Matthew Garrett: How-To Drive Developers From OS X To Linux

[jimbohale]

There are a million reasons why people don't use Linux and we already know them all.

The question I have is why practically nobody uses OS X. The marketshare is pathetic. Basically in the noise.

It's a safe assumption that more people use OS X than Linux on desktops and laptops.

[Del_]

Exactly and as a result if BSD had been first to market instead of Linux we'd all be using a BSD instead,

Might be, but again you are only talking about the kernel. It is one data-point. There are countless BSD and GPL projects around, lots of data to conclude from that you seem to be ignoring for some reason.

Moreover, you seem only to focus on the needs of software houses (who do not want to commit to open source). They are a noticeable but still minority of the businesses on this planet. For the rest, copy-left is superior by removing vendor lock-in. That also goes for all but the market leader among the software houses. It gives the smaller firms a chance to compete against the dominant player. Like we see Suse, Ubuntu, CentOS and Unbreakable linux do successfully against Red Hat (yes I do know CentOS was acquired by Red Hat).

The BSDs avoid copy-left for a pragmatic reason as opposed to a religious one.

I thought you were the native English speaker among the two of us, still I have a hard time following your choice of words here. As I have already told you, for me copy-left is pragmatic, and I do not believe I am alone. The FSF lists all relevant permissive licenses as free licenses, the only distinction is whether they are compatible with GPL or not: https://www.gnu.org/licenses/license-list.html How exactly does that have anything to do with religion or fanatism? It also seems that we have a different interpretation of the word pragmatic. When linux distributions pick whatever free software component they believe best does the job regardless of whether it is permissively licensed, that is pragmatism. When BSD distributions choose to spend large amount of the resources removing copy-left software simply because it is copy-left, even giving their users an inferior product, that is intolerant. It could even be considered fanatical. One side sees many licenses as OK, and realize that several licensing models can serve a purpose. The other side fanatically tries to enable proprietary extensions across the board to accommodate a few large software houses (that are in no need of help), rather than enable smaller software houses a chance to enter the market.

There are a million reasons why people don't use Linux and we already know them all.

The question I have is why practically nobody uses OS X. The marketshare is pathetic. Basically in the noise.

Apple exists on the desktop for one sole purpose, because Microsoft needs them to avoid anti-trust. Check out the event: http://www.wired.com/2009/08/dayintech_0806/ Actually, the desperate push from Microsoft to keep Apple on the desktop did not succeed, Microsoft was convicted and sentenced to be split up. Windows in one part and Office in another. However, George Bush for some reason did not want to carry out the ruling. All Microsoft needs to do is to cripple Office on Macs, and Apple's market share on desktop goes down the drain. They only half-cripple it to keep Apple exactly where they want them.

Like always they know that the real threat is from linux on the desktop, and like with netbooks during Vista, it now really hurts them again, http://www.bloomberg.com/news/2014-0...er-rivals.html Let us see if they again can succeed by dumping prices.

[Luke_Wolf]

Might be, but again you are only talking about the kernel. It is one data-point. There are countless BSD and GPL projects around, lots of data to conclude from that you seem to be ignoring for some reason.

Okay and you're completely missing the point I was getting at which is this: Permissively licensed software can coexist happily with both competing permissively licensed projects, and proprietary forks, and that while proprietary forks don't opensource everything they do tend to give something back, and both support companies and software licensing companies both coexist alongside each other.

Moreover, you seem only to focus on the needs of software houses (who do not want to commit to open source). They are a noticeable but still minority of the businesses on this planet. For the rest, copy-left is superior by removing vendor lock-in. That also goes for all but the market leader among the software houses. It gives the smaller firms a chance to compete against the dominant player. Like we see Suse, Ubuntu, CentOS and Unbreakable linux do successfully against Red Hat (yes I do know CentOS was acquired by Red Hat).

The only people who matter in the OSS community are those who contribute back (be it through development, UI design, bug reporting, QA, advertising, etc), everyone else has a null effect. Of those who contribute back, software developers, UI designers, and QA people are objectively the most important. Guess what software houses are built on?

The primary difference between Copy-Left and Permissive communities is whether you decide to include these software houses and potentially get some contributions back, or if you cut them out of the community and force any software house that is foolish enough to try to use GPLed software in their proprietary project to release their code. Most programmers are pragmatic as opposed to being fanatical about their licenses so the number of developers doing volunteer work won't significantly change between permissive vs copyleft, support companies like iX Systems are proven to work with permissively licensed code, and users are license agnostic. So the only people left who actually care about the license are fanatics and individuals who have the option of integrating an open source project or creating their own version. The former may be regarded as a constant. So what remains is those individuals who will either integrate an open source project or make their own version.

