I’m not sure why people keep on spew tripe about there being too many distros. Give it some thoughts folks, the issues is NOT the number of distros it’s the varying ways in doing things; moreover a lack of standards and/or not following them.
Project forking usually happens due to another avenue which isn’t being explored or lack of maintenence of a project.
You, as well as many others who voice a similar complaint, are only looking at one aspect of the problem. Sure, having many different GNU/Linux distributions means that not every program will work every time on every one without problems. Using that argument, however, one could say that it would be better if GNU, Linux, Hurd, the BSDs, SkyOS, BeOS, (etc etc the list goes on for a long, long time) didn’t exist.
” If you want to use cool features of one Linux distro over other then its as good as learning a new OS. ”
Right, but the other distro didn’t have the features. You don’t have to get the second distro. However, the choice is there for you to make. If you want more features, you are free to use one with more features. Thats why there are different OSes and different distributions. They all do things in a certain way that they feel is superior. If you don’t think their way is superior, don’t worry – your first choice is still there.
“Whats the point after all because they all share the same kernel.”
Well, I’d imagine the point would be the “cool new features” you mentioned just one sentence ago.
“Why have *TOO MANY* different flavors of linux?”
Again, same reason. Each flavor exists to fill a niche that wasn’t filled from the existing distributions.
“Forking is a big problem in Open source.”
Really? Whats the problem? We hear all the time “forking is a problem” but noone has actually pointed out a free software project that forked and caused harm to the community. Why? Because the with free software, the forks can use each other’s code to continue development and only change the parts they disagree on.
“I wanted to release a product in OSS license but did not do because i don’t want my project to get forked. All i want is knowledge sharing.”
So in other words, you want a bunch of people to do exactly what you want them to do in a project you control but you don’t want to give them the freedom to make it do what they want to do if you don’t like it. That isn’t freedom at all. That is unpaid contract work.
“Wonder if Linux kernel gets forked like this, what would happen?”
There already are many “forks” of Linux. Different versions that are customized for different things. Just like the distributions use many of the same programs, though, the different Linux “forks” use the same main development tree and only change what is needed. Nobody is harmed and there are more options to fit better with what different people need.
“Incompatible drivers…Development costs will increase and eventually free software will start costing more than windows (which according to some still costs more than windows on a longer term). ”
Free Software has been around for a long time. Not yet has this happened. Not even once. As I pointed out in the previous paragraphs, because its free software both projects can use the same code base and only make changes when they need to. For example, most of the BSDs have a Linux Compatibility Layer. Its easy to build because Linux is free software and they can just use Linux code. Likewise for any projects that fork from any other free projects. They can maintain compatibility if they need to.
Sometimes, of course, the forks are so different that they cannot maintain compatibility. This is true. However, if they are so different that they cannot be made to be compatible then the fork was obviously needed.
“This is what i like most about Windows. No pain to configure things which should work by default.”
Thats correct, there is no “pain” to configure things. You get things The Microsoft Way. As has been proven in the past, The Microsoft Way, as much as they may want it to be, is not the way for everyone. This is more true in business environments than for desktop use, but it applies in many different scenarios. To me, it isn’t a pain to configure things. As a matter of fact, with most modern distributions little to no configuration is required – but the option is there. When I installed Debian (supposedly one of the harder distributions to configure) with the Sarge net-installer it autodetected all my hardware and configured everything correctly leaving me nothing to configure at all. Video, sound, codecs, etc – everything just worked.
So, when it comes down to it, what you have is another choice. Do you want The Microsoft Way, or do you want something you can customize completely and have it your way? The choice is yours. If you want to use Windows I have no problem with that. Just remember that other people prefer to have the option and it is unethical to take that option away from them.
LOTS of distros are not a problem, and the slightly different ways they each are configured are not a problem either, just pick one that works best for you and stick with it, it dont hurt to learn about the “under the hood” stuff with your distro of choice or with other distros in an extra disk partition used for testing purposes…
if you want user friendly go with Fedora or Mandrake or Novell/SuSE…
if you want more control go with Debian or Slackware…
and if you are feeling really brave try Gentoo or any of the other source based distros <eg> Linux From Scratch offers the untimate in control by letting you “roll your own” distro…
Most linux configuration is very similar among distributions, with small differences. This is fine, it actually does make administration harder, but it makes automated take overs harder as well. Generally speaking, if you stick with a distribution like RedHat you can buy manuals to do most anything it can do; and it stays the same (Note: Fedora is not RHEL).
Monocultures are stupid. Competition is good. And forking is quite rare in linux. It took years of bad leadership in xfree for a real fork, xorg. And now most have simply switched to xorg. The kernel hasn’t forked, at least nothing successful. GNU utilities don’t fork. GCC hasn’t forked. Gnome has one semi-fork, but it’s really just distribution with extras (dropline). KDE has not forked.
Choice and forking are two different things. And a lot of those linux distributions exist to better serve other locales than English-US.
So you run off and live in your monoculture. Enjoy many tutorials and your spyware. Enjoy it.
That was the most incoherent rant I’ve seen in a while. I couldn’t force myself to get through all of it so I’ll just stick to this piece:
…that having too many distros in Linux sucks. The steps to configure this stuff is so different on every major linux distro. This plainly should be unacceptable.
While I think people making up their own new distribution just for the hell of it is absolutely stupid, I don’t lose any sleep over it. PAM, nsswitch, and winbind work the same on every modern distribution. If your complaint is that the GUI tools aren’t all the same, learn how to really configure things and then it will be the same.
The lack of an adequate ‘LSB’ framework which specifies what a baseline Linux distro should support, how it should support it and a complete reference implementation is a major problem with regard to interoperability etc. in the Linux world.
You can whine all you like about ‘forking’, ‘too many distros’ etc. but its not going to change in the next 5 years at least, because nobody capable of actually doing this work is at all interested in seeing it happen, as far as I can tell.
So you might as well just use an OS that actually fits your needs e.g. Windows or MacOS X. Otherwise, pick a distro, stick with it, and pester them to support these features in a way you, as the customer can use.
Forks usually happen because of bad project leadership, lack of maintenance, to explore different directions than the original project’s goal, etc. It’s a myth that fork magically occur just because the source is available.
Forking is not necessarily bad. People fail to realize this.
– How long do forks live? Do forks that live for a week matter?
– What about forks that have practically no users?
– What if the fork becomes more sucessful than the original?
– Etc.
Geez, USE your brain and think about it.
Maybe I should write an entire website about this. I’m sick and tired of the “OMFG a fork!! bad bad bad!!! AARGH”-attitude.
You can buy a scoop of vanilla and then turn right around and buy a scoop of chocolate… and they TASTE DIFFERENT!
This plainly should be unacceptable.
Whats the point after all because they all share the same main ingredients.
Flavoring is a big problem in ice cream parlors. I was going to fix dessert, but I did not do so because I did not want my project to get chocolate syrup poured all over it.
All I want is recipe sharing. Wonder if milk gets flavored like this, what will happen? Incompatible recipes, cooking costs will increase and eventually ice cream will start costing more than yogurt!!
This is what I like most about yogurt. No pain to flavor things which should be flavored by default.
For instance I think it would do more for the cause if people would offer a subset of packages or tools for an existing specific distro rather than reinvent the wheel by creating a new distro to offer new tools.
I think the reason this does not happen is because egos are at stake, people don’t want to say “I build tools for Red Hat” they want to say “I have my own Linux distro, with some better tools than Red Hat”
I definitely don’t believe all forks are bad, or that all forks are good. It is a gray area.
There are many quirks in some distros I find very annoying, I am glad I am not stuck having deal with them and I have forks to thank for it.
Actually i didn’t wanna say that forking is always bad but unless it is forked to target a different segment, it is bad. One big problem with forking is when a product is forked due to bad leadership or developer ego. This causes same product to get developed by two different teams. Sometime they adopt features from each other, sometime they reinvent wheel. At the end of the day we could have a better product if they were not forked because ofcourse developer energy is not synchronized.
This problem becomes even more worst when these forks provide little value over each other but are incompatible with each other in certain ways. The worst is when instead of a single product the whole distribution gets forked. Now i can’t say i am familiar with Linux, i can only say i am familiar with distro X of linux. If somehow one distro doesn’t fulfill my needs, i will have to go and learn anotehr distro (and frankly its a pain, i want to use my time on more productive things than where the hell a particular distro stores its config files).
Sometime i feel Linux distributions want to keep their distro incompatible with each other to lock their users in, something which Microsoft does too to keep Linux away from the network where windows is.
Think about it, lets say all distributions use LSB. We now have a base system which is standard and on top of it, a distribution can provide stuff which has value in it. This way when i move to a new distro, even if i am not familiar with its GUI tools and other stuff, i can still use the distro based on my LSB knowledge. Similarly this would make sure that once i write an application for *Linux* it runs on *Linux* and not x distro of Linux only.
Finally you guys know can read emacs vs xemacs discussion and you can put yourself in a corporate customer shoes and ask him if he would love LSB or not. I can then run free Linux distro on workstations and RHEL on servers and use same automated scripts to manage the whole system. Or i can even install different Linux distro on workstations based on particular need but i won’t have to worry that i have now 10 different OS in my corp which i have to manage.
Believe it or not, Windows does make people feel more comfortable with this because they know once they learn how to fix one workstation they know how to fix all of them.
