Slashdot Mirror
Linux Approaching A Fork In The Road?
Hai|_ Hai|_ writes "Linux vs. Linux? ZDNet has an interesting article about the potential for a forking of Linux distributions into incompatible OSes. " Nothing we haven't seen before, the article just basically says "Linux could fork". Well duh, welcome to open source. The bigger questions are 'Will it' and if it did, 'Would it matter?'
Consider, Linux can handle SCO binaries. Solaris software can be ported with little more than a compile. Many Windows and DOS programs will run )Wine and DOSEMU). For a bit more money, NT, Win98, and all of their software will run under vmware. Linux is able to mount MANY other OSes' filesystems.
From PC based server to mainframe to handheld, to embedded device, just recompile and you're ported. Try that with any other OS (BSD might do it, I don't know, but certainly no closed OSes).
The reason for that is that the community wants it that way. Even if someone does fork Linux significantly, someone else will want to make a patch to allow for compatability.
At this point, the only significant vendor that seems to actively fight this unforking is MS with it's 32 bazillion poorly documented API entry points and a lawyer behind every tree. (If you don't think they're poorly documented, then why cant the NT (Win2000) and Win98 development groups manage to agree on them).
I really don't understand the article's claim that the backup software failed because of library locations. The libraries can be anywhere as long as the ld.so.conf is correct unless the software is linked against absolute locations for some bizarre reason.
[rant] I just have to get this off my chest. Caldera is the most MS like of all distros based on the horror stories I've heard. It even has a sort of registry (with all of the problems that entails). I'm all for making configuration easier for beginners, but come on![/rant]
Yeah! It's a very flexible windowing system, and has stood the test of time.
X! Sheesh. Forget X, if we're going to get rid of a letter, let's start with the bloody useless C. It has two pronunciations, both of which are handled already by another letter.
Look to the Indonesian national language, Bahasa Indonesia: they've done a smart thing and reassigned C to the 'ch' sound. We should just follow through and reassign X to 'sh'. And while we're at it, annex Y and use it for 'th'. Bingo: we've gotten rid of three useless letters and made "h" mostly unambiguous.
At the LinuxWorld Expo in NYC at the start of Feb he said that from a technical perspective virtually everything about fragmentation is good. Unix gave it a bad name. But the problem lies in being different just to be different. However it is important to be able to create a version customized to your needs, one size does not fit all, yadda, yadda, yadda. Just so long as your changes are recirculated and you don't introduce gratuitous incompatibilities, fragmentation is simply not an issue!
:-)
It was...a unique answer.
Cheers,
Ben
My usual seat in the cluetrain is at A HREF="http://pub4.ezboard.com/biwethey.ht
It looks like a troll-moderator or astroturfer-moderater has moderated the parent down, presumably because it is unflattering to Microsoft.
This is inappropriate.
Someone with moderator priveleges today please correct this.
The Future of Human Evolution: Autonomy
Linus owns the Linux trademark. The way I see it, if Linus tacks on some type of compatibility designation like "Linux 1.0 Compliant" for certain distros then many will feel like they have no choice for consumer type OSes than to follow Linus' lead.
You can still fork the code for embedded products or linux appliances, but if you're marketing a Consumer or Server OS then you'd better not be the oddball who isn't Linux 1.0 compliant.
LK
"Hi. This is my friend, Jack Shit, and you don't know him." - Lord Kano
Whatever happened to a little depth in journalism? You would think that before someone would undertake writing an article like this, they might actually do a little research...!
It seems patently obvious to me that the writer has neither heard one of Eric Raymond's recent talks, nor read the essay Homesteading the Noosphere -- or at least if he did, he completely missed the part about "cultural taboos in the hacker community." As a writer myself, I'm completely put off by this shoddy, "let's get a few sound bites" type of work trying to pass itself off as print journalism: if you want to write news that way, get a job in television, where depth and research quality always play second fiddle to the terse and pithy.
Oh, and a little technical competence is probably in order too: Linux is a kernel. Library layouts are a feature of a particular distribution (although the LSB is trying to fix that) -- different distributions can do whatever they want with them without forking "Linux". If the article was supposed to be about incompatibilities among distributions, write about that... but a little checking among sysadmins would reveal how easy many of those issues are to overcome; and if there really is a compatibility problem running Legato on Debian, how about actually talking to someone from, say, either Legato or Debian about it instead of a third-party admin having a problem using it?
