Re: 2.1.83: Sound, SB16, Modules, MIDI?

James Mastros (root@jennifer-unix.dyn.ml.org)
Wed, 4 Feb 1998 20:52:12 -0500 (EST)

On Wed, 4 Feb 1998, Jim Bray wrote:
[...]
> How does one go about doing this? I used to do a lot of kernel-hacking
> back in the early '80s, and have provided the odd patch here and there,
> but I don't do politics well and developed the sense that Linux was sort
> of divided up into fiefdoms and that I had little chance of getting stuff
> into the kernel.
It depends. There are people who are generaly recognized as uber-hackers...
Alan, Bill, Linus (in alphebetical order by first name -- not in order of
importance <G>). There contributions are generaly going to make it into the
offical tree, sooner or later. There are also maintainers of vairous peices
(mostly drivers). They can do pretty much whatever they want to those
peices, and Linus will take patch-sets from them. But Linus can change any
code he wants, though if you are modifing a mantained driver, it is
considered more civilized to go through them.

> Combine that with natural sloth and aversion to
> flame-wars and one has a formidable barrier.
<G>

[...]
> I've offered to touch up the readme.modules (several times) but have
> gotten no response.
The general fealing is that ideas are cheap. Working patches are much more
dificult. Write a patch, and post it on linux-kernel. Comments will start
pooring in, more then likely.

> My suggestions about the module stuff would either
> have to be backwards-compatible or would require touch-ups to every
> modular driver; in either case I would be hesitant to wade into it without
> a go-ahead from those who own those areas of code.
The normal way seems to be:
1) Write the New Way in, breaking all the drivers.
2) Change one driver as an example.
3) Write a sepperate patch that steps on everyone's toes, or declare
everything obsolete <G>.

> I am willing, as I
> said, to at least do some research on it if the Linux equivalent of the
> group managers offer me some political coverage or whatever it is one is
> supposed to have to do this, which I make no pretense of understanding.
Nothing is needed. Write the code, then worry about getting it in the
offical tree.

-=- James Mastros

-- 
   "I'd feel worse if it was the first time.  I'd feel better if it was
   the last."  
   	-=- "(from some Niven book, doubtless not original there)" 
	    (qtd. by Chris Smith)