I can then run free Linux distro on workstations and RHEL on servers and use same automated scripts to manage the whole system. Or i can even install different Linux distro on workstations based on particular need but i won’t have to worry that i have now 10 different OS in my corp which i have to manage.
Or use the same distribution for everything. A good free, all around distribution that comes to mind is Debian. Maybe Ubuntu for client desktops and Debian stable for the servers.
Even if you do use two different distributions for your clients and servers, do you really perform the same maintenance tasks on the clients as you do the servers? Doubt it.
A much better reply, Wolf. This time I actually agree with you. I think your first post was just worded poorly.
In any case, it would be excellent if every distribution (where applicable, obviously some specialized ones won’t be able to) stuck to the LSB. It would be great if people didn’t fork distributions when all they needed to do was fork a program. I agree there are too many distributions out there. I just don’t agree that there shouldn’t be segmentation. Segmentation, competition, it is good.
If I am not mistaken, I believe most mainstream distros follow both FHS and LSB pretty closely. The promlem is that those specifications don’t cover every aspect of Linux software.
I believe there should maybe be a Linux workstation specification that would be compatible with FHS and LFS but cover some of the things they don’t. If distros don’t follow the spec exactly, it is not a problem. It would be more of a guideline than a rule.
is the same on all distro, the same config files and so on. you just need to learn how to use a either a good ascii text editor (lot of them out there) or a browser aimed and localhost:901 and use the swat web interface
gui tools should not be the worry of the app makers unless they a building a gui app. if its a server your trying to configure then it will be much more flexible if all you need is to ssh to it and use a text editor as this can be done over any connection, includeing a slow dialup. let the distro or desktop people worry about makeing guis for the rest of the stuff, or take a look at webmin (in many ways, swats big bro. it enbables you to configure a whole linux distro useing a browser)
i compiled MPlayer from source with the Essential codecs package and i can play ANYTHING (both audio & video) that Windoze Media Player can play and ANYTHING (both audio & video) quicktime can play too, (including mozilla/netscape browser plugins)…
Grip or Sound-juicer can rip mp3 or ogg files from music CDroms…
and gphoto2 pulls photos from my digital camera and with sane/Xsane my flatbed scanner works great too…
so your idea of Linux not haveing a good media player is WRONG!!!
There is plaenty of good and free support for Linux & every major Linux distro keeps their products updated with security patches and any bug fixes that are needed, which is better than any of those expensive phone calls to M$FT or Windoze Update that requires InfectedX (ActiveX)…
as far as Redhat or Novell/SUSE goes they are tergeting business that need professional support…
but if you want free there are some distros that offer FREE like Debian & Ubuntu are two for example…
i just wanted to debunk your ignorant rant for the public good, not that you would follow up on any of this since you are obviously a M$FT fanboy looking to do some trolling…
Different flavors of icecream would be more applicable to an analogy of something like desktop background color or font size, nothing close to the analogy of the forking problem. Linus himself is on record saying that forking is a problem. Someone hit the nail on the head with the word “ego.” Ego is a huge problem in the linux community. The point of OSS is to make a product that does something easily and effeiciently, which many OSS developers do, but there are too many hack developers out there doing their own thing for the sake of the mentioned “ego.” Hail to the well planned OSS projects…check your ego at the door to the rest
Yes. People weren’t happy with the progress being made with gcc so it was forked into egcs. Cygnus was a driving force behind the fork as well as the development of egcs.
Alright man, i am sorry for the harsh words. I just get ticked off when people compare Software and specially an OS or distribution to cars or other things. Like hey why we have so many models of cars?
The point is each car works almost the same way for us. In computers things becomes more complicated and time consuming and everyone has gone through the pain of spending hours on configuring something which they should be able to do in a minute.
As for different games, the thing is i can live by playing one game because i don’t have to interoperate between them. In software it just isn’t the same ball game.
I don’t mind having different flavors of something but all i want is that if they are all serving the same purpose, please don’t make them incompatible. It may be my gut feeling, but i feel that this too many distro confusion is actually hampering the growth and adoption of Linux at desktop.
How odd, I thought the point of OSS was freedom to do as you like with the code. Quick, someone tell RMS, Theo, ESR, etc they are barking up the wrong tree!
OSS has nothing to do with freedom. Read the GPL and tell me you can do anything you want with the code. You can’t. BSD license is about freedom.
Of course it is. If Linus said it, it must be the gospel truth!
One of the posters said it well when he advised OSS folks to check their egos at the door. Apparently, you guys can’t even accept a simple fact: forking is generally a bad thing. Fine. Whatever. Keep crowing about how forking promotes more flavors of ice cream — but continue to ignore standardization.
Read the GPL and tell me you can do anything you want with the code. You can’t. BSD license is about freedom.
This argument has been rehashed too many times. Both licenses promote freedom. The BSD license gives developers the freedom to do what they want, including the freedom to make their code not free (taking some freedom away from others in the process). The GPL removes that freedom, but in doing so ensures the freedom of others is not lost when a project forks or GPL code is used in another project.
Both licenses trade a certain freedom for another. Both are necessary, as they are mutually exclusive. Both licenses are just as good, depending on what your purpose is for licensing your code under said license. Neither is “more free” – they both trade away one type for another. Saying things like “The BSD license is more free” only starts flame wars. Since you are, apparently, smarter than me you should know this. Unless, of course, your point was to start a flame war?
Keep crowing about how forking promotes more flavors of ice cream — but continue to ignore standardization.
That is a false dichotomy. I posted why I believe forking promotes better software. I also think it is bad when the forks ignore standards. You would be wise to not post such inflammatory comments.
… would you dare to give one example where there are major differences among the distros (in terms of where configuration files are stored and so on)? In my experience it’s essentially package management and GUI frontends that are really different.
Both licenses trade a certain freedom for another. Both are necessary, as they are mutually exclusive. Both licenses are just as good, depending on what your purpose is for licensing your code under said license. Neither is “more free” – they both trade away one type for another.
Nonsense. The BSD license doesn’t dictate how you can use the software — or how you can use derived products. It’s simply more free. You guys resist this criticism because it conflicts with the primary motivation of the GPL — to keep people shackled to the “community”.
Saying things like “The BSD license is more free” only starts flame wars. Since you are, apparently, smarter than me you should know this. Unless, of course, your point was to start a flame war?
My motivation is irrelevant. If you consider the post to be flame bait, exercise some self-restraint and ignore it. But don’t pretend that merely disagreeing with you constitutes flame bait.
That is a false dichotomy. I posted why I believe forking promotes better software. I also think it is bad when the forks ignore standards.
Your reasoning is flawed. Forking doesn’t promote better software.
1. It creates unnecessary and multiplicative effort.
2. It makes it more difficult to enforce standards across multiple codebases.
3. Changes in one fork can inadvertently (or intentionally) fail to make it into other branches (there are many examples of this in practice) which leads to propagation of defects.
4. It greatly complicates coordination.
You would be wise to not post such inflammatory comments.
Nonsense. The BSD license doesn’t dictate how you can use the software — or how you can use derived products. It’s simply more free. You guys resist this criticism because it conflicts with the primary motivation of the GPL — to keep people shackled to the “community”.
The BSD license does, on the other hand, take away a freedom that the GPL gives me: The right to always be able to see any modifications made to my code. The BSD license doesn’t give me that freedom. By this point alone you are wrong in saying that the BSD doesn’t give up any freedoms.
Your reasoning is flawed. Forking doesn’t promote better software.
1. It only creates unnecessary effort in circumstances when a fork is not warranted. If you had been following the conversation you would realize that I, too, am against needless forks.
2. That depends on how different they are. If their purpose is to fill a niche then they don’t really care about some things other standards might care about. What about NetBSD which runs on so many platforms? Having to follow a standard base would undoubtedly leave it unable to do so. As I said, forking when it is needed to fill a niche is good for the software. Most forks that don’t change much simply use the forked project’s updated code each time the project updates and modify their own code to fit it.
3. Yes, some projects have poor organizational skills. That is their fault, not the fault of the fork. There are proprietary projects all over the world that do the same thing. They aren’t forks, but they do the same thing, so you really ought to count their time as wasted much more than the “wasted” time of the projects who choose not to communicate.
4. Yes, it does. Some of us are capable of that. Most of the leaders of more popular free software projects are. Thats why the projects become popular. Not everyone is, but in the proprietary world noone is allowed to share without a transfer of money. Some sharing is better than no sharing.
I would be wise to ignore you.
And yet you didn’t. That says quite a bit about you, especially in combination with the petty name misspelling. Do you not see how immature that makes you seem? It makes you lose whatever respect you had gained from the few valid points you made. Remember that despite how commonly we see immature posts, the large majority (those who usually, intelligently, choose not to post so they don’t have to deal with the children) of people involved in the free software community are adults with jobs, wives/husbands, children, and lives. These people will take one look at your veiled hostility and see that you are clearly biased.
Or, keep posting just as you are. I don’t mind replying. Its break at university so I have lots of time to do whatever I want. It isn’t really difficult to respond to you when you do things like cast generalizations of “you guys all do this” and trying to say that without question the BSD license provides every single freedom the GPL does and more when that clearly isn’t true (in that regard, by the way, I think you ought to step back and think about why you are defending the BSD license. It, as other licenses, is around because it fit a niche. The GPL is around because it fit a niche the BSD license didn’t – making sure that the code is always free. It doesn’t make the BSD license any less viable to those who wish to use it, and I certainly never implied that it was a bad license).
