Slashdot Mirror
Linux 'Awfully Cathedral-Like' - Java's a Bazaar
jg21 writes "LinuxWorld draws attention to a curious use of ESR's The Cathedral and the Bazaar by the Sun Microsystems exec who currently talks about Linux more than he does even about Java. Apparently Sun's President and COO Jonathan Schwartz said at a press briefing last week that Java with its JCP is more like ESR's Bazaar than Linux, which he dismissed as being "awfully cathedral-like" since Linus is the final arbiter (or Great Dictator), and not a committee." But be sure you don't mis-use the word Java in this Bazaar or the Mall Police will totally get you.
It seems to me that using generalisations on what development leadership strategy is best is wrong. I mean, look at the totally different method Linux has compared to Apache compared to any other successfull project. The deciding factor for success for each of these very different strategies is in how well it fits the people involved and how well it gets the best results through. One size does definitely not fit all.
... and I can definitely affirm that Java is bizarre. B-dum-chee! Thank you! Tip your waiters!
Seriously, Schwartz's bias is clear. The Java Community Process which involves committees of experts and interested parties does indeed yield enhancements to the Java API that are nicely featured and well thought out. But getting on those committees in the first place requires surmounting quite a hurdle. And in the end, Sun itself remains every bit as much a "final arbiter" to the core in which any enhancement runs, the virtual machine.
Linux on the other hand, I can muck around in the code myself however I like. I can include other people's patches that Linus *does not* approve, or I can even change it myself (though between you and me, don't expect it to do a damn thing other than crash).
Exactly. If Linus' version (for whatever reason) became so outdated and unnecessary anyone else could fork it off (from any point) and maintain it however they wanted.
If someone thinks that a panel of people is so much better at making descisions for the future of "Linux" so be it. Enjoy maintaining the kernel. Honestly, Linux has been doing amazingly well with Linus at the wheel and I really can't see it changing anytime soon.
Yeah, there's tiffs here and there about what gets put in and what doesn't but it's his fork and he can maintain it however he wants.
Someone saying something "bad" about their competition!
In other news, MS claims use of Linux violates 1m of their patents and has been known in the state of california to cause cancer.
Mod me down with all of your hatred and your journey towards the dark side will be complete!
The Linux kernel is extremely fragmented, and the only reason Linus' kernel remain relevant is because he's shown himself to be pragmatic enough about what he'll include that people find it worthwhile trying to sync with him where possible.
How it could be more Bazaar like is beyond me - various strains survive purely based on merit, and features appear or disappear based on what gets popular or what doesn't get any traction. At any point there can be a total chaos of available versions solving any number of different problems.
Linus just happens to keep being that guy that built a name by being the original author, and keep his reputation by getting his version good enough for enough people to keep the "customers" coming. If he starts screwing up, someone else will take all his "business" and he'll end up being ignored. At the same time there keeps being enough niches for tons of other versions, because not everyone has the same goals.
Contrast that to Java, where no matter what happens, Sun is the final arbiter.
The cathedral means developing inside a small circle and releasing only in great intervals. The bazaar means releasing all the time and letting lots of people submit patches. By that definition, the JCP is certainly more cathedral-like than Linux.
(Note that the cathedral/bazaar difference doesn't refer to free vs. non-free; the FSF's early free software was developed in a more of a cathedral model.)
I find when corporate CEOs openly attack the opposition it's from a position of fear and weakness more than anything. When you're attacked by your competitor, you're doing something right. Maybe I'm wrong, but I don't hear Jobs attacking Linux and it's in direct competition with OSX in the server market. What did he do instead? He embraces it. I'd love for Sun to enter the desktop market more like I think they want to, but they have to give up on the "let's replace MS" dream.
Luck favors the prepared, darling.
Except do you think one man (ie Bill G) is the only one who decides what goes in or out?
