Towards A Real Business Model For Open-Source Software

[Sprewell]

monraaf, that's a silly argument. I could equally well say it's all non-GPL code like Apache and lighttpd and X11 that is keeping linux alive. Linux takes a lot of code from BSDs (just like the corporations that kraftman bemoans) and GPLs it, which artifically inflates its codebase while the BSDs can't do the converse with GPL code. This is allowed because of the true freedom of the BSD license. Both GPL and BSD OSs benefit from other licenses right now, but only one will survive.

kraftman, your netcraft link shows Apache market share, which is licensed according to the BSD-alike Apache license. Your arguments are so goofy they're hilarious. I don't think you know what a straw man is. Google runs a modified linux kernel on their servers, which they largely do not provide the source for, since they don't have to under the GPL. Many other hosts do the same for all kinds of GPL software on the server, that's why some people are now pushing the Affero GPL. You know nothing about OS X and keep ignoring the actual info we're providing, so please stop talking about it. As for taking out OS X, they're next for my hybrid model to take out after linux cuz they actually have some market share.

Apopas, nice try putting words in my mouth but I never said OS X is successful on the desktop, I merely noted that it is much more successful than desktop linux. It is tough to measure the actual number of servers using each OS, but the sales numbers seem to indicate the opposite of Ballmer's throwaway line in an interview. I suggest you actually look at the numbers you're comparing for linux vs Apple: that $36 billion figure includes anything that even has the word linux attached to it, including closed-source software like Oracle for linux. That means it's hardly a measure of the success of GPL software, but even if you insist on that measure, Apple by itself has more revenues and the Mac ecosystem revenues are probably 2-3 times bigger, so BSD still wins by a lot.

[Wyatt]

I don't feel like digging through this entire thread searching, so I'll just play ignorant Joe Blow and ask it straight: Does anyone have any real statistics on Linux usage? Note that old statistics are automatically suspect, as are any statistics that don't disclose methodology. Show me stats that are current, gathered using a methodology that has any semblance of a chance at all of coming up with a representative sample.

[kraftman]

monraaf, that's a silly argument. I could equally well say it's all non-GPL code like Apache and lighttpd and X11 that is keeping linux alive.

Apache, X11 and lighttpd are mainly running on Linux, so I can say they're living thanks to GPL - Linux + GNU utils. They're also available for *BSD, but *BSD are much less succesful same time then Linux is, so Linux (GPL) is driving and keeping them alive (or at least keeps them in a good shape).

Linux takes a lot of code from BSDs (just like the corporations that kraftman bemoans) and GPLs it, which artifically inflates its codebase while the BSDs can't do the converse with GPL code.

Which of the "a lot of code" Linux already took? While Linux can benefit (in theory of course, because there's nothing interesting to take imho) being GPL from BSD licensed projects and those projects can't benefit from the GPL projects how this makes BSD better? There's no logic in this.

This is allowed because of the true freedom of the BSD license. Both GPL and BSD OSs benefit from other licenses right now, but only one will survive.

This is allowed, because BSD license allows to do what everyone want to do with the code. It's not a true freedom. I don't get this. How BSD license does benefit from GPL (you said it can't do the converse)? I assume you agree the GPL helped *BSD to gain some popularity?

kraftman, your netcraft link shows Apache market share, which is licensed according to the BSD-alike Apache license. Your arguments are so goofy they're hilarious.

You're funny. And this Apache runs on Linux. The main problem is with people like you (bsd faboys), because they focus only on their damn, lovely license and blindly tries to proof the worse is better. If there's no better server then Apache then why would I like to run a worse one? It was already said, but it seems I have to repeat - license is not everything, the GPL license gives some guarantees like the code will always stay free as in freedom, it won't support competitors and it also guarantees contributors under very smart conditions. All of this allows GPL projects to be far more successful then *BSD ones - Linux vs *BSD. This if few BSD or Apache lincesed projects are more successful then GPL ones has nothing to this discussion. If there are two identical projects and developer groups and one is using GPL and another one BSD, there's incredibly bigger chance the GPL project will be more successful. If Apache group was more skilled etc. or some GPL group didn't bother to make an Apache like server then it's something natural Apache is more successful. It also doesn't mean it's successful thanks to Apache like license. The same can be said about Firefox etc.