In any case, my friend, enjoy your holidays. I look forward to less hostility in your next post.
“At the end of the day we could have a better product if they were not forked because ofcourse developer energy is not synchronized.”
This is assuming the developers of the fork would be willing to work on the original project at all! If the fork happened because developers are not happy with the leadership, and you prevent people from forking, then developers will just leave. So in the end the original project will weaken.
One of the posters said it well when he advised OSS folks to check their egos at the door. Apparently, you guys can’t even accept a simple fact: forking is generally a bad thing. Fine. Whatever. Keep crowing about how forking promotes more flavors of ice cream — but continue to ignore standardization.
You have mistaken an opinion for a “simple fact”. In fact forking has been a good thing for OSS (egcs, xorg, etc). Forking in itself has nothing to do with standards. Xorg and and XFree both conform to the X11 standard. That’s the thing about OSS, as long as you conform to standards, a fork should mean little in the grand scheme of compatibility. A fork just allows people to get features they want if a developer or project refuses to incorporate them into their program.
Why is it so difficult to integrate Linux/UNIX into Active Directory? Because Microsoft went and made a non standard LDAP implementation out of Active Directory. I can do single sign on perfectly well using an OpenLDAP server, which does its best to adhere to the standards. It’s secure, it’s fast and it stores all the info I could ever need. Besides, it also authenticates Linux, FreeBSD, MacOS X (with a slight nudge in its mappings) perfectly well. The only problem is Windows, which requires the added layer of complexity we call Samba 3.
You can’t blame Linux for not adhering to some bastardized proprietary version of a standard protocol.
“Nonsense. The BSD license doesn’t dictate how you can use the software — or how you can use derived products.” The license itself doesn’t, but it lets developers dictate how it is used. If someone takes code, adds to it and closes the source, no one else is able to use his improvements. This may be good for the developer who added the improvements, but it is quite bad for the people who developed the code he piggybacked on, or anyone who wants to take his refinements, and refine them further. I really haven’t decided which license I prefer, but I can see the merit in each side’s arguments.
Forks can most certainly be a good thing. egcs was forked from gcc, and later it was decided egcs was superior and it replaced (and became) gcc. Sure, ideally it would have been better had the fork not happened, had the people who worked on egcs worked directly on gcc. Sometimes it just isn’t possible though. I’m not very familiar with the circumstances, but I imagine the gcc people didn’t want to make some changes that the egcs people favoured. Egcs was able to fork it, and prove themselves right.
It doesn’t necessarily even have to do with egos. It may just as well have had to do with one side being right, and the other wrong (or less right), but no one really knew which side was which until both were able to put out products that could be compared with each other. Without forking, there would have been no way to test egcs vs gcc to see which was better, there would have been only one, and it might not have been the right one.
…I’m afraid that, in fact, Preston St. Pierre is smarter than you. In any case, he didn’t see the need to resort to immature attacks and demeaning lables (i.e. “zealots”) to make his point. A very well-made point, might I add.
As for the name-stealing troll, who posts under Rayiner Hashem’s name, the best thing to do is to ignore him until OSNews introduces its new anti-troll features (I can’t wait!)
The BSD license does, on the other hand, take away a freedom that the GPL gives me: The right to always be able to see any modifications made to my code. The BSD license doesn’t give me that freedom. By this point alone you are wrong in saying that the BSD doesn’t give up any freedoms.
I didn’t say that the “BSD doesn’t give up any freedoms”. I said that the BSD license grants “more freedom”. Forcing people to give up their code modifications isn’t “freedom”, no matter how you want to spin it. That’s like saying “Communism is about freedom because, at the end of the day, you donate your work to the betterment of the state.” Which is patently ridiculous. Simply handing somebody a pile of code and saying, “Here, go forth and play without any restrictions” (the BSD model) is the essence of freedom. GPL shackles them to “the community”. I could care less how philanthropic you consider your “community” to be. When you force me to do something, that isn’t freedom.
1. It only creates unnecessary effort in circumstances when a fork is not warranted. If you had been following the conversation you would realize that I, too, am against needless forks.
Not true. When you create a fork, you now need to maintain source code in more than one place. By definition, that requires more effort.
2. That depends on how different they are. If their purpose is to fill a niche then they don’t really care about some things other standards might care about. What about NetBSD which runs on so many platforms? Having to follow a standard base would undoubtedly leave it unable to do so. As I said, forking when it is needed to fill a niche is good for the software. Most forks that don’t change much simply use the forked project’s updated code each time the project updates and modify their own code to fit it.
You would be better off not forking, if all that you’re doing is a one-off. As for your NetBSD example, I completely disagree that proliferation of platforms requires a fork. That’s nonsense. I write software that runs on very wide variety of platforms, and only the code beneath the abstraction layer needs to be modified as the platform changes.
3. Yes, some projects have poor organizational skills. That is their fault, not the fault of the fork. There are proprietary projects all over the world that do the same thing. They aren’t forks, but they do the same thing, so you really ought to count their time as wasted much more than the “wasted” time of the projects who choose not to communicate.
You missed the point. Having to make mods in more than one codebase increases the likelihood that you will make mistakes. I’ve seen this time and time again in various projects, where somebody fails to apply a patch that was needed, and it wreaks havoc. Results in unpatched security vulnerabilities, etc.
4. Yes, it does. Some of us are capable of that. Most of the leaders of more popular free software projects are. Thats why the projects become popular. Not everyone is, but in the proprietary world noone is allowed to share without a transfer of money. Some sharing is better than no sharing.
Collaboration between human beings is difficult. Particularly for geeks with poor social skills. You’re underrating the difficulty of this aspect of forking.
These people will take one look at your veiled hostility and see that you are clearly biased.
Duh. I AM biased. We’re ALL BIASED. I see things from my own perspective. So do you. If you can’t at least admit your own bias, you’re kidding yourself.
I think you ought to step back and think about why you are defending the BSD license. It, as other licenses, is around because it fit a niche. The GPL is around because it fit a niche the BSD license didn’t – making sure that the code is always free. It doesn’t make the BSD license any less viable to those who wish to use it, and I certainly never implied that it was a bad license).
I’m defending the BSD license because I prefer freedom. Freedom from having to hand my modifications to code back to the OSS community. Freedom to do whatever I want with the code. Perhaps give it away. Perhaps commercialize it. Whatever. The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about. And then people have the audacity to tell me that I should be glad about it. Rrrrrright.
You are, once again, only looking at the freedoms of people who take your code. You are not looking at the freedoms of the people who wrote the code. The GPL gives the person who wrote the code the knowledge that their code and any modifications of their code will always be available to them. Of course, I already said that, so at this point its pretty obvious you are speaking without taking anything I’ve said into consideration.
Not true. When you create a fork, you now need to maintain source code in more than one place. By definition, that requires more effort.
You didn’t say more effort. You said unnecessary effort. Effort is needed to do new things. If the project fork does nothing, it is unnecessary effort. If the project fork does something needed, it is necessary effort.
You would be better off not forking, if all that you’re doing is a one-off.
In some cases, yes, and I said that already. Not all forks should happen.
You missed the point. Having to make mods in more than one codebase increases the likelihood that you will make mistakes.
Yes, it will. What makes you think I don’t know that? Dear boy, all you’ve done so far is give me circumstances in which a fork is not a better option than not forking. You are stalwartly refusing to admit that other scenarios exist than your own, however, no matter how much evidence I give you to the contrary. I admit readily that there are places where software shouldn’t fork. You simply refuse to admit that there are places software should fork.
Collaboration between human beings is difficult. Particularly for geeks with poor social skills. You’re underrating the difficulty of this aspect of forking.
Ok, since you know more about me than this, explain to me something: If it is difficult for two projects consisting of four people each to work together, then how is it so much easier for one team of eight people? Code version repositories take care of most of the hard work. As an example, security patches to one will work on the other unless that exact piece of code has been modified. Usually it isn’t.
Now if, for example, the projects split apart for personal reasons of the people leading them there would be no collaboration – but at the same time, there is no way that these people would ever be able to work together on the one project, so a fork, while requiring more work, still gets more done than no forking would have.
Duh. I AM biased. We’re ALL BIASED. I see things from my own perspective. So do you. If you can’t at least admit your own bias, you’re kidding yourself.
I am biased insofar as everyone is biased. The difference is, I try to not let it seep into what I say. I look up facts to base my arguments on. When someone proves a point, I accept that point. When someone says something intelligent, I listen to them. I admit that there is more than one “right” or “wrong” in most circumstances because life isn’t binary. I don’t stubbornly argue my points just for the sake of argument. You, on the other hand, have exhibited none of the above qualities.
I’m defending the BSD license because I prefer freedom. Freedom from having to hand my modifications to code back to the OSS community. Freedom to do whatever I want with the code. Perhaps give it away. Perhaps commercialize it. Whatever. The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about. And then people have the audacity to tell me that I should be glad about it.
Ah, here is where the truth really comes out. You want to take open source code and make it your own without giving back. For you, the person who takes the code, the BSD license grants more freedom. For me, who writes the code that you wish to take, the GPL grants more freedom. Are you still unwilling to admit that? If the past has taught us anything, I guess so.
This line really strikes me as insightful into the kind of person you are:
The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about.
