Valve's Dota 2 Game Can (Still) Run Natively On Mir

[Awesomeness]

Because of the user base, Ubuntu is the largest distribution out there

That's a claim Shutleworth regularly pulls out of his ass, without ever backing it up with verifiable data.
Even if that were remotely true, it says nothing about desktops in use. If his numbers are somehow calculated via access to Ubuntu's package repos, all Kubuntu, Xubuntu, and even Linux Mint users would be included as well ? none of them eyeing Mir.

[uid313]

Meh, I wish they just went with Wayland instead of Mir.

Maybe they could port Unity to Wayland?

[gutigen]

I'm doubtful SteamOS will ever switch over to Mir, but i guess we'll find out. The thing is, I don't think they're going to want to bring in Unity to SteamOS, and without that they are likely looking at a port/maintenance work on Mir of an unknown size/scope. Unless Canonical commits to providing that support for them, I doubt it happens.

At Debconf this year Valve has said that they do not see any benefits from switching to Wayland in SteamOS at this moment. I guess Steam is just going to work through XMir in first place too.

[Awesomeness]

Meh, I wish they just went with Wayland instead of Mir.

They do. SDL2 supports Wayland as well. In fact the support for Mir is mostly just copy and paste for Wayland support code (as both are based on EGL).
SteamOS is repackaged Debian. There is not even a prospect of Mir support in Debian.

[Cerberus]

That's a claim Shutleworth regularly pulls out of his ass, without ever backing it up with verifiable data.
Even if that were remotely true, it says nothing about desktops in use. If his numbers are somehow calculated via access to Ubuntu's package repos, all Kubuntu, Xubuntu, and even Linux Mint users would be included as well ? none of them eyeing Mir.

Mir is somewhat of an persona non grata at the moment in the Linux community because it supposedly fragments Linux, not that Linux isnt fragmented already in many ways but that is not the topic at the moment, but ultimately it will all depend on the performance and compatibility, if Mir ends up being better than Wayland I am not so sure no one else is going to adopt it, after all Linux is supposedly all about choice, and if Mir performs better other distributions need to include it as an installable option at least, that unique situation will reveal just how much hypocrisy is there in the Linux community, just a little, a lot or none at all.

Because if solution 1 is better than solution 2 you need to offer solution 1 to your users as an option, if you refuse to do that you are refusing on purely "political" grounds, something that should not have place in the open source community. I like it how Arch Linux deals with it, supposedly a complex and highly praised geeky distribution but yet it sees no issues with including Unity and Ubuntu font patches in its repositories for the users that want them. Thats open source as it should be, users want it? Give it to them. They didnt refuse to include Unity and Ubuntu font patches because of "political reasons" like that Martin guy from KDE who said that he will not support Mir for political reasons, the audacity to deny users something that might end up being a better choice for them, Arch Linux approach is the way, not Martin's way. As a long time Linux user I am quite disgusted at such things where people put their personal beliefs before the community project they are working on. What if I use KDE and I want Mir for some reason? I cant have it because some douchebag said so out of nothing less than political reasons. So KDE users will have to hope for Canonical to patch KDE to work on Mir or that Mir gains traction and forces KDE to support it. Sure they can use Wayland but if Mir performs better most will want the thing that gives them more performance.

[FireBurn]

Mir is somewhat of an persona non grata at the moment in the Linux community because it supposedly fragments Linux, not that Linux isnt fragmented already in many ways but that is not the topic at the moment, but ultimately it will all depend on the performance and compatibility, if Mir ends up being better than Wayland I am not so sure no one else is going to adopt it, after all Linux is supposedly all about choice, and if Mir performs better other distributions need to include it as an installable option at least, that unique situation will reveal just how much hypocrisy is there in the Linux community, just a little, a lot or none at all.