I don't think you know what a straw man is. Google runs a modified linux kernel on their servers, which they largely do not provide the source for, since they don't have to under the GPL.

There are many others then "just" Google, but Google is running Linux (some proof they're using non GPL patches?) + some kernel patches (according to what you said). You're missing the obvious point which seems to be something natural (I will consider you're right about mentioned patches):

Linux + patches = Linux on Google servers - still Linux, because they only used some patches, a success thanks to Linux,

*BSD + patches + proprietary userspace applications and technologies - OS X - a success thanks to proprietary userspace applications and technologies, of course you can probably call it's *BSD, but this will not be a pro *BSD argument, but rather anti *BSD and pro proprietary Apple addons.

Many other hosts do the same for all kinds of GPL software on the server, that's why some people are now pushing the Affero GPL. You know nothing about OS X and keep ignoring the actual info we're providing, so please stop talking about it.

Damn, thanks for confirming what I was talking about!

As for taking out OS X, they're next for my hybrid model to take out after linux cuz they actually have some market share.

Yep, as a desktop which is aiming at desktops since the beginning they have some market share

That means it's hardly a measure of the success of GPL software, but even if you insist on that measure, Apple by itself has more revenues and the Mac ecosystem revenues are probably 2-3 times bigger, so BSD still wins by a lot.

OS X wins here and not the *BSD. FreeBSD, OpenBSD etc. nearly don't benefit from OS X popularity (and if they it's only, because Apple was "generous") and they're not able to compete with it, so they're looser on desktops, because it's also one way road BSD → proprietary.

[kraftman]

You know nothing about OS X and keep ignoring the actual info we're providing, so please stop talking about it.

Oh, I know some things about it. It's the beginning of the end mister They/you even started to attack Linux, because of the fear. Linux is catching up rapidly and this is why folks like you want to aid Apple by additional bsd licensed code, but the GPL says: keep your dirty hands off

[Sprewell]

kraftman, wow, the stupidity of your arguments plus the constant repetition of arguments I've already debunked is breathtaking. Keep making these idiotic arguments, because you really make the GPL zealots look bad with your illogic. Linux keeps non-GPL software alive? Even though the BSDs have been around and running them for much longer? What?! If you don't know about the BSD relicensing issue with Theo DeRaadt, where even the FSF said GPL devs can't just take BSD code and ditch the BSD license, like some GPL guys tried to do, then you're clearly ignorant of the real issues here. Sharing code makes BSD better because it has led to a much more successful desktop like Mac OS X. I see, anyone can do what they want with BSD-licensed code... but that's not true freedom. Funny how GPL zealots constantly hold contradictory thoughts in their mind in order to rationalize their silly license. The BSD license still benefits from GPL software for the same reason monraaf gave: we run GPL-licensed software like GNOME or git on BSD also, just as linux would be useless without permissively-licensed software like Apache or X11. I see, so Apache only runs on linux? Funny how you say GPL is far more successful than BSD right after using Apache, which is under a permissive license like BSD, as your evidence of popular open source software. The GPL has done well in certain niches, but if you actually look at how much BSD code is used vs GPL, BSD kills GPL, just as Mac OS X kills desktop linux. As for why those projects are successful, a lot of it has to do with corporate funding. Companies who use a service model, like IBM or Red Hat, are more likely to fund GPL code, while companies who want to incorporate code into a product, like Apple with their BSD userland or WebKit or llvm, are more likely to fund BSD code. The BSD products win because their economics scale.