They have huge numbers of developers, testers, researchers etc to decide what should go in and how. The colour of the taskbar in Windows wasnt just randomly picked by Bill "hmm, lets make it blue", they would have had researchers and testers etc to decide what it should be. The same goes with every little detail in Windows. Its not based on what some guy think would be best, but typically by what the research shows would be the best So, there is a committee deciding what should go into Windows, its just that its pretty much Microsoft employees - but at least they seek outside consultation etc.
I.O.U One Sig.
As the sainted Lindsay Marshall pointed out to ESR
at a conference some years ago, cathedrals (which
we know a bit about in Europe) weren't built like
ESR thinks. They were built over the course of
generations, by a sequence of random people, and
if you had the money to put up (say) a side-chapel
for your recently deceased son, you could do so.
In that sense, they are precisely like Linux: a
set of guiding lights, an overall architecture,
and a framework into which anyone with time and
money can put their additions. If you go to one
of the larger, more complex cathedrals in Europe
you'll see they changed massively in plan and
intent over the some hundreds of years they took
to build.
ian
Its amazing how someone can take a great concept and paraphrase it to complete bullshit.
ESR's distinction has nothing to do with closed or open, good or evil, or any moral judgement you ascribe to it.
The Cathedral is a small group working in isolation to a common and predefined goal. "In isolation" meaning not involving collaborators outside the group during development. ESR himself says "It's fairly clear that one cannot code from the ground up in bazaar style. One can test, debug and improve in bazaar style, but it would be very hard to originate a project in bazaar mode. Linus didn't try it. I didn't either." Open source projects general start in Cathedral style.
The Bazaar has everything open for collaboration from anyone during development. Some small group chooses and manages what does and doesn't go into the final "product", but there are only loose and informal goals. The product gets pushing into the shape of whatever anyone and everyone want it to be for them.
You can't modify Linux to do what you want. You can take the Linux source and make a derivation that does what you want, but its not the Linux that the rest of the world uses. Its not product development. Its not the Bazaar. The Bazaar is about contributing to a product, not forking it. The Bazaar is managed, it just doesn't look that way. The source tree isn't open for just anyone to modify, only to read, and to suggest modifications.
Java and Linux present an interesting case to which to apply TCATB.
Java uses a Cathedral style -- development on a revision is performed in isolation by a small group working to specified goals, then the result is released (with source code, but maybe not under your favourite license). But the determination of Java's goals uses the Bazaar style -- everyone gets to make their suggestion and have their say. Depending on community support (either in terms of being vocal or by contributing reference code or technically beneficial suggestions) the desired features may or may not be implemented during the next Cathedral phase.
Linux on the other hand uses Bazaar development. Anyone can hack on the code and contribute changes. But near the top there are a small group who are managing what changes do or do not make it into the official kernel, and ultimately Linus makes the final choice. So assuming that Linus and the patch managers have their own predetermined goals for Linux, the patches they admit to the official kernel tree are more typical of a Cathedral model, in that they are committed by a small group working towards a common and predetermined goal. Of course the argument can be made that Linus and co. don't have specific goals. I believe the truth is somewhere in between -- the goals of the patch managers change from time to time, but are (in the short term) generally predefined.
i-name =twylite [http://public.xdi.org/=twylite], see idcommons.net
There is a point to maintaining pruity in your programming language. For example, we have an application at work that some knuckleheads in New York wrote against Microsoft Java about 2 years ago. Now the thing doesn't work with Sun Java, doesn't work with IBM Java and it doesn't even work with the new Microsoft Java. I'm forced to uninstall java on every machine in my location and install a 2 yr old M$ build.
Linux works without purity because it's not designed to be pure. It's designed to be taken apart and reoutfittied as necessary.
The whole comparison thing is Apples to Oranges.
You can also obtain and modify Java's code as you wish (see http://java.sun.com/j2se/1.5.0/download.jsp) but you can only *distribute* your modifications for the purpose of "research" (so not as part of a commercial product for example).
1 6#linux_is_an_open_source) has aspects of both the cathedral and the bazaar.
r /cathedral-bazaar/index.html) to be an essential read, but it's terminology is IMHO too obscure to be used effectively in discussions like this; I find well-known terms like "dictatorship" (Linux kernel), "meritocracy" (Mozilla.org, "Individual Expert"s on the Java JCP Committee) and "feudal" (GNOME.org) are clearer.
Java is "bazaar"-like because the JCP provides a mechanism for groups and individuals to create proposals to evolve or extend Java which are ratified by a committe (again of groups and individuals, essentially chosen in a meritocratic manner). This could be compared with Mozilla's team of super-reviewers.
Jonathan's point is that Linux (the kernel) is cathedral-like because decisions about changes to the kernel are made exclusively by Linus Torvalds.
Java has open processes for becoming a member of the change committee (see http://www.jcp.org/en/participation/membership) and for submitting proposals (see http://www.jcp.org/en/procedures/jcp2#1).
"Linux" in the broadest sense (see http://blogs.sun.com/roller/page/ColmSmyth/200411
I really find Eric Raymond's seminal CATB article (see http://www.catb.org/~esr/writings/cathedral-bazaa
http://blogs.sun.com/ColmSmyth/
Linux is developed the way it is because it works, after a fashion.
Java is developed the way it is because it works, after a fashion.
Now which method is better is impossinle to tell since Java and Linux do very different things.
If they were both operating systems, you might compare with a bunch of benchmarks, like number of computers installed with it, market share, job vacancies administrating it, whatever, and draw some conclusions. But they're not, so you can't.
This is a bit like saying my way of making ice cream is better than your way of making sports cars.
Perhaps Schwarz should put out the new open source Solaris' with his preferred bazaar-like development model and show Linus and the rest of us how it's done.
I think your dining venue analogy is interesting, but flawed. The proprietary vs free issue is not about how the meal was prepared it is what you are entitled to do with it once you have it. In a proprietary restauraunt you must do with the meal what the chef said. If they demanded that you must have peas, potato and beef in each mouthful then that's what you must have. The fact that you have the capability to eat all your peas first and then do beef and mash together is irrelevant you must consume the meal as the seller intended and God forbid if you wanted to take any excess home!! In a free restaurant you would be entitled to enjoy the plate in front of you as you see fit. Sure you can take the chef's recommendation and indeed that recommendation may be valid but it is up to you. Alternatively, in a free restaurant all the meals come with the recipe so you can take it away with you and roll your own if you want to (replacing the nutmeg in the taters with the cinnamon you prefer).
But the choice of consumption is the real distinction. Not that you get to roll your own (I only ever do that when I cannot get a package), but that once a particular meal has been delivered, the consumer has the unfettered right to consume it as they see fit, in whatever way they see fit.
"The first thing to do when you find yourself in a hole is stop digging."
This is in a nutshell, FUD.
Sun is watching it's market share of Unix spiral downwards. Sun's solution to this problem isn't to innovate but to go after the competition.
It's the classic bare assed emperor...
Yes Francis, the world has gone crazy.
when you manage to delegate more important stuff, contribution comes from greater people. status has weight.
Not exactly correct... you cannot draw the conclusion that delegating important stuff to other people necessarily means that they are greater people.
"when you manage to delegate more important stuff, contribution comes from a greater number people."
Which may have merit in making something better, but it might not either. I remember sayins such as "design by committee" and "too many cooks spoil the broth" that tend to warn against having too many people trying to do something. Basically, the statement is a tautism in that if you delegate more stuff out to more people, then more people have contribution into it. It doesn't really mean much else.
Don't forget these people job is to generate news to there product/company - even, or specially, if there is nothing new about it.
I don't have the time to read this type of non event things but sounds pretty much like nothing over nothing.
You're missing the point. Bill G's minions are like priests of some secret sect who won't let you into the secret chamber where the sacred texts lie. Linus is the best merchant in the bazaar, so people "buy" from him. Bill G is the grand inquisitor and the sacred texts are updated at his whim, albeit with the input of his secret priests.
