FAQ: Questions not to ask on linux-kernel.

C. Scott Ananian (cananian@lcs.mit.edu)
Sat, 16 May 1998 05:41:57 -0400 (EDT)

----------- THE FAQ YOU'VE BEEN WAITING FOR: -----------

--> Questions not to ask on linux-kernel <--

Silly algorithm/implementation questions:
1) Why don't we toss gzip and move to bzip for everything?

bzip is *much* more CPU-expensive and needs external buffers to decompress.
gzip can decompress in place, which it why we use it to compress the
kernel image. The sources are bzipped on kernel.org as a
convenience. gzip is faster, gzip is preferred. If the bzipped
images cause too much griping, they will be removed.

2) Why don't we internationalize all the kernel messages?

Because we prefer maintainable code.
This has been discussed to death many times before.

Silly kernel compile/config questions:
3) Why won't Linus put kernel X.Y.Z in directory linux-X.Y.Z instead
of putting everything under linux/ ?

Because he doesn't want to. He's the boss. End of thread.

4) I have kernel problems. I'm using gcc 2.8.x/egcs/pgcc.

These compilers have "issues". Compile with and see if your
problem goes away before complaining about it.

5) I get funny messages about objdump when I try to compile the kernel.

Upgrade binutils the way linux/Documentation/Changes tells you to.
You really do have to do
rm `which encaps`
like the documentation says. We write these things for our sanity,
not for our health. 'nuff said.

6) Why does ifconfig report strange numbers after I upgrade to 2.1.x?
In 2.0.x it worked.

Upgrade net-tools like linux/Documentation/Changes tells you to.
'nuff said.

7) Why do I have to keep commenting out SMP=1. It's *so*

Michael Chastain is fixing this, as part of a general kernel
makefiles overhaul. It takes time, he needs testers, fixing
everything across all the platforms linux supports is non-trivial.
Be patient. It'll be done by 2.2.

Silly kernel features questions:
8) Why doesn't linux have a spiffy boot logo?

Answer 1:
It does. Look through the archives for any one of several
different patches.
Answer 2:
Because no one has taken the time to implement it cleanly.
It has ceased to be an interesting topic for hand-waving.
Implement something, and we'll talk about it.
Answer 3:
Because it can be done better in init or lilo.

9) I want {Unicode,UTF-8,native charsets,klingon} in the kernel.

Enough already! Everyone has different goals/wants.
'<insert language name> should have an 8-bit encoding' only amuses
us the first time we read it. If your post on the topic does not
*directly relate* to a *specific* kernel API (citing code!), then
most of us don't care. If you design the subsystem, you get to
make the rules. Ted T'so has declared that ext2fs "by default"
will use UTF-8, but no one's forcing you to rename your files.
End of discussion.

10) Why isn't GGI in the kernel?
Because we're at the tail end of a development kernel cycle/very start
of a stable kernel cycle. Linus had some issues with the first
GGI/KGI/EvStack code he saw, but that's not unusual for the first
implementation of any kernel system. The GGI folk are
rearchitecting their code, and they've promised to come back when
we're into the next development cycle (kernel 2.3.x) and let us take
another look. Until then, discussion is closed.

11) Why isn't there a STREAMS implementation in the kernel?

Because it would make (all of) the kernel very slow. You can
implement it in user-space, if you need to. It *will* be slow.

Miscellaneous flame-wars:
12) The gnome-kde debate does not belong on linux-kernel.
13) The emacs-vi debate does not belong on linux-kernel.
14) Porting linux to run under windows is not interesting.
15) Whether a given intel CISC chip "is really RISC" is not

Final thoughts:
16) Linus is boss. When he speaks, the discussion should end.
17) Linus is wise. When he asks for changes to your patch, he's
usually right.
18) Linus is busy. Patches do not get incorporated immediately and
sometimes things get overlooked. Be persistent, but not
impatient. Getting your patch well-tested by daring folks on
linux-kernel is a good way to help it along. (Think: "my patch
*shouldn't* break anything" versus "8 people have tested this on
3 different architectures and no one has reported a problem".)

19) Newbies should not be flamed.
20) Old-bie flames should be taken off-line.
21) If you want to see a feature, implement it instead of complaining
about it.


I've probably forgotten some of our favorite flame-food. Oh, well.

All responses to this message should be private.
I certainly don't want this to start *yet another* flame war on

Feel free to save a copy of this and forward it to anyone who tries to
bring up a question from the list in linux-kernel.
@ @
C. Scott Ananian: cananian@lcs.mit.edu / Declare the Truth boldly and
Laboratory for Computer Science/Crypto / without hindrance.
Massachusetts Institute of Technology /META-PARRESIAS AKOLUTOS:Acts 28:31
-.-. .-.. .. ..-. ..-. --- .-. -.. ... -.-. --- - - .- -. .- -. .. .- -.
PGP key available via finger and from http://www.pdos.lcs.mit.edu/~cananian

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu