Slashdot Mirror
Linus Explains his Patch Policy
An anonymous reader writes "For everyone who has been wondering the method behind Linus's seeming madness of accepting or dropping patches, he has finally given a thorough explanation. A must read for anyone who wants to get their favorite feature into the next release of the kernel."
Remember the kid in school that would always say, "My ball, my rules"?
Take note that Linus decided to remind us nine times that it is his tree. I am a big fan of Linux, but not so much of Linus. The way he wrote that letter made him seem a bit childish.
I just wanted to get my thoughts out there. There is no need to mod me down.
Welcome to the land of the free...pay toll ahead...no photography...please open your bag...
I assume that most linux users know how to build a kernel and in the same respect how to apply patches to that kernel (this isn't exactly rocket science.. it wasn't made to confuse you), so are you all really too lazy to build in the patches you want?
transmission_err
I have to say I agree with The Man.
My company (which sells a commercial product to run under Linux) have produced several enhancements to the kernel and have been able to get some of them into the Linus' Tree, some were not accepted, but is now incorporated into a well known Linux Distribution.
It all boils down to what I would call the Mitnik Factor (Tm). Namely how good your social skill is, i.e. how good you are to convince Linux in a PROFESSIONAL way that the patch you have made actually will add a value to the general kernel release and that the whole community will be better off with the patch in Linus' Tree rather than outside of it. (Now that is ofcourse the hard part)
If you mod me down, I *will* introduce you to my sister!
Well, because I was looking for something that was, for lack of a better term, less arbitrary.
Sure Linus does not dictate what each distro has to include, but he is a very influential force, and his statement is pretty much an endorsement of petty personal favoritism.
I am not in the operating system business. I write applications. I would just as soon NOT have to worry about whether a particular user has a particular patch in 'his personal tree' or not. That is just additional support headaches from my standpoint.
Finally, someone who refuses to snivel. I'll bet he's got a strong backbone too!
I hate to say this in such a generalized term, but he's very right that no one is entitled to have their patch accepted. Americans think everything is an entitlement. That isn't so and the rest of the worlds going to get really pissed of and blow something up.
Common Errors in English
I don't know what this patch is or what it does, and I really don't care.
No-one's patch is entitled to be incorporated into Linus' tree. It is his tree, and he puts stuff in there that he feels is the best. Would you really want Linus putting something in his tree which he didn't feel good about or was unsure of? When Linus puts something in his tree, that's his certification that he thinks it's good and useful. Its his word on that in a sense. The minute he starts putting stuff in because people pester him, his word that something is good and useful to his knowledge becomes useless.
Chances are that if the patch is good, Linus will accept it provided he's given enough time to properly evaluate it. Linus is a human being like the rest of us. He can't thoroughly evaluate hundreds of patches coming in a week before the feature-freeze deadline. Try to give him the same breathing room to do a job you'd give anyone else. Also, remember, Linus doesn't have to do anything. He's doing this voluntarily as a service to the public. If you think you're patch is good and useful enough to be incorporated, and Linus rejected it, then go out and prove that its good. Put it in you're own tree or convince a vendor to do so; then people will use it, and if its good, word will get around. Once that happens, more likely than not, Linus will put it in his tree.
I've submitted about a hundred articles to Slashdot, many of them on what I thought were good "your rights online" issues. Do you know how many submissions of mine have been accepted? 1. It was on Creg Ventor, the man who used his own DNA to help sequence the human genome; ironically, I thought that was one of my worst submissions. Yet, believe it or not, you don't see me whining to the editors of Slashdot or in the discussions about it. I realize that many many many other submissions have been made, that the editors have to choose what they feel is best, and that they have to create a variety; I also realize that they're human beings.
Other people would do well to do the same in regards to patches.
social sciences can never use experience to verify their statemen
Figured I'd post a quick summary of the underlying issue.
There is a patch that has strong vendor support (like vendors have already signed contracts involving services from this patch).
This patch is a service offered on many other commercial unixes (Irix, Solaris, AIX, etc..)
Linus considers this patch:
a) to be dangerous
b) to be difficult to test
c) likely to have the most problems on the x86 platform which is Linux's home platform
d) supporting it might add long term maintainability problems to the kernel
The kernel hackers whom Linus trusts seem to agree with his assessment.
What Linus wants is
a) for the vendors to support this patch over a long period of time on a wide range of systems.
b) For there to be some evidence that Linux users (as opposed to Linux vendors) actually want this feature.
So what you have is a fight between big guns: Suse, United Linux, IBM.. and Linus.
Mac OS could be counted as a *BSD, but that is a very recent trend. The Mac following wasn't built because of the new *BSD foundation. OTOH, I know of no one that is using Linux, FreeBSD or OpenBSD that is using it because it the foundation of another semi-emulated OS environment. As an example, I haven't seen anyone that is using Linux and wine/winex/vmware/win4lin/basilisk/etc... as their main operating environment ( sans Linux apps).
Still, if Apple were to release an X86 version of MacOS 10.2, I would buy a copy to try as my main desktop. It would be to try Mac apps, not run FreeBSD specific apps, I would just use FreeBSD if that was my only goal and I assume the goal of most Mac users is to run Mac apps, not FreeBSD stuff. That makes for some stretching to count MacOS 10.x as FreeBSD. Do Mac people run two OSes? If they have to pick between describing their computer as running MacOS or FreeBSD, which would most of them pick?
I have been running *NIX type OSes for 17years starting with VMS on a microvax, basically because I like things that work. Linux seems to like the followthe same combination of requirements. *BSDs tend to have less support for the hardware in my systems than Linux and they both pale to Windows. Sheer support numbers don't mean as much to me and Linux is the happy medium.