The presence of a multitude of competitors building on the same code keeps Linus honest and forces him to get by on his merit rather than on his appointment as Archbishop of Linux.
microsoftword.mp3 - it doesn't care that they're not words...
It all comes down to community involvement. And both Linux and Java communities do a very good job at that.
Btw...
Have you ever wondered what would have happened if a more organizationally-minded person ran the kernel development?
Linus is very authoritative, and has yet to form an official public community/legal entity that develops and protects the kernel in the 10 year+ that he has been doing it.
What happens if he gets hit by a bus?
Heck what happens today when large factions of kernel developers disagree with him? ( Kernel debugger )
I am not saying Linus is doing a bad job; but couldn't the Linux kernel as an organization be a lot further than it is today?
Based on upvotes, Ageism is the only "-ism" Slashdotters care about and think isn't SJW
In short there's a big bazaar around Linus cathedral, and the cathedral is open to visitors to talk with the high priest.
Analogies don't equal equalities, they are merely somewhat analogous.
Whatever the claims about "Community Process", Sun runs Java and Scott McNealy runs Sun when it really comes down to it. I would suggest asking long term Sun folks(the folks that built that company and were there over 15 year ago) what they really think of that means of governance.
Say what you will about Java and Sun, but here's how I see things:
I'm much more productive writing Java code than C or C++ code, at least for the kinds of applications I build.
Java is well supported. Most often, how well a language is supported is just as important as how good the language itself is.
Sun has done an excellent job listening to the community and making sure Java continues to grow.
Java is perhaps the only serious competitor to Microsoft's .NET, and in my opinion, if .NET "wins", we all lose.
Suggesting that Scott McNealy has some kind of low level control over Java's growth is ridiculous.
All in all, I would say Java is an excellent technology with a bright future, and to fear it because "OMG OMG, evil dumb stupid Scott McNealy controls Java, OMG OMG, it sux0rz, it's proprietary, run for the hills!" is foolish.
Linus really is the Great Dictator. In fact, Linus has outright rejected or slowed the progress of many things (stable binary driver API, the initial rejection of kernel pre-emption as "not needed" which is now in the official kernel, etc.).
Completely disregarding Sun in the discussion, is the point still valid?
Of that pony-tailed oik Schwartz.
He is such a jack-off it makes me think that he is an industry sleeper - someone sent to destroy all credibility of Sun.
M$ has thier own monkey boy, a semi-self-styled nut who says anything he wants, and M$ quizzically apologises for him, and does that half eye roll, well he is nuttier than squirrel shit look and hopefully get away with it.
What the fuck are they trying to prove, Linux is an OS, Java is a devleopment platform, what is the point all this rhetorical FUD? Does it make sense man?
I think not. Now to compound matters the sub blurb on this book is:
Musings of open source blah blah by an accidental revolutionary.
WTF? WHO-TF more like... Also, he is a gun nut. Just what we need. ITS GNU NOT GUN you nozz.
http://www.catb.org/~esr/guns/.
Did this make sense to anybody else? Is this Sun's take of M$ OS costs more? Is it just my sugar deprived brain thinking this is all too wierd?
#hostfile 0.0.0.0 primidi.com 0.0.0.0 www.primidi.com 0.0.0.0 radio.weblogs.com
In the same way, you can put additions in Microsoft Windows, or in Sun Java. But, in order to do so, you must be a big corporation, you must pay, and it must be done according to Microsoft's or Sun's specs.
I tend to think not. "There's more than one way to do it" is a phrase I associate with the Perl language--probably the best example of an open development process applied to language development. I don't recall Larry Wall micro-managing the process, enforcing special licensing requirements on developers of Perl programming tools or sending nasty-grams to developers for using the name "Perl" in a project title.
By and large, the source for Perl and its libraries is wide open (I believe the "artistic license" allows the code to be "stolen" for proprietary use..not sure though). Despite the "lack of discipline" Perl is quite consistent between platforms. I've written some pretty fancy Perl that runs without modification on my Linux box as well as in the ActiveState perl on Microsoft Windows.
