Shuttleworth Challenged Over Mir Comments

[Alex Sarmiento]

Mir is CLA, Qt is CLA? Same exact thing, right? Therefore Aaron is a hypocrite, right?

Not exactly.

Qt has that open source agreement. Some say it doesn't have any teeth. It seems people who actually know what they're talking about (ie, lawyers) feel that the agreement is valid.

Well okay. For the sake or argument, lets assume the agreement has absolutely no teeth. That is to say, if Apple bought Digia tomorrow, they could entirely lockup Qt without releasing Qt under a BSD style license.

That leaves us with the real difference. Qt is LGPL2.1 where as Mir is GPLv3. There's a *world* of fucking difference between those licenses. There's a reason we're seeing a lot of platforms shift away from gcc since it migrated to the GPLv3. It isn't some stupid BSD ideology either as the BSD's would have stayed with a GPLv2 licensed gcc (this is why FreeBSD and OpenBSD are using gcc 4.2, the last GPLv2 licensed gcc, interestingly, NetBSD seems to be using newer gcc releases).

Even if you think the GPLv3 is the greatest license in the world (I don't really take issue with the GPLv3, but I can understand why some companies are adverse to it), the main argument against CLA's is asymmetry. What asymmetries do we see with Qt? You can link closed source applications against Qt, but you can't statically link them without a license. The source to any changes to Qt itself must also be made available, but closed applications can remain closed. These really aren't a big deal.

So maybe those small asymmetries are still too much for you and Aaron is still a hypocrite. There's another important point. Qt is a just a toolkit, Mir is a display server. In X, if you don't like Qt, use GTK. You can even run GTK and Qt apps side by side with no penalty other than a small increase in memory footprint. Your application will still run on any desktop in X as well. A display server is a much more foundational part of the stack. Wayland and Mir are much more mutually exclusive. We'll see some toolkits that will support both transparently, but we'll see how this approach works out. I don't think it's going to go well since most see Mir for what it is and don't want to support it.

Most technical minded and/or experienced linux people immediately saw Mir for what it was. Mir is clearly power grab. If Mir had been released under the MIT license like Wayland, I think we'd be seeing _far_ less hate. Instead, Mir is under the GPLv3 with a CLA. No way you cut it, the MIT license is far more permissive than the GPLv3. If Mir wins, we've given far too much control of a fundamental part of the stack to a single entity. This would be bad in general, but it's even worse when you realize there's no technical basis for Mir over Wayland.

I'm busy so I may not be able to respond, and I definitely won't respond to trolls.

OMG what a pathetic attempt for FUD.

You can still create you own closed source applications using QT, GTK or whatever and run it under MIR without any problem.

You can fork MIR and modified it all you want, and distribute your own version all you want. Any distribution can use MIR all they want and Canonical can not do anything about it. And all your contributions still belong to you, no matter what. What canonical is asking to you is to let them create private derivatives of you work and gather any good contribution not committed to upstream . However, if you want to create any derivative work and close it, you'll need to talk with Canonical first . There is no such control with Wayland. Intel and Samsung can fork wayland all they want, including your contributions and close it without returning anything back. That highlights some hypocrisy then, regarding the open source community

The Canonical's CLA makes sense for Canonical. All they are doing is adding some protection to their signature projects, which is perfectly fine. The Wayland's license makes no sense for Canonical. It makes sense, however, for companies like Intel and Samsung. This alone makes a case for Canonical to create MIR: The license and the CLA, even if you forget about the API approach instead of the protocolar one .

The spirit of free and open source software remains more intact with Canonical whilst retaining differentiation for signature projects against any competition who might want to create a tivo-like project. I am sorry but i think that this is a very clever decision from Canonical.

[hiryu]

Your post is one long dibble of hypocrisy and BS, you sir are the biggest hypocrite of the week and a anti-Canonical troll.

I'll bite this one for my own amusement. It's just too easy.

See, if you're going to call BS/hypocracy/whatever, you actually have to back those claims up. Just saying something doesn't make it so. And by back it up, I mean with facts and logic, not just inept hand-waving such as calling me a hater.

[hiryu]

OMG what a pathetic attempt for FUD.

You can still create you own closed source applications using QT, GTK or whatever and run it under MIR without any problem.

Never said you couldn't. But lets say company A wants to use Mir in their mobile device. Maybe they need to add some proprietary code in order to make it work. Or perhaps there are some patents that they are unable to share. This means they either pay canonical for a license or use another solution altogether.

If Mir is forked then only the terms of the GPLv3 are available for said fork. This means that the fork will end up being useless for a lot of cases.

The license obviously makes sense for for Canonical. The issue there is that it makes sense for absolutely no one else. This is why just about no one else is using it.

But you're missing the point entirely. My point is that there is a huge difference between Qt and Mir in their licensing. it's disingenuous to compare directly. Qt can be forked under a permissive license and isn't a foundational part of the stack. Digia just doesn't have that big of an advantage.

So to call the KDE devs hypocritical over a pretty different situation is a strawman regardless of whether you love or hate Mir.

Sorry for any typos, etc. I had to type this response from my phone.