As for Google's modified linux kernel, I was able to find info about it in 5 seconds with a simple google search, I'm sure you could do the same. I see, so when Google patches linux and doesn't share their proprietary changes, all the credit goes to linux and the GPL, but when Apple adds proprietary code to BSD, that success is due to the proprietary code? Hilarious how you GPL fanboys will assert these ridiculous things to distract from the fact that GPL code mainly runs on the server, where sharing isn't required by the GPL anyway. Did I confirm that you know nothing about modified GPL code on the server or about OS X, which is it? Apple only went BSD with desktop OS X in 2001, so desktop linux has been tried much longer and failed. Though I'm sure Apple's pre-existing installed base helped, that wasn't the main factor in its demolition of desktop linux. Since OS X uses a BSD core, it is a BSD, a BSD using a hybrid model that has already destroyed your beloved desktop linux. Haha, yeah, linux is catching up so fast that all the desktop linux vendors went out of business, it's forever "catching up!" We don't want your dirty GPL code anyway: it's just a shame that smart guys like Linus happened to choose the GPL cuz all that code will have to be thrown away because of the silly GPL. I don't want to aid Apple, I said they're next to be taken out after linux, because Steve Jobs has built a far too proprietary stack on top of their BSD core. Keep it up, kraftman, you really exemplify the silliness and ignorance of a GPL fanboy with your responses.

[kraftman]

kraftman, wow, the stupidity of your arguments plus the constant repetition of arguments I've already debunked is breathtaking. Keep making these idiotic arguments, because you really make the GPL zealots look bad with your illogic.

You didn't answer or didn't post any counterarguments to what I already wrote, but you're the one who's repeating same bull all the time. It makes bsd zealots to look even worse with yours logic or rather without it

Linux keeps non-GPL software alive? Even though the BSDs have been around and running them for much longer?

That's what I said. As you said BSD OS's benefit from other licenses and while we're talking mainly about the GPL I assume you meant the GPL. While Linux (GPL) is far more popular, it droves and keeps in good shape X11, KDE, Gnome development and it keeps BSD still alive. It doesn't matter if BSDs have been around, because they didn't ever bring such attention like Linux and some other GPL projects.

What?! If you don't know about the BSD relicensing issue with Theo DeRaadt, where even the FSF said GPL devs can't just take BSD code and ditch the BSD license, like some GPL guys tried to do, then you're clearly ignorant of the real issues here.

Why they can't? Can you answer? TDR was bitching bsd license, because people were sucking what they ever wanted from OpenSSL and give nothing back.

Sharing code makes BSD better because it has led to a much more successful desktop like Mac OS X.

Nope. OS X is a proprietary desktop and BSD are not. If you want a proprietary Linux then it's even bigger shame Phoronix posted such idiotic bull. OS X isn't much more successful and it's not a *BSD merit it's more successful then Linux. It's more successful, because of reasons I already posted, but you're still repeating same bull. OS X is damn poor OS with poor market share as for OS aiming at desktops since beginning. In contrary, Linux has great market share in the server side (this is one of the first areas which Linux was aiming for) and HPC.

Primary Linux goals - server and HPC - great success. Its model rocks.

Primary OS X goals - desktop - very poor success. Its model sucks.

I see, anyone can do what they want with BSD-licensed code...[/QUOTE]

It's not the first time there's no logic in your claims. If this is true then why did you mentioned TDR before?

but that's not true freedom.

Of course it's not. If anyone can do what he wants in some country, so including killing it's a true freedom in your logic?

The BSD license still benefits from GPL software for the same reason monraaf gave: we run GPL-licensed software like GNOME or git on BSD

They're still living thanks to this.

ust as linux would be useless without permissively-licensed software like Apache or X11.

They can be GPL as well and I bet they will be even in a better form right now.

I see, so Apache only runs on linux?

According to statistics it mainly runs on Linux straw man.

Funny how you say GPL is far more successful than BSD right after using Apache, which is under a permissive license like BSD, as your evidence of popular open source software.