I think Python and Ruby are also quite open and neither seems to have problems with forking or undue platform incompatibilities. I don't think that it matters if it is open or proprietary, headed by a "benevolent dictator" or designed by committee in terms of design stability and compatibility. If the leaders are responsive and genuinely considerate to users and other developers needs then it will result in success.
If it was indeed true that Linus was becoming beligerent or uncooperative (I see no indicatoin of that) then the license allows disgruntled users to produce a fork that addresses those needs. If Sun and/or the 900 or so merchants in their "bazaar" become disconnected then...well you're SOL and have to invent your own alternative/derivative and it had better not have a coffee-related name (and it has happened--I do not think it was MS' fault alone that it created an incompatible Java and its own alternative in C#--Sun created the environment that allowed it to happen). Open projects are also not immune to the problem--XFree86 for example. In the end though things sort themselves out and we get Linux and Perl and Apache and Mozilla/Firefox and a whole host of successful applications (even closed ones--MS Office and Visual Studio.NET are pretty successful by many measures by and large because they cater to the user's needs).
Sorry, but that simply isn't true (in general). You personally may have been welcoming and helpful, and that is commendable. Sadly, a lot of your peers are anything but, as the number of "luser" rants in a typical FOSS IRC channel or discussion forum will testify. The attitude of a significant number of self-important 3l337 Hax0rz in this respect has been one of the biggest and stupidest things holding back the FOSS community since forever.
Mod me (-1, Flamebait) if you must, but know that in doing so, you'll only prove my point.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
RTFA. The original article was posing Linus Torvalds as the dictator of Linux-and Java as some kind of nice democracy. Also, read the link in my comment please. There is interesting stuff happening well outside the C/C++/Java/C# world.
I read the article. It was nonsense, like your post. You don't refute nonsense with nonsense.
I also read the link in your comment. You conveniently ignored one of the points in my post:
Java is well supported. Most often, how well a language is supported is just as important as how good the language itself is.
I'm sure Ruby is very "cool". It is not, however, nearly as widely supported as Java.
That's rather unfair. Of course they're going after the competition, as any smart marketing organisation always will. But accusing the people who have contributed so significantly to the state of IT today, through Java, Star/OpenOffice and of course Solaris, of not being innovative is just asinine.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
You may have been friendly and I'm thankful for that. But as some of the other posters have indicated the typical attitude has been anything but friendly ... it's been downright hostile.
... they weren't written by technical writers who presume you know nothing .... instead they are written by developers (who might be great coders) but can't write. Hence, they make a bunch pf presumptions which lead users going to one How-To after another.
I've been a software engineer for over 15 years and used lots of technologies, but when it comes to any linux forum I have personally called every dirty name in the book, and I hardly think I or anyone else deserve that.
I can't being to tell you how many times I have been told to "look it up", RTFM, Google-It, or anything similiar. The problem is that these How-To's are completely incomplete
I for one intend to do my part to help Linux Newbies, and I intend to be patient and charitable with them (unlike how I was treated).
Star was a 3rd party product which was bought by Sun. I am happy they GPLed it, allowing for OO.org, but that does not make it a Sun innovation. ferthermore, OO.org and StarOffice is not a innovation, mearly an implemntation.
Oninoshiko
Sun could turn the standard over to an independent committee. They don't want to do that. You can argue the merits (or lack thereof) of their position but that's a different conversation and isn't comparable to Linus' control of the kernel (which is arguably an implementation of the POSIX standard.)
For the billionth time, Java IS run by a standards body - the JCP. Sun has a vote on future changes to Java, just like many other companies - such as IBM. JSR's are as valid a standard as anything POSIX.
Do you think IBM (or other companies) would have got so on board with Java if the process for changing the language was not open?
"There is more worth loving than we have strength to love." - Brian Jay Stanley