You can fork MIR and modified it all you want, and distribute your own version all you want. Any distribution can use MIR all they want and Canonical can not do anything about it. And all your contributions still belong to you, no matter what. What canonical is asking to you is to let them create private derivatives of you work and gather any good contribution not committed to upstream . However, if you want to create any derivative work and close it, you'll need to talk with Canonical first . There is no such control with Wayland. Intel and Samsung can fork wayland all they want, including your contributions and close it without returning anything back. That highlights some hypocrisy then, regarding the open source community

The Canonical's CLA makes sense for Canonical. All they are doing is adding some protection to their signature projects, which is perfectly fine. The Wayland's license makes no sense for Canonical. It makes sense, however, for companies like Intel and Samsung. This alone makes a case for Canonical to create MIR: The license and the CLA, even if you forget about the API approach instead of the protocolar one .

The spirit of free and open source software remains more intact with Canonical whilst retaining differentiation for signature projects against any competition who might want to create a tivo-like project. I am sorry but i think that this is a very clever decision from Canonical.

[erendorn]

Mir is CLA, Qt is CLA? Same exact thing, right? Therefore Aaron is a hypocrite, right?

Well, even without going technical, Qt business model is to be usable by as much projects and platforms as possible, while Mir is created to fit Ubuntu and Unity as much specifically as possible.

Open source is good for:
1) knowing what you run
2) fixing issues yourself and tinker
3) sharing efforts and reusing code

Projects that are counterproductive with respect to point 3 always get flak from the community (eg: Android, Tizen, Mir, etc..).

[brosis]

To me the whole situation with Mir vs Wayland license is simple to express within analogy:

Mir:
"If you attempt to date my sister, I will kill you.
But if you pay just enough, I will close an eye.
In fact I secretly date her myself."

Qt:
"If you want to date my sister, you must pay.
In fact, I date her myself.
Otherwise, I will not allow. You can only talk (as in free speech...),
but don't worry - if I try to keep dating monopoly on her, and disallow her to talk to everyone,
I will be obligated to free her."

Wayland:
"I don't care what relations you may have"

Also, I don't understand what kind of stuff you, people, thinking Canonical values GPL are smoking.
Canonical distributes and uses a LOT of closed source software.
It officially integrates own closed source software within own solutions.

There is no sensible difference to Qt, for this matter.

With Wayland, everyone can take the server however and close it/do whatever he wants,
but simply free of charge. One must pay to "big brother" in case of Digia/Canonical, instead. Thats the only difference...

[dee.]

Your post is one long dibble of hypocrisy and BS, you sir are the biggest hypocrite of the week and a anti-Canonical troll.

[dee.]

3) sharing efforts and reusing code

Projects that are counterproductive with respect to point 3 always get flak from the community (eg: Android, Tizen, Mir, etc..).

Why do you group Tizen in with that definition? They are a Linux foundation sponsored project and reuse & share lots of code: glibc, Wayland, systemd, EFL, Mesa, libhybris, GNOME... please explain.

[prodigy_]

If Mir is forked then only the terms of the GPLv3 are available for said fork. This means that the fork will end up being useless for a lot of cases.

Yeah, useless for those who wanna close the code and to those who want to abuse FOSS for making money. Nobody cares.

I hate Mir probably even more than any average FOSS community member next to me and I still find your posts outright disgusting. Attempting to use the community disappointment with Canonical as a veil that should obscure how groundless your anti-GPLv3 remarks are exposes you not only as a troll (we have enough of your kind here) but as a very dangerous troll.

So let me just reiterate what others have already said: your anti-GPL rhetorics is based on nothing, you have zero credibility and the only thing you're really trying to do here is spreading FUD. Begone.

[mrugiero]

Yeah, useless for those who wanna close the code and to those who want to abuse FOSS for making money. Nobody cares.

I hate Mir probably even more than any average FOSS community member next to me and I still find your posts outright disgusting. Attempting to use the community disappointment with Canonical as a veil that should obscure how groundless your anti-GPLv3 remarks are exposes you not only as a troll (we have enough of your kind here) but as a very dangerous troll.

So let me just reiterate what others have already said: your anti-GPL rhetorics is based on nothing, you have zero credibility and the only thing you're really trying to do here is spreading FUD. Begone.

No, useless for those who are in that group and aren't already a big company able to pay Canonical.

Also, the quoted claim is true, GPLv3 is useless for a lot of cases, where they need or want to close up the source code. What is arguable is if we really want to care about those use cases. I do not. I mean, if they don't care about us (and the ones who give them the code), why should we care for them?
However, there are people who are willing to give their code and do not expect anything back, and those people are far better using a BSD-like license, and they should be respected as well.

[erendorn]

Why do you group Tizen in with that definition? They are a Linux foundation sponsored project and reuse & share lots of code: glibc, Wayland, systemd, EFL, Mesa, libhybris, GNOME... please explain.

Tizen is the reason that the community driven Mer project exists.
In the Meego era, Nokia and Intel had both learned to work with the community, and the community was quite happy to base itself on the corporate funded project (Meego) without forking.

Tizen was brought without announcement (you know, "we promise we are truly committed to this project" and then 9 months later "hello, new shiny project that replaces whatever you have been doing until now"), with complete indifference to community input and massive developed-in-the-dark code dumps from Samsung.

Most of all, the open governance seen in the Meego project was removed for a nice tighter control: to be a tizen association member and be able to have a say in where Tizen goes, it costs 120 000$ a year.

Apparently things are changing for the better, with the Tizen SDK evolving to FOSS licenses, and Tizen 3.0 targeting a more open governance.
But I remember pretty well how the community reacted to the Tizen announcement back then, and it certainly wasn't with gratitude or happiness.