Because if solution 1 is better than solution 2 you need to offer solution 1 to your users as an option, if you refuse to do that you are refusing on purely "political" grounds, something that should not have place in the open source community. I like it how Arch Linux deals with it, supposedly a complex and highly praised geeky distribution but yet it sees no issues with including Unity and Ubuntu font patches in its repositories for the users that want them. Thats open source as it should be, users want it? Give it to them. They didnt refuse to include Unity and Ubuntu font patches because of "political reasons" like that Martin guy from KDE who said that he will not support Mir for political reasons, the audacity to deny users something that might end up being a better choice for them, Arch Linux approach is the way, not Martin's way. As a long time Linux user I am quite disgusted at such things where people put their personal beliefs before the community project they are working on. What if I use KDE and I want Mir for some reason? I cant have it because some douchebag said so out of nothing less than political reasons. So KDE users will have to hope for Canonical to patch KDE to work on Mir or that Mir gains traction and forces KDE to support it. Sure they can use Wayland but if Mir performs better most will want the thing that gives them more performance.

Your response is a little naive, Linux isn't about choice per se, it's about freedom. For the GPL parts of the stack those freedoms allow people to do whatever they like with the code so long as if any changes are distributed the source code is made available to under the same licence. This allows people and companies to take code and improve on it or create folks - all which creates choice.

I think where your argument fails is assuming that all non technical reasons are then political reasons. Most distributions don't want to support two separate stacks, most probably don't have the resources to fund or QA two separate stacks. Even if Mir was faster it wouldn't necessarily make it the better choice. Currently only Canonical are developing it - a company that doesn't have a good track record or supporting projects that fall out of favour - and also requires a CLA to contribute code back up stream. Wayland on the other hand is being developed by the folks who maintain the old X stack, who're comprised of many companies and independent developers.

Speaking from a bugzilla point of view, if a distribution doesn't offer Mir in its repositories then it won't (or at least shouldn't) receive but reports for it. Bug reports take time to triage - even if it's only to mark it as not our bug. I hope this helps explain why what might seem like a political choice is actually technical or perhaps at least financial.

[przemoli]

Mir is somewhat of an persona non grata at the moment in the Linux community because it supposedly fragments Linux, not that Linux isnt fragmented already in many ways but that is not the topic at the moment, but ultimately it will all depend on the performance and compatibility, if Mir ends up being better than Wayland I am not so sure no one else is going to adopt it, after all Linux is supposedly all about choice, and if Mir performs better other distributions need to include it as an installable option at least, that unique situation will reveal just how much hypocrisy is there in the Linux community, just a little, a lot or none at all.

Because if solution 1 is better than solution 2 you need to offer solution 1 to your users as an option, if you refuse to do that you are refusing on purely "political" grounds, something that should not have place in the open source community. I like it how Arch Linux deals with it, supposedly a complex and highly praised geeky distribution but yet it sees no issues with including Unity and Ubuntu font patches in its repositories for the users that want them. Thats open source as it should be, users want it? Give it to them. They didnt refuse to include Unity and Ubuntu font patches because of "political reasons" like that Martin guy from KDE who said that he will not support Mir for political reasons, the audacity to deny users something that might end up being a better choice for them, Arch Linux approach is the way, not Martin's way. As a long time Linux user I am quite disgusted at such things where people put their personal beliefs before the community project they are working on. What if I use KDE and I want Mir for some reason? I cant have it because some douchebag said so out of nothing less than political reasons. So KDE users will have to hope for Canonical to patch KDE to work on Mir or that Mir gains traction and forces KDE to support it. Sure they can use Wayland but if Mir performs better most will want the thing that gives them more performance.

Would agree, but Martins way is about Martins ability to support Mir or lack thereof. As Mir would require switching distribution for Martin. Martin do not want to switch distribution.

Canonical should make Mir in a way that do not force Martin to switch distro of Martin's choice, and then Martin way will be compatible with Mir.



* All based on my impressions after reading Martins blog posts about Mir.

[anda_skoa]

Canonical should make Mir in a way that do not force Martin to switch distro of Martin's choice, and then Martin way will be compatible with Mir.

Even easier to accomplish than that.