You don’t care about the people who write the code, then, you just want to be able to steal their code and use it to your own advantage without giving anything in return. You think nothing of stealing all that code and you freely admit you don’t give a damn about the people who worked long, hard hours building that code. Yet, you still don’t see why people would want the GPL there to protect the code that you so much want to steal. The GPL is doing its job and protecting these developers from having their code stolen by you, and you are angry about it.
Well, I don’t care if you are angry about it. Its their code. The GPL and the BSD license both give one freedom in exchange for another. The people who write BSD code don’t mind if you steal their code, so steal it, but the people who write GPL code do care, which is why they used the GPL license – to protect their freedoms. You can’t see that, I guess, because of how biased you are, but you should make the attempt anyway. It would certainly bring merit to the rest of what you say instead of making you look like a raving madman who can do nothing but say “BSD is better, BSD is better, BSD is better.”
I was hoping you would see the advantages to logical argument, civil speaking, and using your intelligence but apparently I was wrong.
it that the GPL gives control(or freedom to control, depending on how you look at it…) to the developer and many freedoms to the user whilst BSD gives all the freedom to the recipient, while reserving none for the developer?
If the goal is desktop linux adoption, then I think too many distros can be a problem. I think OS adoption can have a self perpetuating effect: the more people use it -> the more people develop for it -> the more people use it. As it is Linux has about 2% of the desktop – now split that up 200 ways. What happens is you never get critical mass. Destop Linux isn’t worth developing on -> so nobody uses it -> so nobody develops for it.
Many Linux advocates don’t understand the importance of applications. Most people don’t run an OS just to run the OS. It’s all about apps. Let’s face it: several standard apps don’t run natively on Linux: Quickbooks (80% of small business market), PhotoShop, MS-Office, Visio, MS-Project, AutoCAD, etc.
It’s difficult enough to get major software manufacturers interested in developing for desktop linux without fragmenting the market 200 ways.
>>> Hope to see a strong LSB compliant Linux one day.
I do not.
From LSB :
—
Supplying an RPM format package is encouraged because it makes systems easier to manage. A future version of the LSB may require RPM, or specify a way for an installer to update a package database.
—
After using the gentoo portage system, I never want to go back to an rpm/deb base distro.
– It’s flexible
– Configurable
– Optimized
– easy to maintain
– Never face dependencies problem (debian, if you uninstall abiword it will uninstall X11, well X11 depend surely on abiword – Mandrake, xmms deinstallation make X to disapear, wow)
– it is more up to date
The problem is that it’s slower to install.
So if LSB want to force a package file format, to bad but this is madness, this standard should not to be follow.
The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about.
You mean in the name of the people who originally wrote that code…
GPL protects the freedom of users more than developers, BSD protects the freedom of developers more than users. Each deals with different aspects of freedom.
Sometimes you have to be restrictive in order to protect freedom (i.e. the U.S. Constitutions, which imposes restrictions on the majority so that the minority isn’t sujected to its tyranny).
You just want a free lunch, i.e. getting free code and not have to contribute anything back.
So if LSB want to force a package file format, to bad but this is madness, this standard should not to be follow.
You want us all to go back to compiling from source and you think that a package is madness? Please, go back to wherever you came from, you won’t help the conversation here. Advanced packages and package management systems are one of the major improvements in GNU/Linux over the past few years. The ability for me to type “apt-get update” and “apt-get upgrade” and have everything upgraded and working in an hour is astronomically useful. While I respect that some people might want to compile from source, trying to say that everyone should always compile from source and that noone should support a package management system is, well, stupid. I’m sorry, but it is. Packages are here for a reason. No need to take a step back. Packages don’t stop you from doing what you want via source. You can even compile from source and turn that into a package, then install using the central package management system for the distro.
To review: Package management > no package management. Source can be used whether there is package management or not, so you get what you want even if there is a package manager. Conclusion: Leave the package manager there and let you, a minority, figure out the source yourself. It only makes sense.
“You can’t blame Linux for not adhering to some bastardized proprietary version of a standard protocol.”
Well said!
It takes some people time to learn that A + B may not lead to C. This is especially true when it comes to proprietary v. open source computer operating systems.
Maybe I was unclear, but my point was, that a standardisation must not force any kind of packaging system. If you prefer to update your system typing “apt-get upgrade” instead of “emerge –upgrade world” is not finally the point, but a basic structure on what any package can relied on.
If LSB want to force me using rpm, I will not support that kind of standardisation and it will not bring Linux any good.
Anyway, it is quite unrelate to the main topic – using user in MS AD in Linux.
I agree that they shouldn’t force one type of package management system on you. All they should do is specify where certain things should be located and have LSB rpms, debs, etc all point to those locations. I’m sad to see they don’t do that instead (although I believe they do it as well).
egcs came out stronger, and then took the name of gcc.
And in effect, the GCC project became stronger. With GPL, it doesn’t have to be the same underlying code to be the same project. ecg (the program) effectively became gcc (the program) and therefor improved the both the gcc program and the GCC project.
GPL protects the freedom of users more than developers, BSD protects the freedom of developers more than users. Each deals with different aspects of freedom.
Rubbish. It depends entirely on how the code is used and how users define “freedom”. Most users tend to associate freedom with capability — not how the capability was provided to them. There is evidence that some MS code (ie. TCP/IP stack) evolved from BSD. The fact that the vast majority of desktop computers are running that derived code gives those users more capability than they had previously. I think that an argument can be made that private entities have leveraged that code under the BSD model more effectively to provide capability than a bunch of GPL’d projects.
Sometimes you have to be restrictive in order to protect freedom (i.e. the U.S. Constitutions, which imposes restrictions on the majority so that the minority isn’t sujected to its tyranny).
Those same words could have been uttered by Lenin and Marx, too. “Look, so we’re repressing you a little — at least you’ll be taken care of…”
You just want a free lunch, i.e. getting free code and not have to contribute anything back.
Don’t knock free lunches. Many huge leaps have taken place on the shoulders of free-lunch providers (ie. Darpa, US government).
You are stubborn and arrogant, you refuse to listen to logic, you insist that developers freedom means nothing (when it is the developers who choose the license), and you continue to flame people.
You are stubborn and arrogant, you refuse to listen to logic, you insist that developers freedom means nothing (when it is the developers who choose the license), and you continue to flame people.
Look, of the two of us, I’m the only with experience managing software engineering teams. I’m the only one who has had to deliver multiple complex projects and juggle human and technology assets. I’m the only one who has seen firsthand what happens when you fork projects and try to keep things in sync. It’s tough. Damned tough. And then, there you are, trying to tell me that it’s a good thing. Wrong. Forking has extremely narrow advantages — but most of the time, it’s a bad idea. You can call me arrogant and stubborn for asserting my real world experience but, compared to you utter lack of experience, I’m willing to suffer such criticisms.
As for BSD vs GPL, you conveniently ignore the fact that the Internet has largely grown into what it is as a result of unfettered sharing of technology by folks like DARPA, the U.S. Government, various universities, Bell Labs, researchers, and countless commercial entities. Most OSS projects are good at one thing: Copying existing commercial technology (Samba, Linux, etc). That isn’t innovation, my friend. That doesn’t promote technology advances. It merely moves pieces around the board. But there’s a reason why it has happened. The investment inherent in commercialization tends to promote rapid development.
When you hand somebody a resource and say, “Here’s a resource that you can use — no strings attached — use it to build something cool”, you’re more likely to see that actually happen in a rapid way when the recipient has a commercial stake in the outcome. As far as technology goes, greed is good. Proprietization is good. Eventually, it serves as an incubator for even greater advances. People tend to forget that Linux wouldn’t even exist if Torvalds hadn’t modeled it after commercial UNIX.
That’s why, I believe, the BSD license works far better. It allows people to utilize code for whatever purpose they find necessary without shackling them to the requirement of revealing their special sauce. And there are many examples of commercial companies (ie. Id, AT&T, IBM, etc) releasing their code into the public domain after it has paid for itself. This is a better plan for evolution than the forced mediocrity endemic in the GPL.
Say what you will. I told you, I’m not discussing it anymore. You know my point of view, and you know the reasons I have taken those views. I look at your reasons and I see holes in them, the same holes every time, but I’m tired of pointing them out time and time again. Enjoy yourself.
I’m not sure why people keep on spew tripe about there being too many distros. Give it some thoughts folks, the issues is NOT the number of distros it’s the varying ways in doing things; moreover a lack of standards and/or not following them.
Project forking usually happens due to another avenue which isn’t being explored or lack of maintenence of a project.
You, as well as many others who voice a similar complaint, are only looking at one aspect of the problem. Sure, having many different GNU/Linux distributions means that not every program will work every time on every one without problems. Using that argument, however, one could say that it would be better if GNU, Linux, Hurd, the BSDs, SkyOS, BeOS, (etc etc the list goes on for a long, long time) didn’t exist.
” If you want to use cool features of one Linux distro over other then its as good as learning a new OS. ”
Right, but the other distro didn’t have the features. You don’t have to get the second distro. However, the choice is there for you to make. If you want more features, you are free to use one with more features. Thats why there are different OSes and different distributions. They all do things in a certain way that they feel is superior. If you don’t think their way is superior, don’t worry – your first choice is still there.
“Whats the point after all because they all share the same kernel.”
Well, I’d imagine the point would be the “cool new features” you mentioned just one sentence ago.
“Why have *TOO MANY* different flavors of linux?”
Again, same reason. Each flavor exists to fill a niche that wasn’t filled from the existing distributions.
“Forking is a big problem in Open source.”
