[ILUG] OT: PCI TV tuner card for 50 bills

[Rick Moen]

Quoting David Giles (pppenguin at utvinternet.com):

> After a while enough interest from users results in manufacturers making
> available closed source drivers, some pioneering ones go open source.
> Results in larger uptake of Gnu/Linux.

This is largely mistaken, and unhistorical.  

The history of Linux kernel drivers (along with non-kernel drivers such
as XFree86 userspace drivers and printer filters) is, by and large, NOT
a matter of sitting around and waiting for hardware manufacturers to
drop boons down from heaven upon us.  In fact, we generally don't look
to the manufacturers for significant assistance at all.

Historically, most hardware drivers used in Linux have been developed
independently, through experimentation and reverse-engineering -- with
either minimal information/help from the manufacturers or none at all.
As a generality, driver code _from_ the manufacturers has been very rare
and has either fundamentally sucked, been unmaintainable over the long
haul, or both.

Exception include such things as the Intel e100 driver, developed
in-house and initially not included in the Linux kernel source on
account of patent encumbrances, but eventually accepted after the
patents were licensed royalty-free.  

So, briefly, that was a worthwhile driver that one sought out on a
manufacturer's site, but was nearly unique as such.  By and large,
manufacturer drivers have been abysmal -- which is why we _don't_ in
fact go around pleading for them.  If you want to do something useful,
_suggest_ to (rather than plead with) manufacturers, as appropriate,
that they release meaningful technical specs and sample driver code for
their new chipsets without requiring NDAs.  That will assist the
_community's_ coding talent in developing real, maintainable, non-sucky
drivers that we actually will _want_, instead of half-assed manufacturer
drivers that would be of little use over the long term.

> Wonders briefly why Linux Users seem to want to attact new users but
> then complain they are the wrong sort when they arrive.

This is the usual tiresome rhetoric of "You can't win OS-advocacy wars
unless you do [foo]", where [foo] is something the speaker wants.
Inevitably, the speaker fails to consider whether his correspondent 
gives a tinker's damn about OS-advocacy.  And I, for one, most certainly
do not.

> My thinking is that the larger the linux user base becomes the more
> manufactures will take notice.

Experience suggests that we really _do not benefit_ from "the
manufacturers taking notice".  I guess you must be new to this topic.
Please see:
http://linuxmafia.com/~rick/faq/index.php?page=kicking#hardwaresupport

> If [manufacturers'] experience with releasing closed source
> drivers is positive they may be more likely to consider going open
> source. 

Experience suggests the opposite is more likely:  The pointy-haired
manager in question will more likely say "What, they want open source
drivers too?  We've already given them drivers.  We're done."

[You spent hours of your life you'll never be able to recover, getting
support going for a 0.50 Euro Lucent AMR winmodem]

> But to _me_ that's pointless and not necessary.

Far be it from me to tell you what your time's worth.  But I'd never do
that.

> In the end you can either accept that there is room for all beliefs and
> try to change people's attitudes by reasoned persuasion or you can put a
> big sign on the door 'Lane swimmers only.

I prefer big signs saying "Free clues here!"  You may have seen them:
They're all over my Web site.

-- 
Cheers,     Founding member of the Hyphenation Society, a grassroots-based, 
Rick Moen   not-for-profit, locally-owned-and-operated, cooperatively-managed,
rick at linuxmafia.com     modern-American-English-usage-improvement association.