All that is needed is a developer or a group of developers on a Mir based system committed to maintaining Mir specific code paths in KDE Plasma.

Comparable to how develoepers running OSX or Windows work on respective code paths in KDE applications.

Cheers,
_

[valeriodean]

Mir is somewhat of an persona non grata at the moment in the Linux community because it supposedly fragments Linux, not that Linux isnt fragmented already in many ways but that is not the topic at the moment, but ultimately it will all depend on the performance and compatibility, if Mir ends up being better than Wayland I am not so sure no one else is going to adopt it, after all Linux is supposedly all about choice, and if Mir performs better other distributions need to include it as an installable option at least, that unique situation will reveal just how much hypocrisy is there in the Linux community, just a little, a lot or none at all.

Because if solution 1 is better than solution 2 you need to offer solution 1 to your users as an option, if you refuse to do that you are refusing on purely "political" grounds, something that should not have place in the open source community. I like it how Arch Linux deals with it, supposedly a complex and highly praised geeky distribution but yet it sees no issues with including Unity and Ubuntu font patches in its repositories for the users that want them. Thats open source as it should be, users want it? Give it to them. They didnt refuse to include Unity and Ubuntu font patches because of "political reasons" like that Martin guy from KDE who said that he will not support Mir for political reasons, the audacity to deny users something that might end up being a better choice for them, Arch Linux approach is the way, not Martin's way. As a long time Linux user I am quite disgusted at such things where people put their personal beliefs before the community project they are working on. What if I use KDE and I want Mir for some reason? I cant have it because some douchebag said so out of nothing less than political reasons. So KDE users will have to hope for Canonical to patch KDE to work on Mir or that Mir gains traction and forces KDE to support it. Sure they can use Wayland but if Mir performs better most will want the thing that gives them more performance.

I'll try to be fast and short on the reply, so...
1) It's hard to image Mir outperform Wayland in a visible manner (i.e. a ~2 fps of difference is a no difference) because the games want to run as fullscreen all of the time. It is a simple requirement for a compositor to do that, then it will be surprising, to me, to have evidence of a big difference (if any) between Mir and Wayland on that task.
2) the gap in performance and power consumption should be evident (it is, look here) between Wayland and X or, if you want, between Mir and X, more than between Wayland and Mir.
So the performance differences between Wayland and Mir could be enough to overcome the CLA and single-distro-project issues? To me, no way.
3) Your statements about Martin's decision on Mir are simple and plain wrong. Please stop to write that shit, just limit yourself to link the blog posts if you are not able to summarize the author's opinion in the right way.

[blackiwid]

Because of the user base, Ubuntu is the largest distribution out there, Debian has a small user base on the desktop compared to Ubuntu, they cant afford to stop supporting Ubuntu, that would make them lose too many potential customers. Regardless of the fact that SteamOS is based on Debian Valve still does not officially support Debian, and SteamOS was to be based on Ubuntu but some licensing issues appeared and they switched to Debian. They were planning to go with Ubuntu all the way. Big companies will always support those with the largest user base, and in the Linux world that is Ubuntu.

Sorry thats just stupid, Steam on Linux is Debianbased its called SteamOS, and debian will use wayland, of course Valve could port Mir to debian, but I doubt that.

The argument you bring here about userbase is just wrong, Linux has maybe 3% Userbase on Desktops and 1% has Ubuntu (you can correkt me with steam-numbers but thats not all linux users only the ones using steam right now), so if steamos with steammaschines will suceed we have fast 10-20% of gamers in Linux or it failed, if it stays at 1-2% steammaschienes sailes failed, and linux as gaming platfform is more or less dead.

So for Valve Ubuntu does not matter in the long run, its ok to give them good support now, so that the old diehard linux users can beta-test their games, and even pay for being able to beta test the ports for the companies, but thats it, when steamos is out this user numbers does not matter.

BTW Ubuntus userbase for ONE distro is big, but its still only 1/3 of the hole Linux desktop users. (but I dont want to fight that number, there are no good sources, so everybody picks the sources that gives them the numbers they want).