Really? Whats the problem? We hear all the time “forking is a problem” but noone has actually pointed out a free software project that forked and caused harm to the community. Why? Because the with free software, the forks can use each other’s code to continue development and only change the parts they disagree on.
“I wanted to release a product in OSS license but did not do because i don’t want my project to get forked. All i want is knowledge sharing.”
So in other words, you want a bunch of people to do exactly what you want them to do in a project you control but you don’t want to give them the freedom to make it do what they want to do if you don’t like it. That isn’t freedom at all. That is unpaid contract work.
“Wonder if Linux kernel gets forked like this, what would happen?”
There already are many “forks” of Linux. Different versions that are customized for different things. Just like the distributions use many of the same programs, though, the different Linux “forks” use the same main development tree and only change what is needed. Nobody is harmed and there are more options to fit better with what different people need.
“Incompatible drivers…Development costs will increase and eventually free software will start costing more than windows (which according to some still costs more than windows on a longer term). ”
Free Software has been around for a long time. Not yet has this happened. Not even once. As I pointed out in the previous paragraphs, because its free software both projects can use the same code base and only make changes when they need to. For example, most of the BSDs have a Linux Compatibility Layer. Its easy to build because Linux is free software and they can just use Linux code. Likewise for any projects that fork from any other free projects. They can maintain compatibility if they need to.
Sometimes, of course, the forks are so different that they cannot maintain compatibility. This is true. However, if they are so different that they cannot be made to be compatible then the fork was obviously needed.
“This is what i like most about Windows. No pain to configure things which should work by default.”
Thats correct, there is no “pain” to configure things. You get things The Microsoft Way. As has been proven in the past, The Microsoft Way, as much as they may want it to be, is not the way for everyone. This is more true in business environments than for desktop use, but it applies in many different scenarios. To me, it isn’t a pain to configure things. As a matter of fact, with most modern distributions little to no configuration is required – but the option is there. When I installed Debian (supposedly one of the harder distributions to configure) with the Sarge net-installer it autodetected all my hardware and configured everything correctly leaving me nothing to configure at all. Video, sound, codecs, etc – everything just worked.
So, when it comes down to it, what you have is another choice. Do you want The Microsoft Way, or do you want something you can customize completely and have it your way? The choice is yours. If you want to use Windows I have no problem with that. Just remember that other people prefer to have the option and it is unethical to take that option away from them.
-Preston
LOTS of distros are not a problem, and the slightly different ways they each are configured are not a problem either, just pick one that works best for you and stick with it, it dont hurt to learn about the “under the hood” stuff with your distro of choice or with other distros in an extra disk partition used for testing purposes…
if you want user friendly go with Fedora or Mandrake or Novell/SuSE…
if you want more control go with Debian or Slackware…
and if you are feeling really brave try Gentoo or any of the other source based distros <eg> Linux From Scratch offers the untimate in control by letting you “roll your own” distro…
Most linux configuration is very similar among distributions, with small differences. This is fine, it actually does make administration harder, but it makes automated take overs harder as well. Generally speaking, if you stick with a distribution like RedHat you can buy manuals to do most anything it can do; and it stays the same (Note: Fedora is not RHEL).
Monocultures are stupid. Competition is good. And forking is quite rare in linux. It took years of bad leadership in xfree for a real fork, xorg. And now most have simply switched to xorg. The kernel hasn’t forked, at least nothing successful. GNU utilities don’t fork. GCC hasn’t forked. Gnome has one semi-fork, but it’s really just distribution with extras (dropline). KDE has not forked.
Choice and forking are two different things. And a lot of those linux distributions exist to better serve other locales than English-US.
So you run off and live in your monoculture. Enjoy many tutorials and your spyware. Enjoy it.
That was the most incoherent rant I’ve seen in a while. I couldn’t force myself to get through all of it so I’ll just stick to this piece:
…that having too many distros in Linux sucks. The steps to configure this stuff is so different on every major linux distro. This plainly should be unacceptable.
While I think people making up their own new distribution just for the hell of it is absolutely stupid, I don’t lose any sleep over it. PAM, nsswitch, and winbind work the same on every modern distribution. If your complaint is that the GUI tools aren’t all the same, learn how to really configure things and then it will be the same.
The lack of an adequate ‘LSB’ framework which specifies what a baseline Linux distro should support, how it should support it and a complete reference implementation is a major problem with regard to interoperability etc. in the Linux world.
You can whine all you like about ‘forking’, ‘too many distros’ etc. but its not going to change in the next 5 years at least, because nobody capable of actually doing this work is at all interested in seeing it happen, as far as I can tell.
So you might as well just use an OS that actually fits your needs e.g. Windows or MacOS X. Otherwise, pick a distro, stick with it, and pester them to support these features in a way you, as the customer can use.
Forks usually happen because of bad project leadership, lack of maintenance, to explore different directions than the original project’s goal, etc. It’s a myth that fork magically occur just because the source is available.
Forking is not necessarily bad. People fail to realize this.
– How long do forks live? Do forks that live for a week matter?
– What about forks that have practically no users?
– What if the fork becomes more sucessful than the original?
– Etc.
Geez, USE your brain and think about it.
Maybe I should write an entire website about this. I’m sick and tired of the “OMFG a fork!! bad bad bad!!! AARGH”-attitude.
GCC once forked into gcc and egcs.
Eventualy gcc came out stronger.
Forking doesn’t have to be a bad thing.
There are too many flavors of ice cream!
You can buy a scoop of vanilla and then turn right around and buy a scoop of chocolate… and they TASTE DIFFERENT!
This plainly should be unacceptable.
Whats the point after all because they all share the same main ingredients.
Flavoring is a big problem in ice cream parlors. I was going to fix dessert, but I did not do so because I did not want my project to get chocolate syrup poured all over it.
All I want is recipe sharing. Wonder if milk gets flavored like this, what will happen? Incompatible recipes, cooking costs will increase and eventually ice cream will start costing more than yogurt!!
This is what I like most about yogurt. No pain to flavor things which should be flavored by default.
For instance I think it would do more for the cause if people would offer a subset of packages or tools for an existing specific distro rather than reinvent the wheel by creating a new distro to offer new tools.
I think the reason this does not happen is because egos are at stake, people don’t want to say “I build tools for Red Hat” they want to say “I have my own Linux distro, with some better tools than Red Hat”
I definitely don’t believe all forks are bad, or that all forks are good. It is a gray area.
There are many quirks in some distros I find very annoying, I am glad I am not stuck having deal with them and I have forks to thank for it.
Actually i didn’t wanna say that forking is always bad but unless it is forked to target a different segment, it is bad. One big problem with forking is when a product is forked due to bad leadership or developer ego. This causes same product to get developed by two different teams. Sometime they adopt features from each other, sometime they reinvent wheel. At the end of the day we could have a better product if they were not forked because ofcourse developer energy is not synchronized.
This problem becomes even more worst when these forks provide little value over each other but are incompatible with each other in certain ways. The worst is when instead of a single product the whole distribution gets forked. Now i can’t say i am familiar with Linux, i can only say i am familiar with distro X of linux. If somehow one distro doesn’t fulfill my needs, i will have to go and learn anotehr distro (and frankly its a pain, i want to use my time on more productive things than where the hell a particular distro stores its config files).
Sometime i feel Linux distributions want to keep their distro incompatible with each other to lock their users in, something which Microsoft does too to keep Linux away from the network where windows is.
Think about it, lets say all distributions use LSB. We now have a base system which is standard and on top of it, a distribution can provide stuff which has value in it. This way when i move to a new distro, even if i am not familiar with its GUI tools and other stuff, i can still use the distro based on my LSB knowledge. Similarly this would make sure that once i write an application for *Linux* it runs on *Linux* and not x distro of Linux only.
Finally you guys know can read emacs vs xemacs discussion and you can put yourself in a corporate customer shoes and ask him if he would love LSB or not. I can then run free Linux distro on workstations and RHEL on servers and use same automated scripts to manage the whole system. Or i can even install different Linux distro on workstations based on particular need but i won’t have to worry that i have now 10 different OS in my corp which i have to manage.
Believe it or not, Windows does make people feel more comfortable with this because they know once they learn how to fix one workstation they know how to fix all of them.
I can then run free Linux distro on workstations and RHEL on servers and use same automated scripts to manage the whole system. Or i can even install different Linux distro on workstations based on particular need but i won’t have to worry that i have now 10 different OS in my corp which i have to manage.
Or use the same distribution for everything. A good free, all around distribution that comes to mind is Debian. Maybe Ubuntu for client desktops and Debian stable for the servers.
Even if you do use two different distributions for your clients and servers, do you really perform the same maintenance tasks on the clients as you do the servers? Doubt it.
A much better reply, Wolf. This time I actually agree with you. I think your first post was just worded poorly.
In any case, it would be excellent if every distribution (where applicable, obviously some specialized ones won’t be able to) stuck to the LSB. It would be great if people didn’t fork distributions when all they needed to do was fork a program. I agree there are too many distributions out there. I just don’t agree that there shouldn’t be segmentation. Segmentation, competition, it is good.
If I am not mistaken, I believe most mainstream distros follow both FHS and LSB pretty closely. The promlem is that those specifications don’t cover every aspect of Linux software.