So the question then becomes: Which is greater the contributions of the license fanatics or the contributions of companies that will use your code in proprietary projects? Copy-left proposes the former, while permissive licensing proposes the latter.

As to this market leader thing, that is a function of distributed copyright and equality as well as the percentage of developers employed by any specific entity. Copy-left doesn't prevent Red Hat from turning linux into it's goals and visions, it's that they only have ~10% of the developers, compare this to Apple who has total control over CUPS and can choose to steer it however they desire.

I thought you were the native English speaker among the two of us, still I have a hard time following your choice of words here. As I have already told you, for me copy-left is pragmatic, and I do not believe I am alone. The FSF lists all relevant permissive licenses as free licenses, the only distinction is whether they are compatible with GPL or not: https://www.gnu.org/licenses/license-list.html How exactly does that have anything to do with religion or fanatism?

The fanatics are the ones who obsess over licenses, and will even go so far as claiming to not run software of one license or the other (even when that's blatantly false), and while they exist on both sides they are primarily concentrated on the GNU side. Anyone else will not be swayed heavily by licenses chosen one way or another unless it interferes with their design goals, as opposed to the technical argument of which software is better.

It also seems that we have a different interpretation of the word pragmatic. When linux distributions pick whatever free software component they believe best does the job regardless of whether it is permissively licensed, that is pragmatism. When BSD distributions choose to spend large amount of the resources removing copy-left software simply because it is copy-left, even giving their users an inferior product, that is intolerant. It could even be considered fanatical. One side sees many licenses as OK, and realize that several licensing models can serve a purpose. The other side fanatically tries to enable proprietary extensions across the board to accommodate a few large software houses (that are in no need of help), rather than enable smaller software houses a chance to enter the market.

Okay so in software development there is this concept of "Design Goals" which are things you're trying to achieve, use cases you're trying to support etc. It doesn't matter how great any particular piece of software is if it doesn't fit into your design goals. Lets say for example that you have a design goal of staying under a certain memory utilization, because you're in a memory constrained environment. So you want to integrate some functionality into your program but you don't want to rewrite your own version from scratch. Now lets say there's a software X that does everything, it's got all of the features you'd ever want, good code quality and it's just plain awesome, however it has high memory utilization as a result of all of it's features. On the other hand you've got software Y, which has very few features and maybe isn't the best code in the world, but it's got a tiny footprint.

Which one is better pragmatically? The answer is software Y. Software X may be better on almost every metric out there however it completely fails to meet the design goals of the project, because it uses too much memory.

The same goes for GPL software on the BSDs, one of the design goals of the BSDs is that there should be "no strings attached" to the base operating system (What goes on in Ports they don't care), the problem is that the GPL is incompatible with this design goal and therefore permissively licensed software that is inferior to GPLed software is still more pragmatic.

[zanny]

It's a safe assumption that more people use OS X than Linux on desktops and laptops.

This is only because there are Apple stores near their house, but no Red Hat / Novell / Canonical store. Or any of those companies distros on notebooks in stores.

This whole debate about the practicality of desktop linux is mute until there are actual linux machines in front of consumers. Wait, there were - they were Android phones, and they sold like mad and took over the mobile space.

[Del_]

The primary difference between Copy-Left and Permissive communities is whether you decide to include these software houses and potentially get some contributions back, or if you cut them out of the community and force any software house that is foolish enough to try to use GPLed software in their proprietary project to release their code.

Hold your horses. Nobody, I repeat, nobody, is shutting anybody out by making one software GPL. *All* major software houses on this planet contribute willingly to GPL projects. You seem to have painted a world view that clouds your judgment. Like Valve is demonstrating, even on GNU/Linux, there is absolutely *nothing* preventing you from distributing a 100% proprietary piece of software using a plethora of GPL tools to do it. Microsoft does it with Skype, likewise they contribute to the kernel under GPL. And you are telling me the fantatics is in the GPL camp, while the BSDs have the pragmatists? Actually, please stop the religious/fanatic name calling, it does not strengthen your argument, and frankly seems to be misplaced. There will be all kinds of people in every camp, and I cannot fathom any value in speculating which camp has most nut heads.

BTW, ixsystems is an interesting tale. You do realize that their main selling point is ZFS, right? And that they were a driving force in preventing FreeNAS to move over to Debian from FreeBSD? They are kept in business over a copy-left project designed to screw most of the open source world. Now I consider myself pragmatic, but I do have some idea of justice, and I do have some emotions. I have a very hard time accepting your only data point as anything but a provocation.

