Hardware Advocacy or Making it easy to do the right thing.
eric.dunbar at gmail.com
Tue Feb 27 04:17:21 GMT 2007
On 26/02/07, Michael T. Richter <ttmrichter at gmail.com> wrote:
> On Mon, 2007-26-02 at 09:04 -0600, Tommy Trussell wrote:
> Surely more manufacturers and vendors will start advertising linuxcompatibility... as this thread points out it will be a real draw insome circles. A few vendors such as newegg.com already include linuxin their keywords, and their customer reviews often include commentsabout compatibility. So there's hope!
> They will not. The problem is one which is fundamentally at
> cross-purposes to the F/OSS movement's philosophy -- particularly the
> "software wants to be free" segment of it.
> Hardware is a strongly-competitive industry full of trade secrets. (One
> of the trade secrets most jealously protected is just how *bad* most
> hardware is under the covers....) The algorithms used, for example, in
> video cards, audio cards and printers tend to be extremely secretive,
> especially the algorithms used to give
everywhere -- by embedding a proprietary blob in a free wrapper. This is
> not a good solution in the long run either because a) the whining of
> militant F/OSS types gets very annoying and likely causes some potential
> Linux supporters to not bother and b) it's a lot of extra work for a fringe
> market again.
> So what kinds of solutions would there be?
> Well, first, we can keep Linux as a hobbyist fringe. (This is a solution
> that many of the more militant wings of the F/OSS movement seem to approve
> of most.) Second, we can lighten up a little and stop hassling the people
> taking the blob approach -- use persuasion to move the blob out into the
> open instead of whining and threats. Third, we can make a stable ABI to
> Linux and stick to it, thus allowing people who maybe aren't quite ready to
> open their source the chance to expand our market for us. Then, again, we
> can use gentle persuasion to move them from pure-proprietary to
> blob-proprietary to openness. Unfortunately the latter two approaches
> require patience and tolerance, something the militant wings -- the
> Stallmanists, if you will -- are not noted for having.
I'm partial to a stable ABI to Linux.
One of the great successes of Macintosh (and Windows) is that you can often
still run ANCIENT apps on current operating systems. For all the touting of
"you will always be able to recompile version xyz of your app for a new
operating system" advantage for Linux, it's not really anything special.
It's definitely not something that's unheard of in the Mac world (or, for
that matter, the evil Windows world ;-P). My father still uses an app from
1992 daily (there's never been a decent replacement for it) and I have
colleagues who constantly use Windows apps copyright 1992 and 1994 on Win XP
and Win 2K.
It would be a significant improvement if vendors could only have to worry
about updating their code once a year rather than every 8 weeks when a new
kernel or library comes out (yes, it's good for development, but it's bad
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the sounder