I believe there should maybe be a Linux workstation specification that would be compatible with FHS and LFS but cover some of the things they don’t. If distros don’t follow the spec exactly, it is not a problem. It would be more of a guideline than a rule.
is the same on all distro, the same config files and so on. you just need to learn how to use a either a good ascii text editor (lot of them out there) or a browser aimed and localhost:901 and use the swat web interface
gui tools should not be the worry of the app makers unless they a building a gui app. if its a server your trying to configure then it will be much more flexible if all you need is to ssh to it and use a text editor as this can be done over any connection, includeing a slow dialup. let the distro or desktop people worry about makeing guis for the rest of the stuff, or take a look at webmin
(in many ways, swats big bro. it enbables you to configure a whole linux distro useing a browser)
Actually I think Wolf had a fairly good point to begin with,
there are too many unneeded differences between different linux distributions, but dont tell him I said that!
*winks
He might just have an aneurism if he realizes someone who is so “plainly stupid” agrees with him!
Hey Anonymous;
i compiled MPlayer from source with the Essential codecs package and i can play ANYTHING (both audio & video) that Windoze Media Player can play and ANYTHING (both audio & video) quicktime can play too, (including mozilla/netscape browser plugins)…
Grip or Sound-juicer can rip mp3 or ogg files from music CDroms…
and gphoto2 pulls photos from my digital camera and with sane/Xsane my flatbed scanner works great too…
so your idea of Linux not haveing a good media player is WRONG!!!
There is plaenty of good and free support for Linux & every major Linux distro keeps their products updated with security patches and any bug fixes that are needed, which is better than any of those expensive phone calls to M$FT or Windoze Update that requires InfectedX (ActiveX)…
as far as Redhat or Novell/SUSE goes they are tergeting business that need professional support…
but if you want free there are some distros that offer FREE like Debian & Ubuntu are two for example…
i just wanted to debunk your ignorant rant for the public good, not that you would follow up on any of this since you are obviously a M$FT fanboy looking to do some trolling…
Have a nice day
Different flavors of icecream would be more applicable to an analogy of something like desktop background color or font size, nothing close to the analogy of the forking problem. Linus himself is on record saying that forking is a problem. Someone hit the nail on the head with the word “ego.” Ego is a huge problem in the linux community. The point of OSS is to make a product that does something easily and effeiciently, which many OSS developers do, but there are too many hack developers out there doing their own thing for the sake of the mentioned “ego.” Hail to the well planned OSS projects…check your ego at the door to the rest
> GCC once forked into gcc and egcs.
Yes. People weren’t happy with the progress being made with gcc so it was forked into egcs. Cygnus was a driving force behind the fork as well as the development of egcs.
> Eventualy gcc came out stronger.
Wrong.
http://slashdot.org/articles/99/04/20/1453234.shtml
> Forking doesn’t have to be a bad thing.
It’s also worth remembering that Cygnus, under Redhat, was responsible for getting gcc to move with their release of 2.96.
quote: “The point of OSS is to make a product that does something easily and effeiciently,”
______________________________________
How odd, I thought the point of OSS was freedom to do as you like with the code.
Quick, someone tell RMS, Theo, ESR, etc they are barking up the wrong tree!
_______________________________________
quote: Linus himself is on record saying that forking is a problem.
_______________________________________
Of course it is. If Linus said it, it must be the gospel truth!
Alright man, i am sorry for the harsh words. I just get ticked off when people compare Software and specially an OS or distribution to cars or other things. Like hey why we have so many models of cars?
The point is each car works almost the same way for us. In computers things becomes more complicated and time consuming and everyone has gone through the pain of spending hours on configuring something which they should be able to do in a minute.
As for different games, the thing is i can live by playing one game because i don’t have to interoperate between them. In software it just isn’t the same ball game.
I don’t mind having different flavors of something but all i want is that if they are all serving the same purpose, please don’t make them incompatible. It may be my gut feeling, but i feel that this too many distro confusion is actually hampering the growth and adoption of Linux at desktop.
Hope to see a strong LSB compliant Linux one day.
Amen
no need to apologize… I was being obnoxious and trying to get a rise out of you on purpose, so I should be the one apologizing.
and for what its worth, I definitely agree, I think it would help in getting more users to switch if there were less differences to contend with.
How odd, I thought the point of OSS was freedom to do as you like with the code. Quick, someone tell RMS, Theo, ESR, etc they are barking up the wrong tree!
OSS has nothing to do with freedom. Read the GPL and tell me you can do anything you want with the code. You can’t. BSD license is about freedom.
Of course it is. If Linus said it, it must be the gospel truth!
One of the posters said it well when he advised OSS folks to check their egos at the door. Apparently, you guys can’t even accept a simple fact: forking is generally a bad thing. Fine. Whatever. Keep crowing about how forking promotes more flavors of ice cream — but continue to ignore standardization.
Read the GPL and tell me you can do anything you want with the code. You can’t. BSD license is about freedom.
This argument has been rehashed too many times. Both licenses promote freedom. The BSD license gives developers the freedom to do what they want, including the freedom to make their code not free (taking some freedom away from others in the process). The GPL removes that freedom, but in doing so ensures the freedom of others is not lost when a project forks or GPL code is used in another project.
Both licenses trade a certain freedom for another. Both are necessary, as they are mutually exclusive. Both licenses are just as good, depending on what your purpose is for licensing your code under said license. Neither is “more free” – they both trade away one type for another. Saying things like “The BSD license is more free” only starts flame wars. Since you are, apparently, smarter than me you should know this. Unless, of course, your point was to start a flame war?
-Preston
Keep crowing about how forking promotes more flavors of ice cream — but continue to ignore standardization.
That is a false dichotomy. I posted why I believe forking promotes better software. I also think it is bad when the forks ignore standards. You would be wise to not post such inflammatory comments.
… would you dare to give one example where there are major differences among the distros (in terms of where configuration files are stored and so on)? In my experience it’s essentially package management and GUI frontends that are really different.
I’ll answer that for him. Slackware uses the BSD style of init, whereas most distributions use the SysV style.
-Preston
> Apparently, you guys can’t even accept a simple fact: forking is generally a bad thing.
Maybe that’s because it’s *not* a fact but just an opinion. See the difference?
Thanks a million, Preston.
Both licenses trade a certain freedom for another. Both are necessary, as they are mutually exclusive. Both licenses are just as good, depending on what your purpose is for licensing your code under said license. Neither is “more free” – they both trade away one type for another.
Nonsense. The BSD license doesn’t dictate how you can use the software — or how you can use derived products. It’s simply more free. You guys resist this criticism because it conflicts with the primary motivation of the GPL — to keep people shackled to the “community”.
Saying things like “The BSD license is more free” only starts flame wars. Since you are, apparently, smarter than me you should know this. Unless, of course, your point was to start a flame war?
My motivation is irrelevant. If you consider the post to be flame bait, exercise some self-restraint and ignore it. But don’t pretend that merely disagreeing with you constitutes flame bait.
That is a false dichotomy. I posted why I believe forking promotes better software. I also think it is bad when the forks ignore standards.
Your reasoning is flawed. Forking doesn’t promote better software.
1. It creates unnecessary and multiplicative effort.
2. It makes it more difficult to enforce standards across multiple codebases.
3. Changes in one fork can inadvertently (or intentionally) fail to make it into other branches (there are many examples of this in practice) which leads to propagation of defects.
4. It greatly complicates coordination.
You would be wise to not post such inflammatory comments.
Nah. I would be wise to ignore you.
Nonsense. The BSD license doesn’t dictate how you can use the software — or how you can use derived products. It’s simply more free. You guys resist this criticism because it conflicts with the primary motivation of the GPL — to keep people shackled to the “community”.
The BSD license does, on the other hand, take away a freedom that the GPL gives me: The right to always be able to see any modifications made to my code. The BSD license doesn’t give me that freedom. By this point alone you are wrong in saying that the BSD doesn’t give up any freedoms.
Your reasoning is flawed. Forking doesn’t promote better software.
1. It only creates unnecessary effort in circumstances when a fork is not warranted. If you had been following the conversation you would realize that I, too, am against needless forks.
2. That depends on how different they are. If their purpose is to fill a niche then they don’t really care about some things other standards might care about. What about NetBSD which runs on so many platforms? Having to follow a standard base would undoubtedly leave it unable to do so. As I said, forking when it is needed to fill a niche is good for the software. Most forks that don’t change much simply use the forked project’s updated code each time the project updates and modify their own code to fit it.
3. Yes, some projects have poor organizational skills. That is their fault, not the fault of the fork. There are proprietary projects all over the world that do the same thing. They aren’t forks, but they do the same thing, so you really ought to count their time as wasted much more than the “wasted” time of the projects who choose not to communicate.
4. Yes, it does. Some of us are capable of that. Most of the leaders of more popular free software projects are. Thats why the projects become popular. Not everyone is, but in the proprietary world noone is allowed to share without a transfer of money. Some sharing is better than no sharing.
I would be wise to ignore you.
And yet you didn’t. That says quite a bit about you, especially in combination with the petty name misspelling. Do you not see how immature that makes you seem? It makes you lose whatever respect you had gained from the few valid points you made. Remember that despite how commonly we see immature posts, the large majority (those who usually, intelligently, choose not to post so they don’t have to deal with the children) of people involved in the free software community are adults with jobs, wives/husbands, children, and lives. These people will take one look at your veiled hostility and see that you are clearly biased.