I explained this. Apache is just one of the many other projects and it runs mainly on Linux, so GPL software. Linux, GCC, KDE, Gnome are far more successful then they're BSD equivalents (if any...). So it's funny how you say BSD is more successful.

The GPL has done well in certain niches, but if you actually look at how much BSD code is used vs GPL

I don't have to speak with the lier do I?

For example, at this moment there are 47,392 projects that use GPL out of the 68,854 total projects that use OSI-approved licenses. That is, about 69% of the roughly 69,000 projects that use OSI-approved licenses, use GPL.

BSD kills GPL, just as Mac OS X kills desktop linux.

Phoronix posted the trolls article. Nice. GPL kills BSD - look at server and desktop market. I don't like speaking to idiots, because it's a waste of time, but I'm quite persevering. It's Linux which is to kill OS X (like it kills *BSD), because OS X was earlier on desktops.

As for why those projects are successful, a lot of it has to do with corporate funding. Companies who use a service model, like IBM or Red Hat, are more likely to fund GPL code, while companies who want to incorporate code into a product, like Apple with their BSD userland or WebKit or llvm, are more likely to fund BSD code. The BSD products win because their economics scale.

Webkit is bsd licensed? It's a kthml fork. I'm not interested what is more likely or not in your opinion, because you already showed you're saying bull The flag BSD products like Freebsd and others are nearly dead. However, you can probably say they have won.

As for Google's modified linux kernel, I was able to find info about it in 5 seconds with a simple google search,

And there's a lot in that tree. Google started with the 2.4.18 kernel - but they patched over 2000 files, inserting 492,000 lines of code. Among other things, they backported 64-bit support into that kernel. Eventually they moved to 2.6.11, primarily because they needed SATA support. A 2.6.18-based kernel followed, and they are now working on preparing a 2.6.26-based kernel for deployment in the near future. They are currently carrying 1208 patches to 2.6.26, inserting almost 300,000 lines of code. Roughly 25% of those patches, Mike estimates, are backports of newer features.

I'm now sure I'm speaking to damn troll and a moron. I see I don't even have to repeat to the rest of your bull. However, If you want to say something next time then quote parts you want reply to. But, you being a troll I don't expect you will do so.

[Hoodlum]

Hoodlum, redundancy is how a market works and is precisely why it works better than any planned system you would prefer: Linus understands this very well.

Oh, please we both know this isn't always the case. Do you want 20 video players that "mostly work" or one that works really, really well? Do you want a horrendous mess of an audio stack (Linux, Competition) or a solid one (Mac, Co-operation)? KDE vs GNOME is another massive redundancy which hasn't pushed the boundaries of anything, only slowed progress. Sure, *some* redundancy has it's place but a lot of it is needless and not particularly helpful, too.

As I've already pointed out, 5 years in medical software is not that long: the released source will certainly be usable, if a bit behind.

I was under the impression this was meant to be an alternative to the service model, not just a model for a really niche segment.

There will also be development on the open source core that will be funded by these closed-source components, development that will be open source from day one. Rapid development simply implies using a smaller time limit for hybrid source, as I already pointed out with the 18-month time limit for web browsers.

So this results in F/OSS always being the poor-man's option. As I previously stated, this does not seem a way forward for F/OSS. The code is essentially locked up for an indeterminate period of time (as long as is necessary to make the required profit to make the model viable) and "eventually released".

If you could use Windows Vista for free or Windows 7 at cost, which would you pick? This seems to be a band-aid specifically for small projects under licences which don't require reciprocation.

As for your claim that "free alternatives to any closed project would be created and improved before the code could ever be released," that is complete and utter nonsense.

Well this depends on the community. If there is great demand, sure. x264 being a good example where no closed project even comes close to competing.

If you're referring to really niche projects again, then it's quite possible the closed project would be the standard. If there are no alternatives this model could help open the code, but it's less desirable than the code "just being open" from the start.

If it's so easy to clone closed-source code, why don't we have pure open source versions of all closed-source software now?