[TeamBlackFox]

There's nothing wrong with ZFS! And I wouldn't have used FreeNAS if they had moved to Debian - Debian is probably the worst GNU/Linux distro in existence. Not only did they mess up the file structure by putting things in the most idiotic places, but they also are slaves to RedHat. ZFS may be CDDL, but unlike the GPL it allows inclusion into proprietary OSes. The LGPL does, and when needed for software that has no permissive software licensed version, I try to stick to LGPL or CDDL, but given a choice, CDDL everyday. And considering the HAMMERFS being the only BSD-licensed competitor, and that the benchmarks for it have yet to really impress me, I'll stick to ZFS for now.

iXsystems is a lot less of a freak than RedHat or Canonical by far - at least they have support for copyfree software unlike the former two. I'm also glad to see PC-BSD moving to their own DE, this will be very good for them as while I like KDE, I'd rather use something with a BSD compatible license.

[Luke_Wolf]

Hold your horses. Nobody, I repeat, nobody, is shutting anybody out by making one software GPL. *All* major software houses on this planet contribute willingly to GPL projects. You seem to have painted a world view that clouds your judgment. Like Valve is demonstrating, even on GNU/Linux, there is absolutely *nothing* preventing you from distributing a 100% proprietary piece of software using a plethora of GPL tools to do it. Microsoft does it with Skype, likewise they contribute to the kernel under GPL. And you are telling me the fantatics is in the GPL camp, while the BSDs have the pragmatists? Actually, please stop the religious/fanatic name calling, it does not strengthen your argument, and frankly seems to be misplaced. There will be all kinds of people in every camp, and I cannot fathom any value in speculating which camp has most nut heads.

*FACEPALMS*
I addressed this earlier. Yes anyone can use GPL programs as ex. a kernel (assuming linking exception) without giving up sourcecode. HOWEVER! As a developer I cannot use GPL Licensed code in my own project without releasing my own source code under a GPL compatible license, which makes it unviable to sell commercial licenses. Further NOWHERE was I talking about simple users. I was always talking about developers utilizing the code as part of their own project.

[Guest]

First off, OpenSSL is not a product of OpenBSD. Despite the "Open" prefix, OpenBSD never was in control of the project.

That's correct. I always linked it with OpenBSD, but the "Open" prefix wasn't the cause. It was, because of some famous bug.

Also FreeBSD is not associated with Apple nor Microsoft officially, I work for Microsoft as a contractor - and I know this. Stop spouting lies. You know nothing of FreeBSD and it sure as hell shows.

Where did I say it's associated with Apple or MS officially? It's BSD licensed, so MS and Apple can take its code. This is what matters. When comes to FreeBSD I know quite a lot about it. I remember their "famous" benchmarks against Linux in old days. They were using the slowest malloc library for Linux.

Dragonfly is very good, but unfortunately it lags in a lot of fashions, plus HAMMER will have to mature some more before I consider using it in a production machine. FreeBSD is the most practical BSD out there right now for a desktop BSD, and it makes some robust servers too.

Yeah, but Dragonfly makes things on its own which is hardly the case in FreeBSD - zfs, Linux threads.

I used Linux for 7 years prior to moving to BSD, and thats because the entire system feels held together with duct tape, and GNU is fanaticist. Thats why I frown at GNU/kFreeBSD, but hey the BSD license is a CYA version of public domain, so I'm all for it. Anything but the GPL.

That's nothing, but FUD which is easy to prove. Nearly all of the software that Linux and BSD share is made with Linux in mind. Thus, if something is held together with duct tape it's BSD. It has to use Linux emulators for some software, it won't get best performance (MySQL), it has problems with some parts in KDE and Gnome.

GNU/Linux owes its success to the fact that, at the time of BSD, AT&T wasn't very pleased to hear that BSD was going to be released open source. And the Linux kernel was developed while the GNU Hurd was at a standstill. Meanwhile, AT&T, like SCO did in the 2000s, they fought BSD for what they saw as theft of intellectual property. This mitigated BSD for almost 3 years, so of course GNU/Linux has won that race.

Around 2003 BSD was still better in some cases, but even then Linux had much bigger support.

However, the BSDs will win out in the long run as GNU/Linux is making itself less and less unique and the GPL is causing the entire open source software community to teeter in a very precise balance. The GPL has long term implications on businesses ability to make money using a GPL product. Plus BSD has had a surge of interest in the userbase, so its not by any means a dying OS family.

Yeah, right. BSD has lost. Linux is more and more popular and successful, there are dozens of companies working on it and BSD is in stagnancy since years. I fail to see where Linux is making itself less unique. There's no competition in most of the markets. Furthermore, BSD isn't unique at all, so your argument doesn't stand. GPL is what makes Linux so successful. BSD is dying OS family and its user and developer base is very tin in comparison to Linux.