Or, keep posting just as you are. I don’t mind replying. Its break at university so I have lots of time to do whatever I want. It isn’t really difficult to respond to you when you do things like cast generalizations of “you guys all do this” and trying to say that without question the BSD license provides every single freedom the GPL does and more when that clearly isn’t true (in that regard, by the way, I think you ought to step back and think about why you are defending the BSD license. It, as other licenses, is around because it fit a niche. The GPL is around because it fit a niche the BSD license didn’t – making sure that the code is always free. It doesn’t make the BSD license any less viable to those who wish to use it, and I certainly never implied that it was a bad license).
In any case, my friend, enjoy your holidays. I look forward to less hostility in your next post.
-Preston
“At the end of the day we could have a better product if they were not forked because ofcourse developer energy is not synchronized.”
This is assuming the developers of the fork would be willing to work on the original project at all! If the fork happened because developers are not happy with the leadership, and you prevent people from forking, then developers will just leave. So in the end the original project will weaken.
One of the posters said it well when he advised OSS folks to check their egos at the door. Apparently, you guys can’t even accept a simple fact: forking is generally a bad thing. Fine. Whatever. Keep crowing about how forking promotes more flavors of ice cream — but continue to ignore standardization.
You have mistaken an opinion for a “simple fact”. In fact forking has been a good thing for OSS (egcs, xorg, etc). Forking in itself has nothing to do with standards. Xorg and and XFree both conform to the X11 standard. That’s the thing about OSS, as long as you conform to standards, a fork should mean little in the grand scheme of compatibility. A fork just allows people to get features they want if a developer or project refuses to incorporate them into their program.
Why is it so difficult to integrate Linux/UNIX into Active Directory? Because Microsoft went and made a non standard LDAP implementation out of Active Directory. I can do single sign on perfectly well using an OpenLDAP server, which does its best to adhere to the standards. It’s secure, it’s fast and it stores all the info I could ever need. Besides, it also authenticates Linux, FreeBSD, MacOS X (with a slight nudge in its mappings) perfectly well. The only problem is Windows, which requires the added layer of complexity we call Samba 3.
You can’t blame Linux for not adhering to some bastardized proprietary version of a standard protocol.
to much freedom is freedom lost, if you dont defend your freedom then it will be taken away from you.
“Nonsense. The BSD license doesn’t dictate how you can use the software — or how you can use derived products.” The license itself doesn’t, but it lets developers dictate how it is used. If someone takes code, adds to it and closes the source, no one else is able to use his improvements. This may be good for the developer who added the improvements, but it is quite bad for the people who developed the code he piggybacked on, or anyone who wants to take his refinements, and refine them further. I really haven’t decided which license I prefer, but I can see the merit in each side’s arguments.
Forks can most certainly be a good thing. egcs was forked from gcc, and later it was decided egcs was superior and it replaced (and became) gcc. Sure, ideally it would have been better had the fork not happened, had the people who worked on egcs worked directly on gcc. Sometimes it just isn’t possible though. I’m not very familiar with the circumstances, but I imagine the gcc people didn’t want to make some changes that the egcs people favoured. Egcs was able to fork it, and prove themselves right.
It doesn’t necessarily even have to do with egos. It may just as well have had to do with one side being right, and the other wrong (or less right), but no one really knew which side was which until both were able to put out products that could be compared with each other. Without forking, there would have been no way to test egcs vs gcc to see which was better, there would have been only one, and it might not have been the right one.
…I’m afraid that, in fact, Preston St. Pierre is smarter than you. In any case, he didn’t see the need to resort to immature attacks and demeaning lables (i.e. “zealots”) to make his point. A very well-made point, might I add.
As for the name-stealing troll, who posts under Rayiner Hashem’s name, the best thing to do is to ignore him until OSNews introduces its new anti-troll features (I can’t wait!)
have any more
I’m curious what feature of LDAP in Active Directory you don’t think is RFC compliant? You didn’t say anything specific.
PS-> I work in the Active Direcory product group and was looking for some feedback, but alas almost all the comments are off topic.
The BSD license does, on the other hand, take away a freedom that the GPL gives me: The right to always be able to see any modifications made to my code. The BSD license doesn’t give me that freedom. By this point alone you are wrong in saying that the BSD doesn’t give up any freedoms.
I didn’t say that the “BSD doesn’t give up any freedoms”. I said that the BSD license grants “more freedom”. Forcing people to give up their code modifications isn’t “freedom”, no matter how you want to spin it. That’s like saying “Communism is about freedom because, at the end of the day, you donate your work to the betterment of the state.” Which is patently ridiculous. Simply handing somebody a pile of code and saying, “Here, go forth and play without any restrictions” (the BSD model) is the essence of freedom. GPL shackles them to “the community”. I could care less how philanthropic you consider your “community” to be. When you force me to do something, that isn’t freedom.
1. It only creates unnecessary effort in circumstances when a fork is not warranted. If you had been following the conversation you would realize that I, too, am against needless forks.
Not true. When you create a fork, you now need to maintain source code in more than one place. By definition, that requires more effort.
2. That depends on how different they are. If their purpose is to fill a niche then they don’t really care about some things other standards might care about. What about NetBSD which runs on so many platforms? Having to follow a standard base would undoubtedly leave it unable to do so. As I said, forking when it is needed to fill a niche is good for the software. Most forks that don’t change much simply use the forked project’s updated code each time the project updates and modify their own code to fit it.
You would be better off not forking, if all that you’re doing is a one-off. As for your NetBSD example, I completely disagree that proliferation of platforms requires a fork. That’s nonsense. I write software that runs on very wide variety of platforms, and only the code beneath the abstraction layer needs to be modified as the platform changes.
3. Yes, some projects have poor organizational skills. That is their fault, not the fault of the fork. There are proprietary projects all over the world that do the same thing. They aren’t forks, but they do the same thing, so you really ought to count their time as wasted much more than the “wasted” time of the projects who choose not to communicate.
You missed the point. Having to make mods in more than one codebase increases the likelihood that you will make mistakes. I’ve seen this time and time again in various projects, where somebody fails to apply a patch that was needed, and it wreaks havoc. Results in unpatched security vulnerabilities, etc.
4. Yes, it does. Some of us are capable of that. Most of the leaders of more popular free software projects are. Thats why the projects become popular. Not everyone is, but in the proprietary world noone is allowed to share without a transfer of money. Some sharing is better than no sharing.
Collaboration between human beings is difficult. Particularly for geeks with poor social skills. You’re underrating the difficulty of this aspect of forking.
These people will take one look at your veiled hostility and see that you are clearly biased.
Duh. I AM biased. We’re ALL BIASED. I see things from my own perspective. So do you. If you can’t at least admit your own bias, you’re kidding yourself.
I think you ought to step back and think about why you are defending the BSD license. It, as other licenses, is around because it fit a niche. The GPL is around because it fit a niche the BSD license didn’t – making sure that the code is always free. It doesn’t make the BSD license any less viable to those who wish to use it, and I certainly never implied that it was a bad license).
I’m defending the BSD license because I prefer freedom. Freedom from having to hand my modifications to code back to the OSS community. Freedom to do whatever I want with the code. Perhaps give it away. Perhaps commercialize it. Whatever. The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about. And then people have the audacity to tell me that I should be glad about it. Rrrrrright.
[BSD is better blah blah blah]
You are, once again, only looking at the freedoms of people who take your code. You are not looking at the freedoms of the people who wrote the code. The GPL gives the person who wrote the code the knowledge that their code and any modifications of their code will always be available to them. Of course, I already said that, so at this point its pretty obvious you are speaking without taking anything I’ve said into consideration.
Not true. When you create a fork, you now need to maintain source code in more than one place. By definition, that requires more effort.
You didn’t say more effort. You said unnecessary effort. Effort is needed to do new things. If the project fork does nothing, it is unnecessary effort. If the project fork does something needed, it is necessary effort.
You would be better off not forking, if all that you’re doing is a one-off.
In some cases, yes, and I said that already. Not all forks should happen.
You missed the point. Having to make mods in more than one codebase increases the likelihood that you will make mistakes.
Yes, it will. What makes you think I don’t know that? Dear boy, all you’ve done so far is give me circumstances in which a fork is not a better option than not forking. You are stalwartly refusing to admit that other scenarios exist than your own, however, no matter how much evidence I give you to the contrary. I admit readily that there are places where software shouldn’t fork. You simply refuse to admit that there are places software should fork.
Collaboration between human beings is difficult. Particularly for geeks with poor social skills. You’re underrating the difficulty of this aspect of forking.
Ok, since you know more about me than this, explain to me something: If it is difficult for two projects consisting of four people each to work together, then how is it so much easier for one team of eight people? Code version repositories take care of most of the hard work. As an example, security patches to one will work on the other unless that exact piece of code has been modified. Usually it isn’t.
Now if, for example, the projects split apart for personal reasons of the people leading them there would be no collaboration – but at the same time, there is no way that these people would ever be able to work together on the one project, so a fork, while requiring more work, still gets more done than no forking would have.
Duh. I AM biased. We’re ALL BIASED. I see things from my own perspective. So do you. If you can’t at least admit your own bias, you’re kidding yourself.
I am biased insofar as everyone is biased. The difference is, I try to not let it seep into what I say. I look up facts to base my arguments on. When someone proves a point, I accept that point. When someone says something intelligent, I listen to them. I admit that there is more than one “right” or “wrong” in most circumstances because life isn’t binary. I don’t stubbornly argue my points just for the sake of argument. You, on the other hand, have exhibited none of the above qualities.
