Ubuntu Announces Mir, A X.Org/Wayland Replacement

[zanny]

So on this count Wayland wins hands-down. They have shipped.

There are only 3 real ways this will end.

1. Canonical, for pretty much the first time ever, produces original complex software that works on time, and does its job well enough to hit all versions of Ubuntu in a working state (aka, not Unity in 11.04). By nature of being a corporate entity pushing adoption, and in collusion with Valve + GPU vendors, Mir sees adoption in the steambox space (in a year) and gets driver support from Nvidia / ATI / Qualcomm / etc. Mir wins, regardless of technical merit, by just having the support infrastrcture coalescing around it. Desktop Linux suffers as Canonical directs Mir to their needs and wants, closes development under the CLA, and stifles innovation in the display server space even worse than the stagnation of X for a decade caused.

2. Mir turns out like most Canonical projects as fluff, delay, and unimpressive results. The consequence is that Ubuntu as a platform suffers, mainstream adoption of GNU loses is once again kicked back a few pegs since distributors like system76 / Dell / Hp can't realistically be selling Ubuntu laptops will a defective display server and protocol, but nobody else has been pushing hard on hardware sold to consumers with any other distro (openSuse or Fedora seem like the runner up viable candidates, though). Valve probably withdraws some gaming support because of the whole fiasco, and gpu drivers don't improve at all because Mir flops and Wayland doesn't get the industry visibility it needs, and its potential is thrown into question by business since Canonical so eagerly just ignored it. The result is we are practically stuck with X for an extradited period of time since nobody is migrating to Wayland because Mir took all the momentum out of the push to drop X.

3. The best outcome is that Mir crashes and burns, Wayland is perfect by years end and can be shipping in mainstream distros, someone at Free Desktop / Red Hat gets inroads enough with AMD / Nvidia to get them to either focus entirely on the open source drivers to support Wayland (best case) or refactor their proprietary ones to work well on Wayland (and better than they do right now on X). The pressure from desktop graphics and the portability of Wayland, given Nvidia supporting it on Tegra as well, might pressure hard line ARM gpu vendors to also support Wayland. The open development and removal of the burden of X mean a new era of Linux graphics, sunshine and rainbows. Ubuntu basically crashes and burns since toolkits and drivers don't support Mir well, or at all, and Canonical being the bullheaded business it is would never consider using the open standard (hic, systemd, git).

Sadly, the second one is the most likely.

[Pallidus]

I blame this all on wayland and their snail pace development process


wayland = 5 years bs

mir = 9 months on the making, released in ubuntu 14.04

[erendorn]

I blame this all on wayland and their snail pace development process


wayland = 5 years bs

mir = 9 months on the making, released in ubuntu 14.04

I thought that ubuntu 14.04 wouldn't be available before 2014, how do you know it will have mir?

[dee.]

I blame this all on wayland and their snail pace development process


wayland = 5 years bs

mir = 9 months on the making, released in ubuntu 14.04

Wayland is the right way to do things, and sometimes doing things right takes time. Mir is nothing but vaporware at this point.

[e8hffff]

I blame this all on wayland and their snail pace development process

wayland = 5 years bs

mir = 9 months on the making, released in ubuntu 14.04

You hit the nail on the head.

I hope Wayland team have had a jolt to get their project done and push it's implementation. I also want Mir a success as its goals are what I need.

[silix]

... and not one directed at canonical's choices (single process wm-shell design, programming language, programming method, etc), rather bashing them for the decision itself
it strikes me odd somehow, especially after hearing about this "freedom" thing for a long time now... oh nevermind...

Wayland is the right way to do things, and sometimes doing things right takes time. Mir is nothing but vaporware at this point.

you may be right on the latter (but know knows, canonical isnt new to skunk works)
but about the former who says Wayland is THE way?
with the most respect for wayland developers and their efforts, but to any problem in sw design, every individual developer may have a different solution and all may be equally valid
the fact that different developers may have different requirements or different approaches, know different languages (and here it's about a C++ monolithic shell+wm+graphic server vs a protocol and a C component for building compositors, with third parties tasked with the latter), follow different methodologies (and ehre it's about test driven development) or have different conceptions, is enough to set the respective solutions apart yet make them equally worthy

the mere fact that Mir is to be developed via TDD can make a HUGE difference development- and possibly functionality-wise, if applied correctly of course
because if applied correctly, it could mean more rapid development times (thus it may be actually possible for an Agiile developed integral solution to be delivered sooner)
because it would mean that the DS's behaviour is already formalized and apriori tested (with corner cases checked, and thus known to work correclty, apriori rather than as afterthought) the moment the code is released - thus Mir would be a formally verified desktop shell (afaik the first ever actually), thus suitable for environments where preverified code is practically mandatory (or the alternative is extensive and expensive auditing before deployment)

with Wayland, if i'm not mistaken, you dont get this - so this alone may make the whole wayland thing unsuitable for those who require, or strive for it

Canonical didn't even try to have any issues they may have had with Wayland addressed.

in the light of the above, if wayland didnt fulfill canonical's aim or requirements maybe it wasnt a no-go in and of itself
but if you have to solve a development problem and stumble upon a non viable solution (or one you deem as such anyway), do you bend to adopt it anyway (knowing and accepting you'll have to make compromises) or do you skip on it directly and implement yours without being restricted by others' choices?
if you do the latter, do you think you have to answer anyone about your decisions or not?

[johnc]

Wayland is kinda DOA if they don't have vendor support.

I'm surprised they didn't from the very beginning rope in AMD and NVIDIA and work with them to create a display server that would make their lives easier.

Then again it is an Intel project, so...

[Pallidus]

"Mir is nothing but vaporware at this point. "


here's the irony: do a search for wayland + vaporware and you will see how many people said the exact same thing about wayland.


here's where I'm coming from:

I hate xorg, tired of the stupidy that is setting up conf files to try different options and xorg needs to die, NAY, should have died already...

enter wayland

for 5 years nothing but promises and "oh it's going to be terrific" " oh it's so great"

bla bla fucking bla

I even remember seeing here and other places that ubuntu 12.04 would be using xwayland/weston what the fuck ever

12.04 didn't, then I heard the same about 12.10 that canonical was desperate to port ubuntu into wayland...

12.10 came and went and now it's 13.04... still no signs of wayland


this is a new fucking world where tech is moving fast, you can't take 5 years to make a gay display manager for fucks sake.

and now these stupid ass wayland devs act all surprised and offended that canonical said 'fuck this' and made their own...

wayland will be ready when 2015? let me fucking lol

it can be the biggest piece of shit ever made, but at least in little more than a years time canonical will release something

[erendorn]

I even remember seeing here and other places that ubuntu 12.04 would be using xwayland/weston what the fuck ever

12.04 didn't, then I heard the same about 12.10 that canonical was desperate to port ubuntu into wayland...

12.10 came and went and now it's 13.04... still no signs of wayland

Indeed, because as the Ubuntu developer said, around (before) 12.04 they decided, secretly, to stop porting ubuntu to wayland and work on their solution instead.
Don't inverse cause and consequence.

[F i L]

this is a new fucking world where tech is moving fast, you can't take 5 years to make a gay display manager for fucks sake.

You have swallowed the Canonical marketing hook, line & sinker. Mir is only at the level Wayland was ~3 years ago, and there's no realistic way they can accomplish their tasks (with the developers listed) in a single year. Most likely the only reason they could accomplish what they did in 9 months is because they are walking on the path Wayland paved.