Linus and crew have found a formula for developing a *NIX type kernel that many others have decided makes a good foundation for their OS distribution.
Kindness is the language which the deaf can hear and the blind can see. - Mark Twain
why would it be such an honor? It's like you people treat Linux as a motherf**king idol. Heck, I like mac for my photoshop, windows for my games, and linux for everything else, but I'm not an elitist.I don't go around with the delusion that just because someone started something, they are God. Sure, Linus is probably a very nice guy, sure, he's a coding genius, but he has his specific interests and that must formulate part of his choices. anyway, my $0.02
</rant>
Celebrate the finer things in life
Deru kugi wa utareru.
"The protruding nail gets hammered."
Seems appropriate. (And note, this is 'hammered' in a non-beverage-related manner.)
The needs of a diverse user base would be better served by a group of developers with some way to discuss and vote on things that cause arguments. It doesn't have to be formal, but anyway one person shouldn't be able to force his/her opinion if everyone else disagrees.
Of course, Linus is saying that this kind of group can just maintain their own branch. But it will take a big event for people to start using a non-Linus branch and for developers to start submitting patches for it. And users will only loose when they found out that one branch doesn't crash and the other one crashes but supports their digital camera.
I guess it depends on weather Linus wants to keep his work as interesting as possible or put up with some annoying meetings and arguments to make the project as successful as possible. Of course, he has earned our blessing to do anything he wants. But I guess in the first case, someone will eventually make a branch with a critical improvement that he overlooked and Linux will end up as fragmented as *BSD.
That's so tacky. Watch your cache miss rate go way up from all those unnecessary context switches.
A useful thing for developers to try is turning down the tick rate to, say, 5HZ. Everything that polls then becomes glacially slow. Fix those things to be event-driven.
As an example, early Netscape (pre-Mozilla) on the Mac had a major polling problem. Every clock tick, it checked every bookmark to see if it needed to be dimmed out, whether the menu was dropped or not. Large bookmarks lists slowed it down to a crawl. Cranking up the tick rate doesn't fix problems like that.
In a project that's open in its true form, where people from all over the world work on the same project without restrictions, there isn't 1 king with more than 1 agenda.
However, in this case, there is: Linus. I fully agree that you can't include every patch supplied by every developer out there, but his trackrecord clearly shows that he refuses patches for other reasons than crappy coding. (read: political reasons).
Admitted, it's the team that should stick together and you can better favor a teammember than some stranger and neglect the teammember, but that has also the disadvantage that the RESULT of such actions is not that much different than what happens at say Microsoft: there, also a team works on Windows and you can submit ideas and patches (if you can, some can since some organisations have the sourcecode) till doomsday, if the team lead doesn't find these patches and ideas up to par, they're refused and ignored.
So: refusing patches because the code is crappy, agreed. Refusing patches because the teamleader thinks it will destabelize the projectteam or for other unknown reasons, partly agreed, as long as you don't call yourself an open source developer, since the end result is just closed source development where the sourcecode produced is downloadable.
Never underestimate the relief of true separation of Religion and State.
You're missing my point. I'm not talking about "clones" or "ripping off." I'm talking about exact functional copies of software, the only distinguishing characteristic of which is that the developer or developers give away the source. It started with the original GNU programs-- feature-for-feature copies of AT&T's utilities-- and went forward from there. If I make a spreadsheet program, that's one thing. If I replicate the precise features and functions of somebody else's spreadsheet program, that's something else.
There's just no innovation to speak of going on in the open source community. Apple, Microsoft, Sun, and other companies are trying like hell to come up with something new. Sun basically redefined the web application over the past few years with Java and related technologies. Apple is trying to design a user interface from a blank slate, and doing a pretty damn good job. Microsoft... well, say what you want about them, but they're trying like crazy to come up with new ideas like Hailstorm and SOAP. Not every idea is a good one, but at least they're new and different.
Let's see some examples of new ideas in the open source community. KDE and Gnome are fighting it out to see which one can be the blandest, least user-friendly desktop environment. Linux, as neat as it is, is caught between trying to catch up to the leading server OS's, like Solaris or IRIX, and trying to catch up to desktop OS's like OS X and XP. It's doing an okay job of both, but not an exceptional one of either. And think of all the brainpower that's being wasted on dumb ideas like the Mozilla sidebar! If only the community rewarded-- through peer validation or whatever you open-source guys use for currency-- original ideas, instead of incomplete implementations of other people's ideas, we might actually see something revolutionary and interesting come out of the open source community. As it stands right now, all I see is a bunch of projects whose names really ought to start with the words "yet another."
Mod me down if you feel that's the right thing to do. This post is definitely off-topic, except to the extent that I'm extending an idea I introduced upthread. And it's flamebait only inasmuch as I will certainly get flamed for it. It's not a troll, but I'm sure people who disagree with me will hold the opinion that it is.
So moderators, do what you must. But know before you do that I'm just saying what lots and lots of other people are already thinking.
I write in my journal
I'm talking about exact functional copies of software
Even that has been done a million times in closed world. You know what simulatio/emulation means? Think in example of the PC-BIOS. PC's only started of when other companies managed to make functionally exact copies of the IBM Bios. And not this is perfectly legal and okay. Making functionally exact copies with the same interface is not coping like copyright law forbids.
Think of another example, the car. Ford started building cars, with gas and brakes and all that. Then other companies started also to make cars, different internals/details, but functionally just the same. Are you saying that was not okay, and ford should be the only car manufacturer out there? Or the first car company that came out with ABS. ABS is good is it? All other car manufactures copied it's behaviour, you say they shouldn't have?
--
Karma 50, and all I got was this lousy T-Shirt.