This is a ridiculous quote. "If water is so abundant why do people die of dehydration" is equivalent.

If proprietary projects have such an innate advantage (as you seem to be suggesting) why is it Linux (as an example) is used almost unilaterally on the TOP500 list, and extensively (and growing) on servers? Are you telling me there is no commercial demand? Your perception of it's "failure" just doesn't seem to apply. Are you specifically talking about niche projects for which there is reduced (or no) demand?

Resources are always limited, I have no fear of such pure open source forks.

This applies to closed source software too. Closed projects aren't a magic bullet to infinite resources.

I don't think you understand the article at all, my whole point is that the service model is a failure: open source has failed precisely because of its service model. The goal is to move to a product model using hybrid source.

Moving toward a model which makes F/OSS a second class citizen seems like a way backwards rather than forwards. I'm sure it will have a place but it doesn't seem that much different than a project-wide code bounty with the caveats that the code is locked up and released "when we can", if at all and F/OSS is eternally condemned to being the last choice - a poor-man's choice.

[Sprewell]

kraftman, hilarious how you just repeat what I said back at me like a little kid, so when I say "you're wrong because of A, B, and C", you simply respond with "No, you're wrong!" Actually it's non-GPL code keeping linux alive, because I can run a BSD without any GPL code just fine, just replace GNOME with E17 and git with subversion, whereas linux would be useless without Apache and X11. The GPL guys can't just take BSD code and remove the BSD license because it's a license, just like the GPL. It just has different requirements, that you must show the BSD license so people know where the code came from. Those GPL devs were such thiefs that they were trying to take BSD code, remove the BSD license, and relicense as only GPL, which even the FSF said they can't do. OS X is a BSD with a hybrid model, where the BSD core is open source and the layers on top aren't, funny how you can't bring yourself to accept this. I don't want a proprietary linux, I simply pointed out it already is proprietary since most linux kernels on the server run with proprietary code modifications, similar to how OS X adds proprietary code on top of its BSD core. I see, so OS X has poor desktop market share with 6 times the share of desktop linux: what does that make desktop linux market share then, horrible? Haha, when I said "anyone can do what they want with BSD-licensed code... but that's not true freedom", I was paraphrasing your contradictory statement that "BSD license allows to do what everyone want to do with the code. It's not a true freedom." Hilarious that you're so dumb that you will contradict what you yourself wrote when I quote it back to you. I see, so you have stats that show Apache mainly runs on linux, where are they? Oh, that's right, I'm the only one who ever has any stats, you just make silly claims that you don't even bother to back up. Please stop using the words "straw man," because you clearly have no idea what they mean. It just makes you look dumber every time you use them. BSD is far more succesful because there are a lot more people actually using BSD code in Mac OS X or other products than those using linux or KDE. Counting GPL projects on sourceforge is a silly metric because there are a bunch of projects with a couple lines of code that nobody uses. GPL fanboys like yourself talk a lot and write almost no code, which is why that happens. Look at actual successful projects like Apache or lighttpd, they're almost all under permissive licenses like BSD/MIT/Apache.

WebKit is actually a perfect example of what happens when BSD competes with LGPL. KHTML was originally LGPL but with WebKit, Apple and Google have turned it mostly BSD. My source analysis shows that it's now about 70% BSD licensed, that's what happens when BSD and L/GPL directly compete, BSD demolishes L/GPL. WebKit will soon be completely BSD-licensed. Haha, typical argument from you: I present an actual economic analysis about service/product models, you don't understand it and ignore it. I agree that linux is currently more popular than FreeBSD, but that's only because the flagship BSDs haven't used my hybrid model yet. Apple already used a cruder hybrid model with Mac OS X and its BSD core and they killed desktop linux, even though Apple started later with Mac OS X. Lol, hilarious that you quote some lines that merely say which kernels Google's using and that say nothing about whether they share their modifications, then act as though that's some big proof of something you said. Do you even understand english? If you do, try reading a bit further down where google says they don't release a lot of their modifications and don't really give a reason why. Hilarious how dummies like you call other people trolls, especially since your silly arguments are more trolling than anything I've ever said. I think this is a good place to end our discussion: I enjoy bashing on a dumb GPL fanboy as much as anyone but it gets boring when you're too ignorant to even make an argument.

