Ivy.NET
Thu Dec 14 16:56:35 EST 2006
>>>>> "ms" == Marc Spitzer <mspitzer at gmail.com> writes:
ms> If some one wants to provide a different alternitive, open
ms> source driver for example, fine but t it is not on the table
ms> yet is it?
Yeah the ``if someone wants to put something on the table I'll
reconsider'' is I think a key motivating strategy from the Linux camp.
The users have this sense of entitlement, which I suspect comes from
the Linux zealots trying to convince people to switch from Windows:
installing on their relatives' and friends' machines and such, who
then say ``but you claimed it was better,'' and now that they have
some shadow of a choice instantly turn into arrogant
customer-is-always-right consumers haggling with the sales guy like
good little capitalists. The people asked to use/fund Linux get to
feel good about themselves. The back-pressure the friends and
relatives exert while it might not be effective on BSD developers, it
_does_ work well on many Linux developers. And obviously Linux is
doing pretty well. I think it's partly because this type of advocacy
worked and not purely for other reasons.
BSD since its beginnings has competed for developers instead of users.
so my first impulse as a BSD guy hearing someone say, ``well if you
guys aren't going to put anything else on the table for me, I'll just
go use Linux'' is, (0) Cool, no problem, (1) That guy's annoying, and
(2) ``Both myself and the linux people have been clear about what
software freedom means, so rather than repeating myself on those
points I'll leave you to choose whatever short- or long-term strategy
you like, confident that you need justify your choice to no one but
yourself.'' The Linux advocate at this point would probably instead
start arguing with you, painting portraits of future dystopias,
villifying corporations, launching ad-hominem attacks.
The BSD style of advocacy feels a hell of a lot more neighborly,
clear-headed, and rhetorically honest to me. But I think it's also
been a failure. For example, Jeff Quast substantially expanded on (2)
above by linking to an OpenBSD paper, which I found really interesting
but I don't think Mark Spitzer's rebuttal acknowledged the arguments
in the paper at all. The paper's out there, is reasonable, even IMHO
compelling, but it's just not effective at capturing attention or
ensuring our platform's future. OTOH I'm sure if like 20 people on
this list all ganged up on Mark like rowdy children and said ``MARC U
R A big st2pidh3ad'' he would respond somehow---eventually it could
get so nasty he would have to either leave or comply. It sucks, but I
think the Linux advocacy approach has a far superior track record, yet
obviously I'm still not willing to adopt it.
Secondly the Linux way of approaching hardware has been to make as
many drivers as possible: the end goal is to ``convert'' machines, so
it has to support everything. No one keeps a list of preferred Good
drivers vs. mediocre drivers and steers people toward hardware with
Good drivers---instead it's ``how many pieces of this laptop I just
bought have some kind of driver attached to them.'' Instead the goal
is to run on Dell's motherboard-of-the-week. They basically set out
to pander to Mark's position, to the best of their ability.
BSD, again since its beginnings with BSD/OS, was never interested in
supporting everything. Instead the BSDI's position was, ``If you want
to run BSD/OS, buy hardware off this short menu of well-supported
things. Please don't try to use other devices because they'll work
poorly or not at all.'' Now, in the present, Theo is saying, ``try to
buy the Taiwanese chips like Realtek gigabit and Ralink
wireless---they'll work the best because they give us documentation.''
very much in the BSD tradition.
This second tradition is also failing us because, as the paper Jeff
cited points out, modern laptops and even servers are so integrated by
their OEM's that it's no longer possible to choose hardware. We have
to resort to buying eBay laptops just to get stuff that mostly-works.
We fill up cardbus slots and have to pack wads of silly USB dongles
and cables to duplicate built-in laptop hardware for which we don't
have drivers. We use underpowered overpriced boards like Soekris just
because the hardware doesn't burry us with major revisions from one
chip stepping to the next.
I really doubt we can survive in this environment too much longer.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 304 bytes
Desc: not available
More information about the talk