I’m defending the BSD license because I prefer freedom. Freedom from having to hand my modifications to code back to the OSS community. Freedom to do whatever I want with the code. Perhaps give it away. Perhaps commercialize it. Whatever. The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about. And then people have the audacity to tell me that I should be glad about it.
Ah, here is where the truth really comes out. You want to take open source code and make it your own without giving back. For you, the person who takes the code, the BSD license grants more freedom. For me, who writes the code that you wish to take, the GPL grants more freedom. Are you still unwilling to admit that? If the past has taught us anything, I guess so.
This line really strikes me as insightful into the kind of person you are:
The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about.
You don’t care about the people who write the code, then, you just want to be able to steal their code and use it to your own advantage without giving anything in return. You think nothing of stealing all that code and you freely admit you don’t give a damn about the people who worked long, hard hours building that code. Yet, you still don’t see why people would want the GPL there to protect the code that you so much want to steal. The GPL is doing its job and protecting these developers from having their code stolen by you, and you are angry about it.
Well, I don’t care if you are angry about it. Its their code. The GPL and the BSD license both give one freedom in exchange for another. The people who write BSD code don’t mind if you steal their code, so steal it, but the people who write GPL code do care, which is why they used the GPL license – to protect their freedoms. You can’t see that, I guess, because of how biased you are, but you should make the attempt anyway. It would certainly bring merit to the rest of what you say instead of making you look like a raving madman who can do nothing but say “BSD is better, BSD is better, BSD is better.”
I was hoping you would see the advantages to logical argument, civil speaking, and using your intelligence but apparently I was wrong.
-Preston
it that the GPL gives control(or freedom to control, depending on how you look at it…) to the developer and many freedoms to the user whilst BSD gives all the freedom to the recipient, while reserving none for the developer?
That would be what I just said, yes.
If the goal is desktop linux adoption, then I think too many distros can be a problem. I think OS adoption can have a self perpetuating effect: the more people use it -> the more people develop for it -> the more people use it. As it is Linux has about 2% of the desktop – now split that up 200 ways. What happens is you never get critical mass. Destop Linux isn’t worth developing on -> so nobody uses it -> so nobody develops for it.
Many Linux advocates don’t understand the importance of applications. Most people don’t run an OS just to run the OS. It’s all about apps. Let’s face it: several standard apps don’t run natively on Linux: Quickbooks (80% of small business market), PhotoShop, MS-Office, Visio, MS-Project, AutoCAD, etc.
It’s difficult enough to get major software manufacturers interested in developing for desktop linux without fragmenting the market 200 ways.
All JMHO, of course.
>> wolf
>>> Hope to see a strong LSB compliant Linux one day.
I do not.
From LSB :
—
Supplying an RPM format package is encouraged because it makes systems easier to manage. A future version of the LSB may require RPM, or specify a way for an installer to update a package database.
—
After using the gentoo portage system, I never want to go back to an rpm/deb base distro.
– It’s flexible
– Configurable
– Optimized
– easy to maintain
– Never face dependencies problem (debian, if you uninstall abiword it will uninstall X11, well X11 depend surely on abiword – Mandrake, xmms deinstallation make X to disapear, wow)
– it is more up to date
The problem is that it’s slower to install.
So if LSB want to force a package file format, to bad but this is madness, this standard should not to be follow.
The GPL takes away much of that freedom in the name of people that I could, frankly, give a damn about.
You mean in the name of the people who originally wrote that code…
GPL protects the freedom of users more than developers, BSD protects the freedom of developers more than users. Each deals with different aspects of freedom.
Sometimes you have to be restrictive in order to protect freedom (i.e. the U.S. Constitutions, which imposes restrictions on the majority so that the minority isn’t sujected to its tyranny).
You just want a free lunch, i.e. getting free code and not have to contribute anything back.
So if LSB want to force a package file format, to bad but this is madness, this standard should not to be follow.
You want us all to go back to compiling from source and you think that a package is madness? Please, go back to wherever you came from, you won’t help the conversation here. Advanced packages and package management systems are one of the major improvements in GNU/Linux over the past few years. The ability for me to type “apt-get update” and “apt-get upgrade” and have everything upgraded and working in an hour is astronomically useful. While I respect that some people might want to compile from source, trying to say that everyone should always compile from source and that noone should support a package management system is, well, stupid. I’m sorry, but it is. Packages are here for a reason. No need to take a step back. Packages don’t stop you from doing what you want via source. You can even compile from source and turn that into a package, then install using the central package management system for the distro.
To review: Package management > no package management. Source can be used whether there is package management or not, so you get what you want even if there is a package manager. Conclusion: Leave the package manager there and let you, a minority, figure out the source yourself. It only makes sense.
“You can’t blame Linux for not adhering to some bastardized proprietary version of a standard protocol.”
Well said!
It takes some people time to learn that A + B may not lead to C. This is especially true when it comes to proprietary v. open source computer operating systems.
Maybe I was unclear, but my point was, that a standardisation must not force any kind of packaging system. If you prefer to update your system typing “apt-get upgrade” instead of “emerge –upgrade world” is not finally the point, but a basic structure on what any package can relied on.
If LSB want to force me using rpm, I will not support that kind of standardisation and it will not bring Linux any good.
Anyway, it is quite unrelate to the main topic – using user in MS AD in Linux.
I agree that they shouldn’t force one type of package management system on you. All they should do is specify where certain things should be located and have LSB rpms, debs, etc all point to those locations. I’m sad to see they don’t do that instead (although I believe they do it as well).
-Preston
GCC once forked into gcc and egcs.
Eventualy gcc came out stronger.
egcs came out stronger, and then took the name of gcc.
egcs came out stronger, and then took the name of gcc.
And in effect, the GCC project became stronger. With GPL, it doesn’t have to be the same underlying code to be the same project. ecg (the program) effectively became gcc (the program) and therefor improved the both the gcc program and the GCC project.
I thought that was a great article and would appreciate more posts to links along those lines. Quick and dirty hands on articles. Fantastic!
Of course maybe it’s too pragmatic, we’d much rather discuss ethereal things that have zilch to do with the topic.
GPL protects the freedom of users more than developers, BSD protects the freedom of developers more than users. Each deals with different aspects of freedom.
Rubbish. It depends entirely on how the code is used and how users define “freedom”. Most users tend to associate freedom with capability — not how the capability was provided to them. There is evidence that some MS code (ie. TCP/IP stack) evolved from BSD. The fact that the vast majority of desktop computers are running that derived code gives those users more capability than they had previously. I think that an argument can be made that private entities have leveraged that code under the BSD model more effectively to provide capability than a bunch of GPL’d projects.
Sometimes you have to be restrictive in order to protect freedom (i.e. the U.S. Constitutions, which imposes restrictions on the majority so that the minority isn’t sujected to its tyranny).
Those same words could have been uttered by Lenin and Marx, too. “Look, so we’re repressing you a little — at least you’ll be taken care of…”
You just want a free lunch, i.e. getting free code and not have to contribute anything back.
Don’t knock free lunches. Many huge leaps have taken place on the shoulders of free-lunch providers (ie. Darpa, US government).
You are stubborn and arrogant, you refuse to listen to logic, you insist that developers freedom means nothing (when it is the developers who choose the license), and you continue to flame people.
This conversation is over. I have made my point.
You are stubborn and arrogant, you refuse to listen to logic, you insist that developers freedom means nothing (when it is the developers who choose the license), and you continue to flame people.
Look, of the two of us, I’m the only with experience managing software engineering teams. I’m the only one who has had to deliver multiple complex projects and juggle human and technology assets. I’m the only one who has seen firsthand what happens when you fork projects and try to keep things in sync. It’s tough. Damned tough. And then, there you are, trying to tell me that it’s a good thing. Wrong. Forking has extremely narrow advantages — but most of the time, it’s a bad idea. You can call me arrogant and stubborn for asserting my real world experience but, compared to you utter lack of experience, I’m willing to suffer such criticisms.
As for BSD vs GPL, you conveniently ignore the fact that the Internet has largely grown into what it is as a result of unfettered sharing of technology by folks like DARPA, the U.S. Government, various universities, Bell Labs, researchers, and countless commercial entities. Most OSS projects are good at one thing: Copying existing commercial technology (Samba, Linux, etc). That isn’t innovation, my friend. That doesn’t promote technology advances. It merely moves pieces around the board. But there’s a reason why it has happened. The investment inherent in commercialization tends to promote rapid development.
When you hand somebody a resource and say, “Here’s a resource that you can use — no strings attached — use it to build something cool”, you’re more likely to see that actually happen in a rapid way when the recipient has a commercial stake in the outcome. As far as technology goes, greed is good. Proprietization is good. Eventually, it serves as an incubator for even greater advances. People tend to forget that Linux wouldn’t even exist if Torvalds hadn’t modeled it after commercial UNIX.
That’s why, I believe, the BSD license works far better. It allows people to utilize code for whatever purpose they find necessary without shackling them to the requirement of revealing their special sauce. And there are many examples of commercial companies (ie. Id, AT&T, IBM, etc) releasing their code into the public domain after it has paid for itself. This is a better plan for evolution than the forced mediocrity endemic in the GPL.
Say what you will. I told you, I’m not discussing it anymore. You know my point of view, and you know the reasons I have taken those views. I look at your reasons and I see holes in them, the same holes every time, but I’m tired of pointing them out time and time again. Enjoy yourself.
-Preston