More Mir Talking Points Come Out Of Canonical

[MartinN]

And why not the other way around? Merge Mir into Wayland-- you know, the solution that actually HAS all the parts "Done right" and ACTUALLY complete instead of a giant TODO list of some of the hardest goals to achieve in a display server, let alone in 6months?

Because you are only allowed one flip-flop on your word per 2 years. It'll take 2 more before Canonical has earned enough credit to flip flop again and go back to Wayland.

[finalzone]

Canonical is just becoming a vampire at this point.

Canonical is a vampire from the beginning. Much like Dracula, the company successfully charmed majority of people to join the club then sneakily bite them.
You managed to break the spell to find out the real thing.

[IanS]

Canonical has a great touch when it comes to UI's, but they don't really have a ton of programming expertise, and the wayland team is far more experienced in creating a display server than canonical's team.

I wouldn't even give them this credit, I have yet to see a vanilla Ubuntu desktop that I could stomach and things only got worse with Unity. Any time I have gone into the *buntu pool for anything more than testing it has been with a derivative (Ubuntu Studio, Mint, Kubuntu, and once with a minimal cli install that I set up with openbox/tint2 and some Xfce bits).

[pythonista]

Canonical is a vampire from the beginning. Much like Dracula, the company successfully charmed majority of people to join the club then sneakily bite them.

Or maybe it has something to do with the fact that Ubuntu generally "just works", and the other distros, even the "stable" distros generally look, feel, and act like somebody's hobby rather than a polished product? I'm sure Canonical has a lot more cred in making usable Linux distros than you do.

I just retried Debian "stable" this morning, and it would hang on shutdown, take forever to boot up, and was in no way more stable or compelling than the much more bleeding-edge Ubuntu 12.04. Then there's the Arch-tards who suggest that Arch LInux somehow should be the dominant Linux distro, even though it's just a toy that is likely to fail to install, and even if it does install, it will pretty much break on every update. I still don't understand the alleged appeal of Arch, but I'm sure some knowledgeable Arch-tard can properly explain that the "I fixed it myself elitism" of Arch trumps Ubuntu's "just works for everybody".

[bwat47]

And regarding ubuntu's laughable claim of having mir and unity-next in an LTS by "2014", let me remind you when in 2010 they commited to having unity on wayland "within a year"

http://www.markshuttleworth.com/archives/551

Only they never did a lick of work on wayland, and never once have even contacted wayland developers, they just expected wayland to do all of the work for them apparently?

[Figueiredo]

i wonder what kind of video hardware accelation is planned on wayland/mir. the over reliance on community projects tends to lead to the current mess we have with vdpau/va-api/xvba. i hope someone is sensible enough to correct this, either on wayland or mir camp

[TemplarGR]

i wonder what kind of video hardware accelation is planned on wayland/mir. the over reliance on community projects tends to lead to the current mess we have with vdpau/va-api/xvba. i hope someone is sensible enough to correct this, either on wayland or mir camp

The same as we have now more or less. All APIs ported to Wayland... For the first couple of years...

I believe that now with OpenCL being more mature and widely used plus being available to FOSS drivers (r600g for example), the best method for the future would be to have software codecs ported to OpenCL. That way the hardware video acceleration would become universal, independent of special hardware or non-free software...

[bwat47]

i wonder what kind of video hardware accelation is planned on wayland/mir. the over reliance on community projects tends to lead to the current mess we have with vdpau/va-api/xvba. i hope someone is sensible enough to correct this, either on wayland or mir camp

vaapi is already ported to wayland afiak

[finalzone]

Or maybe it has something to do with the fact that Ubuntu generally "just works", and the other distros, even the "stable" distros generally look, feel, and act like somebody's hobby rather than a polished product? I'm sure Canonical has a lot more cred in making usable Linux distros than you do.

Only credit for Canonical is the PR department and Mark Shuttleworth fame status. Ubuntu itself was decent, Mandrake, SuSE, Mepis did far better job in term of usability and contribution to FOSS.

[Figueiredo]

The same as we have now more or less. All APIs ported to Wayland... For the first couple of years...

I believe that now with OpenCL being more mature and widely used plus being available to FOSS drivers (r600g for example), the best method for the future would be to have software codecs ported to OpenCL. That way the hardware video acceleration would become universal, independent of special hardware or non-free software...

That would certainly be a very nice fallback, but dedicated hardware is already in place and, afaik should be much more power efficient than doing over opencl, this is specially relevant in power constrained devices. Ideally I think there should be a single API to be targeted by every application and supported by every driver, pretty much how is done in windows (DXVA). In android it seems to be OpenMAX..

It makes sense that canonical also adopts the same API as android in order to leverage its driver stack, thus the display server should also support this API.

I don't know if any of the above has had any weight on the decision to develop mir, but one thing is true. Wayland is intel's project, and canonical is seeking to support a miriad of ARM SoCs, so it makes sense that they do not adopt a DS controlled by competitors. By developing it in house they can have a common ground. What do you guys think?