I'm really surprised actually; Charlie Babcock has some of the best print credentials of anyone writing for ZDNet, yet this article read like something written by a fresh journalism school graduate writing TV news in Missoula. <CONSPIRACY>It almost makes me wonder if there is an agenda of some sort in play here.</CONSPIRACY>
This is my opinion and my opinion only. Incidentally, IANAL.
MOO;IANAL.
There used to be a picture linked here.
Debian, Redhat, LinuxPPC. I'd call that three "forks". Yes, they share code, and aggressively, but the fact is that, in any *practical* sense, Linux has been "forked" since Linux68k first showed up. That's what, close to a decade now?
My blog: http://www.seebs.net/log/ --- My iPhone/iPad app: http://www.seebs.net/seebsfrac/
Two out of three Dining Philosophers think the world needs more forks, and thus await The Great and Destined Forking of Linux with bated breath. And growling tummies.
--
Sheesh, evil *and* a jerk. -- Jade
Standard responses to follow:
a) It hasn't happened yet
b) If it does... will it matter?
c) Windows already has 9+ versions
Did I miss anything? We should have a FAQ for this!
Chris Dibona's job title is "Linux Community Evangelist".
-B
In my opinion, it pretty much already *has* forked. But the question should read, "Does it really matter?" As long as the vast majority of software comes with source code, and the OS has a standard compiler, then fragmentation is not a problem. Unixes have been fragmented for years. But since you simply downloaded the source code, did a "make config; install", it didn't really matter. And for the same reason, it won't be a problem with Linux as long as two things happen. Source code to major components must always be available to be rebuilt on "non-standard" systems, and software must be written with portability in mind.
---
"Go Metallica. Die RIAA." -- Linus Torvalds
Because Linux has a bunch of different uses, there is a forking tendancy. I don't want palmtop Linux to be the same as mainframe Linux. If you're good at one, you can't be good at the other.
The important thing is to have a common ABI. If the same software runs on two boxes, I don't care whether they are two different flavors of Linux, or even if they are both Linux!
The value of an OS is proportional to the size of its user base, since that defines the developer base. The size of the user base depends on ABI level compatibility--the ability to use the same program on different machines. Thus, the tendancy here is for the ABI to stay similar, because changing your own ABI cuts you off from the pack. Tendancies towards forking the ABI will come from two ends: those wishing to sabotage Linux (that is, other OS vendors), and those who extend the ABI to showcase some feature that only they have. The former can be recognized and fought; the latter will eventually extend the ABI, but allow for non-compliant software in the short term.
If ABI forking becomes a problem, the solution may be to have a centralized ABI authority. Such a group would define several ABIs (such as Linux Embedded, Linux Desktop, Linux Server), and define what services are available and how to access them. The group would then define a reference platform that provided those ABI services and only those ABI services. Any app that could run on a given reference platform gains a "logo" approval, and any platform that supports the reference ABI gets the same (whether it's Linux or not).
--The basis of all love is respect
But that doesn't mean that the threat of forks is significant. Many small forks can, and will, happen. The danger arises only when more than one of them commands a large audience.
Will the Linux community follow two different forks en mass? I am not so sure.
Geeky modern art T-shirts
The only thing that I can think of is that the commercialization of Linux might cause some compatibility problems in the short term. But, I mean how serious are they? It's like if you install Redhat you get a Gnome desktop as default, most of the others use KDE, but you can always switch to whatever Windowmanager you like by editing .xinitrc (I like Window Maker, myself) and installing the proper stuff off of the disk if it is missing.
I mean if you install the proper libraries into Debian(in the required directories), couldn't you get the program in the article working? Of course, if the program was closed source you might have a problem figuring out what was wrong, but then that's more of a closed source problem than a Linux problem, I think.
Closed source Linux vendors may, in fact, just use distribution X as a standard, and if you are using distribution Y (which doesn't install certain libraries by default) it might cause a headache if the people at the closed source vendors tech-support hotline don't know what they are doing and can't tell you that you need ncurses or whatever. So, this means good training to the closed source vendors tech support, not that there is a problem with Linux. If I had a problem like this, I wouldn't be complaining to Linus but to the person who sold me the incompatible closed source software.