[Del_]

*FACEPALMS*
I addressed this earlier. Yes anyone can use GPL programs as ex. a kernel (assuming linking exception) without giving up sourcecode. HOWEVER! As a developer I cannot use GPL Licensed code in my own project without releasing my own source code under a GPL compatible license, which makes it unviable to sell commercial licenses. Further NOWHERE was I talking about simple users. I was always talking about developers utilizing the code as part of their own project.

That didn't come quite clear to me, you seemed to be making much more wide ranging claims. Still you are wrong of course, still making too wide ranging claims. Let us have a closer look at a couple of relevant examples. Samsung opened up their exfat driver over GPL, there seems to be no serious consequences from that incident. In particular, Samsung still distributes its proprietary parts of Android as closed source. In other words. no big deal including some GPL in their most important proprietary code. Microsoft did the same, if I remember correctly it was their migration tool to win7. The migration tool was subsequently opened up, but of course Microsoft's core products remaind closed source. Again no biggie, and certainly no big loss to Microsoft. Linksys needed to open source the firmware for their wrt54g over GPL, needless to say the biggest sales success of any network router followed, and they are trying to repeat the pattern today. Google included a couple of GPL lines into their Dalvik engine. Again no biggie, they simply rewrote that part. Actually, Google could very well have gotten all of Java for free using GPL, they could still keep all proprietary parts of Android closed (gmail, maps, play, etc.) Again, no biggie. So you see, your notion that GPL is the be all end all for proprietary developers if they choose to include it in their software is at best disingenuous. Actually it is a minor problem where it occurs, even yielding positive effects in the few cases where it has more wide ranging effects. The GPL is there simply to level the playing field between contributors. I do believe that is necessary in many cases, while you seem to take the extreme liberalist view where you on principle reject restrictions on code. I am afraid there is no way our worlds will meet on this, we will simply need to see how it all works out. I will continue spending my time on what I believe is the future, and you are luckily free to do the same.

[GermanyChris]

Might be, but again you are only talking about the kernel. It is one data-point. There are countless BSD and GPL projects around, lots of data to conclude from that you seem to be ignoring for some reason.

Moreover, you seem only to focus on the needs of software houses (who do not want to commit to open source). They are a noticeable but still minority of the businesses on this planet. For the rest, copy-left is superior by removing vendor lock-in. That also goes for all but the market leader among the software houses. It gives the smaller firms a chance to compete against the dominant player. Like we see Suse, Ubuntu, CentOS and Unbreakable linux do successfully against Red Hat (yes I do know CentOS was acquired by Red Hat).
I thought you were the native English speaker among the two of us, still I have a hard time following your choice of words here. As I have already told you, for me copy-left is pragmatic, and I do not believe I am alone. The FSF lists all relevant permissive licenses as free licenses, the only distinction is whether they are compatible with GPL or not: https://www.gnu.org/licenses/license-list.html How exactly does that have anything to do with religion or fanatism? It also seems that we have a different interpretation of the word pragmatic. When linux distributions pick whatever free software component they believe best does the job regardless of whether it is permissively licensed, that is pragmatism. When BSD distributions choose to spend large amount of the resources removing copy-left software simply because it is copy-left, even giving their users an inferior product, that is intolerant. It could even be considered fanatical. One side sees many licenses as OK, and realize that several licensing models can serve a purpose. The other side fanatically tries to enable proprietary extensions across the board to accommodate a few large software houses (that are in no need of help), rather than enable smaller software houses a chance to enter the market.
Apple exists on the desktop for one sole purpose, because Microsoft needs them to avoid anti-trust. Check out the event: http://www.wired.com/2009/08/dayintech_0806/ Actually, the desperate push from Microsoft to keep Apple on the desktop did not succeed, Microsoft was convicted and sentenced to be split up. Windows in one part and Office in another. However, George Bush for some reason did not want to carry out the ruling. All Microsoft needs to do is to cripple Office on Macs, and Apple's market share on desktop goes down the drain. They only half-cripple it to keep Apple exactly where they want them.

Like always they know that the real threat is from linux on the desktop, and like with netbooks during Vista, it now really hurts them again, http://www.bloomberg.com/news/2014-0...er-rivals.html Let us see if they again can succeed by dumping prices.

You do understand that AAPL's market cap is roughly twice that of MSFT and Apples spare cash is over 50% of MSFT's market cap. I think if you were to objectively poll Apple users you'd find that MS Office is not installed on a majority of their computers. MS does have a strangle hold in the corporate world because of SharePoint, Exchange, and Office with the increase of BYOD that has potential to change and change quickly.

In my little AO which is DA Apple devises have been Approved by Army IA to live on our networks, the same applies to OS X 10.6 forward and we're not known to be progressive in any way.