Hoodlum, I don't presume to know how many video players is plenty. If someone wants to create yet another video player or OS, I'm just glad for the competition that gives me more choices. However, you're right that it would be better if they could share more common code, which is the good argument for open source that underlies my hybrid model. Hilarious how when I point out how my hybrid model can be tailored for each segment, you then say it only applies to one of those segments. Are web browsers part of the niche segment of medical software too? I don't care what's the way forward for F/OSS (as I don't care about the F for "Freedom") plus pure F/OSS already is the poor man's option, nothing changes that. However, my hybrid model will result in ten times more open source code being written, since it will be funded by the closed sections. What do you find indeterminate about a 18-month or 5-year time limit? You really need to think about your arguments before writing them: one post you say 5 years is too long, the next post you say it's indeterminate, which is it? Not sure I understand your Windows Vista/7 analogy, but there will always be people who will pay more for the latest OS. It is true that open source devs usually clone popular software like h.264 or Flash but because there's not much money coming in, what they can do is limited. A hybrid model will give them more resources because of the money coming in from a product model and hybrid code will actually be able to compete with proprietary code like Windows, not merely being happy to have 1/80th the share like the desktop linux fanboys. You're thinking too much in terms of the current closed/open source dichotomy so you're happy that there are a few purely open source projects like linux or KDE, even though most people only use closed-source software and as a result there are orders of magnitude more closed lines of code. What I look at is the percentage of lines of code across all software that is open, which is what this hybrid model will grow a lot. When all software is hybrid-source and 70% of it is open, that's a lot more open source code than today, even though very few individual programs will be purely open source then.

Your water analogy is what is ridiculous: you are the one who asserted that we would have open source forks of all hybrid code, even though you now admit that's not the case with closed-source today. Linux is not used unilaterally on the top 500, it has almost 78% share, mostly because the OS doesn't matter for supercomputers, it's the hardware that matters. So they usually spend a ton of money on the hardware and usually just grab the cheapest OS to put on it, which is linux. I never said there is no commercial demand for linux: I said carving out 1-20% market share in a few isolated markets or dominating a fairly useless niche like supercomputers is not a success. Funny how you keep saying that open source only doesn't do well in niches, when the truth is that pure open source has only done well in a few niches. Nobody said open or closed source has infinite resources. Hilarious how you quote me saying resources are always limited, then say that means I think closed source has infinite resources. You are the one who implied something like that when you said that pure open source would magically clone all hybrid code, despite very little money coming in through its weaker support model. If F/OSS were to become a second-class citizen, that would be a step up from it's fourth-class citizen status now, not backwards. What part of having an explicit 18-month time limit means releasing "when we can" to you? Funny how you continue to just make up shit. F/OSS will always be the poor man's choice because it's free, that's all poor men can afford to pay for. However, my goal is to lower the cost of software for those paying for it also, by having more sharing and innovation come from the open source parts of a hybrid codebase, and of course the poor can use the much greater volume of quality open source code generated from such a hybrid model also.

[Wyatt]

This conversation hurts my mind. Sprewell, you can bluster all you want about how your model is amazing, but you're not going to have proof for a while. Being a high-handed asshat about it in the mean time isn't actually helping anything. Your behaviour is what most would call "trolling" though a real troll would hardly invest himself so thoroughly in his own rhetoric. And you're spending a lot of effort defending your great idea on the forums of a site for Linux users.

What did you think would happen?

[kraftman]

However, If you want to say something next time then quote parts you want reply to. But, you being a troll I don't expect you will do so.

You have failed. It's a waste of time diving in such mess and reality is proving opposite to what you're claiming. I can't only realize why does Phoronix posted such bull?