All the creatures will die, And all the things will be broken. That's the law of samurai. (Jubai, 1605)
To see the lack of forking in OSS all you need do is look at the older generation of projects. The great Emacs/XEmacs revolt led to two completely incompatible.... oh okay then, two occasionally incompatible versions of the editor. While the code was different beheath the surface they can still use each other .el files on most occasions. That split is probably the most famous in OSS, most other projects result in the killer App destroying the competition rather than two competing forks.
fetchmail and sendmail, both had competitors, and especially with fetchmail there were chances for a fork when the philosophy behind it changed. Thanks to the massive climate of "don't branch on pain of death" with pervades OSS it didn't happend (read the Cathedral and the Bazaar for more on that).
Compare this to the world of closed source Unixes and the fatanstically non-interoperable plethora of mail tools that have come out for Windows over the year.
Splits happen when there is a clear difference of opinion that is supported by a significant proportion of the development community. This differs with the "upgrade or die" philosophy that keeps many vendors afloat.
OSS can result in forking but doesn't tend to result in popular forks. And even when it does it doesn't have to lead to incompatiability between the branches.
An Eye for an Eye will make the whole world blind - Gandhi
X-ING A PARAGRAB
Didn't read it, so i don't know how relevant it is to this particular thing... but it's not offtopic so please don't mod me down i'm a karma whore... heh.
--
linuxisgood:~$ man woman
Restating the obvious since nineteen aught five.
Nobody is yet making money off Linux. All above companies are losing millions on Linux products.
A deep unwavering belief is a sure sign you're missing something...
From now on, I think we should simply assume that, whenever any Slashdot article's headline ends with a ?, it should be assumed that "No, of course not!" is implied after it.
Examples:
The End of Unix? (No, of course not!)
Linux Approaching A Fork In The Road? (No, of course not!)
Do Geeks Have A Political Voice? (No, of course not!)
Hmm, maybe that wasn't such a good idea...
--
Win dain a lotica, en vai tu ri silota
2000-03-20
15:35 GMT
Dr Evil Submarine Base
Somewhere in the Marriana Trench.
Dr Evil:
"... We will subvert the Open Source movement, fork Linux, go IPO with Dr EviLinux(tm) and make millions of dollars!! BWA HA HA HA HA HA HA HA HA HA!!"
[Stunned and embarassed silence around the table]
Number Two:
"[Clearing throat] Er... Dr Evil, sir... You cannot really fork Linux... It's controlled by that finnish guy Linus Torvalds... And the largest Linux companies in the world are actually more sought-after on Wall Street than (gasp!) our own Evil Corporation(tm)!"
Scott Evil:
"Yeah, you idiot! And any moron knows that Linux is protected by the GPL anyway! If you do something good with Linux, you have to release the code, and all your competitors can then incorporate it in their own distros! You are just an imbecile!"
Mini Me:
[Jumps on table and slaps Scott]
Scott Evil:
"OUCH! [incensed]I am going to kill you, you little piece of sh..."
Dr Evil:
"Zip it Scott! Mini-Me, I said no slapping while Daddy is working! Number Two! What is that [quote with fingers]Evil Corporation(tm)[quote with fingers]?"
Number Two:
"Well, you know, the software company that has some problems with the DoJ roght now? [whispering] The one that is based in Seattle? 65K+ bugs in their last released product?"
Dr Evil:
"Riiiiiiiiiight... That Evil Corporation(tm)"
[Thinks for a little while]
Dr Evil:
"Very well, then, I'll just have to come up with another plan to conquer the world and make plenty of money. Something which involves large nuclear weapons or some such. This time-traveling is so embarassing -- you just cannot keep up to date with all these latest technologies. Come Mini-Me! Daddy is going to teach you how to torture small animals while he thinks about all this"
[Dr Evil & Mini-Me exit stage left]
The right to offend is far more important than the right not to be offended. (Rowan Atkinson)
OK so if today all the linux developers go NUTS and decide to support some other OS doesn't the fact that the source code of their code is available provide the best garuntee for a business ?
I mean any software company could tank, at any time. But if an OSS software project loses its developers all you need is time before some other developers take off where they left off (IF in fact the software does have any market potential or penetration). In every example given in the article the potential to enter the market of developing linux software or maintaining its software is freaking huge!!!
