5 ms·
From the article, and I'm sorry if this seems condescending, I would truly love to have a conversation about this, but I believe your question is to vague: * T
by oneweekwonder 14y ago
From the article, and I'm sorry if this seems condescending, I would truly love to have a conversation about this, but I believe your question is to vague:
* The input event handling partly recreates the X semantics and is thus likely to expose similar problems to the ones we described in the introductory section.
* The shell integration parts of the protocol are considered privileged from our perspective and we'd rather avoid having any sort of shell behavior defined in the protocol.
[1]: https://wiki.ubuntu.com/MirSpec#Why_Not_Wayland_.2BAC8_Weston.3F https://wiki.ubuntu.com/MirSpec#Why_Not_Wayland_.2BAC8_Westo...
- shmerl 14y agoWhat worries me is fragmentation of drivers. It can be a very serious problem. Possible scenario: Nvidia wait for broader Wayland adoption to release their driver for Wayland. Comes along Canonical and drops Mir on our heads. Nvidia scratch their heads and say - forget Wayland. Not sure if they'll say forget Mir as well - they have better things to do than to deal with this mess, but all this doesn't sound good already. Consider games development. People wait for better drivers (Wayland)? They aren't coming. Should they say forget this mess (Linux)? Or Canonical think they should refocus all driver development on Mir like Android did on mobile creating a horrible fragmentation again?
- oneweekwonder 14y agoI don't take the driver situation into consideration and it seems that Ubuntu is contacting closed source manufactures to allow driver support for mir. I like your scenario because I believe a interesting element is that Valve is trying to port most of its Steam games to Ubuntu and they do have muscle with Nvidia and AMD. And with the rumours that the steambox called Piston would be based on Ubuntu. So in my opinion it looks bad for distro's not using mir and good for distro's using mir. If that is a good thing altogether is another debate. And as mentioned above profesional game development to Ubuntu is coming at a rapid pace. Maybe at the price of another horrible fragmentation as you call it.
- shmerl 14y agoI don't consider Ubuntu usurping the focus of game development and drivers a good thing. And notion like "using Mir - good for you, not using Mir - too bad, no drivers" doesn't sound good as well. Ubuntu is already quite isolationist. This will only make matters worse.
- oneweekwonder 14y agoWhy would you say Ubuntu is taking game or driver development by force or illegally? As far as I can tell there is no other competition in this space, so it is actually very open and its not like Ubuntu is not willing to share. I agree with the notion, but we still need to see if it will play out that way. For now its only on the table, with a lot of other cards we are not even aware of.
- shmerl 14y agoI didn't say anything about "illegally". It's all open - "do what you want". It's the question whether it benefits the global Linux community, or creates unnecessary internal competition / fragmentation and spreading of resources. Competition for drivers can start (if they won't be compatible), look at Android targeted hardware and see how sick it is now (it's SurfaceFlinger only, try getting Wayland drivers and good luck with that). We don't need the same story repeated with Mir.
- oneweekwonder 14y agoSorry I did a define on usurping and ended up being pedantic because I don't believe its true. I understand all the concerns about how it will impact the global community; But isn't it generally seen that competition is good, and who's resources is it spreading or fragmenting? * A distribution who tries to cater to everyone? * Other companies employees who now need to understand and support technology it doesn't want to use? * "Linux" users in double quotes because they want to be able to administrate all their different boxes with different distro's using the same old tools of 20 years ago? * FOSS coders who spend their free time fixing bugs and adding features to projects they like and use? Call me ignorent and please point out why. But I don't see any of the above as resource issues. The only problem is the "Linux" users and I feel if they start to see themselves as distro's users instead of a name given to a Kernel and Foundation, that problem will sort it self out. By your account it seems like it already started and Ubuntu might be able to bring in a second or third option. By the looks of it the greater community wanted Ubuntu to rather spend time in bending tools to do what it needs to do, to benefit the community. Without realising Ubuntu is a product of a company that has goals and can not always take the long road of community inputs, but at least it tries do be open as much as it can. Which gets criticized a lot because it is a leader in a "linux" environment it mostly created from scratch.
- wmf 14y agoIt looks like both Wayland and Mir are supposed to use KMS drivers.
- shmerl 14y agoLet's hope.
- prg318 14y agoDo you have a link to this assumption? I would really like to know.
- shmerl 14y agoI found these: * https://wiki.ubuntu.com/MirSpec#Mir_on_the_Free_Graphics_Driver_Stack https://wiki.ubuntu.com/MirSpec#Mir_on_the_Free_Graphics_Dri... * https://wiki.ubuntu.com/MirSpec#Mir_on_HW_Supported_By_Closed_Source_Drivers https://wiki.ubuntu.com/MirSpec#Mir_on_HW_Supported_By_Close... we are in contact with GPU vendors and are working closely together with them to support Mir and to distill a reusable and unified EGL-centric driver model that further eases display server development in general and keeps cross-platform use-cases in mind. I hope the later means Wayland compatible drivers, because if not - it's going to be bad.
- noahl 14y agoMy interpretation (and apologies if this is obvious to you) is that they will define certain interfaces between the driver and the outside world, and any manufacturer will be able to provide something (a library or a daemon, I guess) that implements those libraries. When they say "cross-platform use-cases", I think it means they want to make it easy to build the Mir driver and Android driver from the same source tree - and quite possibly the Windows and OS X driver too. So hopefully they will have some quite general interfaces. I think the ideal scenario is that the driver interface they define has a very general interface, and then there's a wrapper that turns that interface into an Android driver, another that turns it into a Wayland driver, another that turns it into an OS X driver, and so on. But I don't know if graphics drivers would accept the performance penalty of even a very thin wrapper.
- gillianseed 14y agoI don't think so. The reason we've had Nvidia and AMD support on Linux for so long is because of the large 3d/sfx and of 'gpu accelerated number crunching presence' Linux has. The end user Linux desktop market has never really been on their radar, perhaps that will change somewhat with the advent of Steam, but really I expect that unless the aforementioned markets adopt Wayland or Mir (with the latter being most unlikely imo) then NVidia and AMD won't target them with proprietary driver support.
- syjer 14y agoFrom the discussion on google+ by the wayland developers [1], it looks like that the reasons given by canonical are quite poor. [1] https://plus.google.com/100409717163242445476/posts/jDq6BAgdpkG https://plus.google.com/100409717163242445476/posts/jDq6BAgd...