Intel is now a dominant hardware manufacturer, and Linux runs easily on it. Even the old guard hardware manufacturers like IBM and Hewlett-Packard sell Intel-based product lines.
Now i can see that IBM sells its NetFinity servers but it is still pushing the AS/400 and RS/6000 series as the REAL servers! And I just dont see Intel based machines being able to take on the BIG tasks that a lot of heavy iron from IBM does. And I just dont see an Intel box working in the place of an Sun EU10K system (*droll*). Not just cause of the hardware, but also cause Linux is still not as developed as the other Unix's.......
"Cream rises to the top. You don't find the cream unless the milk is out there."
Penguin says :
I'm the cream of the crop... i rise to the top.
Non-Deterministic Finite Automata
Let's look at situations where forking has managed to get past a missing/closed-minded maintainer, or just been used to introduce revolutionary ideas:
;-)
NetBSD designed for portability.
OpenBSD designed for security.
FreeBSD designed for performance.
Emacs and XEmacs, both the same interface, but each with a separate design philosophy for the various developers.
Forking of Windows into the "professional" (NT) and "home user" (3.1/9x/WM).
Exactly which of these have caused great turbulence for the people involved with those programs?
None. Although the NetBSD people are a bit touchy about Theo, still
Once we are allowed to compete on our merits, opensource darwinism allows the best to be created. We've seen what monocultures do before, and will likely do in the future. Heck, nature has forked people in the past. Why do you think we have different skin/hair colours and other differences?
---
--
Internet Explorer (n): Another bug -- that is, a feature that can't be turned off -- in Windows.
The distributions have been forked for quite some time. This may be an issue for end users who don't ever want to have to compile a kernel or install upgrades to libraries. But the underlying source has not forked. In fact, with the remerging of egcs and gcc, one of the biggest potential forks has disappeared.
For developers, as long as the kernel, compiler and libraries don't fork, does it matter if the distributions release at different times with different versions? Does it matter that they focus on different audiences? Not really. I'd probably have trouble finding a program compiled for one Linux distro that I can't make run on another one. I might have to get different versions of libraries and recompile the kernel, but would it be impossible to make it run? Probably not.
So, is there anything wrong with Corel, Redhat, Mandrake, Caldera and others fighting it out for desktop market share? If they are perceived as incompatible, it may slow acceptances of Linux among end users. And it will hurt their individual market shares. But to developers, it isn't going to slow the progress of development on the tools that brought us to free software in the first place.
The net will not be what we demand, but what we make it. Build it well.
Actually, it was DOS that they were afraid it wouldn't run.
Since the hardware manufacturers can change Linux, they no longer have that worry. The temptation to build proprietary hardware (for whatever purpose) re-enters the picture.
(knock knock knock) HELLO! McFly! Have you heard of Wintel? Does the term "duopoly" mean anything to you? The hardware may be "open" in a sense, but Intel holds patents over a very large part of it.
Besides, proprietary hardware has been found to be a Bad Idea. Like Communism, it just doesn't work. Linux's popularity will have no effect on that.
Then you have the source for Linux, but it's only X-box Linux that won't run on anything else. So you've got the source, what good is it?
It's a lot of good. It gets merged into the kernel and the good stuff is kept while the crap is thrown out. Then it's just plain ol' Linux that now supports X-box. Of course, it's improved faster than the forked version, and the people who were doing the forked version realize that the kernel proper with X-box support is moving faster and just start sending patches to Alan. He approves many of the patches and the X-box people are made maintainers of the X-box part of the code. They also submit patches to the regular kernel and everybody benefits.
That's what good it does.
I see no reason why Linux should 'fork' at all into incompatible OS's...
The nature of the GPL is such that any any alterations to code (including the kernel) have to be released, making it possible to keep any strands that shoot off included in Linus/Alan's 'official' kernel, or not if they were pants.
Yes, Open Source software allows for people to take code where they want to, but in the case of Linux, where there are obvious focal points (Linus Alan etc) for what is or isn't the 'official' kernel, it seems unlikely to me that it would happen.
Even if it did, I don't think it would really matter, porting software between platforms that would inevitably be very similar would be no great hassle on the whole.
--
Listening for the sound of the coming rain...
Forks are only important in closed source software.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
Duh, he was obviously talking about the letter X.
Most of the time you can use Z (xenophobia == zenophobia) and for other cases just spell it out (ecksray and eckscrement).
With only 25 letters in the english alphabet there would be more time left to teach our children useful stuff like how to troll slashdot.
Slightly off topic but I found it funny that they refered to Chris Dibona as an evangelist.
From Dictionary.com:
evangelist \E*van"gel*ist\, n. [F. ['e]vang['e]liste, L. evangelista, fr. Gr. ?.]
A bringer of the glad tidings of Church and his doctrines. Specially:
(a) A missionary preacher sent forth to prepare the way for a resident pastor; an itinerant missionary preacher.
(b) A writer of one of the four Gospels (With the definite article); as, the four evangelists, Matthew, Mark, Luke, and John.
(c) A traveling preacher whose efforts are chiefly directed to arouse to immediate repentance. The Apostles, so far as they evangelized, might claim the tittle though there were many evangelists who were not Apistles. --Plumptre.
I can see Chris now, all gussied up in his ceremonial Tux outfit, watching over his Linux flock, lined up in the pews of the Church of Linux and Open Source Advocates.
More race stuff in one place,
than any one place on the net.
Why is it that articles of this sort talk about the perspective of investors, corporations and IT managers (Microsoft collaborators mostly)? Linux has managed very well indeed without these assorted "people" and will continue to do so, fork or no fork. Berst and Wall Street have so far had a perfect score of "zero" when it comes to Linux, and I suspect this is more of the same.
Fight Spammers!
I mean come one... NT 4.0 why is it that some stuf wil work with service pack X,but not with Service Pack X+1... or something equally annoying... It seems that MS cant even keep a tree clean and that one product for them can branch....
Since you mention NT...   ver. 4.0 is what?   Basically an updated 3.51 with a Windows 95 GUI.   But are they really compatible?   No.   And worse, why create yet another 32-bit operating system for Win95 ver. b/c and Win98 that is not recognizable by NTFS, MS's own earlier 32-bit operating system??   Fork.   But that's acceptable I guess and gee... maybe not really a fork and....
As long as Linus maintains the kernel, then a "fork" will not happen.   You will have differentiation between distributions based on what apps ship with them, how the distro gets installed, and what stuff shows up as icons on your X window manager, but "forking" (in the negative connotation that this article portrays) ain't gonna happen.
-- Win2k: "It's not so much that it's only 65,000 bugs, it's just that they stopped at 65,535 to prevent an overflow."
My theory is that it's to divert media (and pro-Linux posters') attention away from the story on WebTV's Melissa-like bug:
Hee hee... YES!   I did hear about that over the weekend but seems that Sony took the hit for it in the press rather than the makers of the OS that powers the backend...
-- Win2k: "It's not so much that it's only 65,000 bugs, it's just that they stopped at 65,535 to prevent an overflow."
Interesting that this morning the FUD is going fast and furious - first "Linux is insecure" and then "Linux is forking".
;-)
I believe that several Linux-based IPOs (like Caldera's) are supposed to debut this week, if not today?   Think there's any relationship between that and the FUD-cranking???
-- Win2k: "It's not so much that it's only 65,000 bugs, it's just that they stopped at 65,535 to prevent an overflow."
The way I see it, all of the above have already forked. One of the biggest anti-linux criticism that Linux lacks application support. Mix this arguement with the idea of forking, and there's not much else to say: if a distribution strays too far from the Linux middle of the road, they lose all application support... unless the good will of 10,000 open source developers will lead them to port their products to an incompatible distribution.
-CSErwin
Linux, the kernel, will not fork in the foreseeable future, since Linus is maintaining his
status as mantainer of the kernel source. And, anyone who doesn't have THE kernel maintained by Linus isn't really Linux.
However, the different distributions, such as RedHat, Caldera, Slackware, have already forked, and have been forked for at least some three years, if I remember correctly. The distributions will have to get in-line with some standard, such as the Linux Standard Base (LSB), so that compatibility among distributions is guaranteed, but that is another entire matter.
This is not to say that compatibility among distributions of packages or applications should not be important. It IS very important. However, Linux, the kernel, is not forking anytime soon. The different distributions, usually based on some
GNU/Linux combination, have already forked, and a long time ago at that, so what's the big deal?