Slashdot Mirror
Ex-Red Hat Employee Matthew Garrett Comments On the State of XMir
First time accepted submitter slack_justyb writes "Matthew Garrett, former employee of Red Hat, comments on the current state of XMir and Canonical's recent decision to not ship XMir as the default display server in Ubuntu 13.10. Noting the current issues outstanding in XMir, the features yet to be implemented, the security loopholes, and Intel's recent rejection to support Mir in general. All of this leading Garrett to the conclusion that 'It's clear that XMir has turned into a larger project than Canonical had originally anticipated, but that's hardly surprising.'"
Ubuntu is barely used by anyone as it is (try to understand that Slashdot users are the vast minority)...XMir's usage rate will now be so low that it won't garner enough support to survive much longer.
Does Ubuntu still ship with non-free firmware? I know they still lure users into running non-free software.
Still suffering from the butthurt he got when Ubuntu sided with Scott James Remnant over him in a technical dispute which then led to MG quitting like a petulant little bitch. Just like what happen when he was with Debian. Now he just takes to shitting on Canonical whenever he can. The fact is, Canonical is concentrating on getting Ubuntu Touch ready and with the technical difficulties with XMir, and made the prudent decision not to dump it as a default on the Ubuntu user base.
BTW, the while he may not work for Red Hat, he's still on the fedora advisory board. Can somebody say "conflict of interest"?
yeh!
FTFA:
Mir could have done the same, but doesn't because of a conscious design decision - in the Ubuntu Phone world, clients stop doing things when they're told to. Ubuntu Desktop is expected to behave the same way.
So they're letting design decisions for their phone interface dictate how they implement their desktop interface. It's the same stupidity that the Gnome developers are engaged in. A desktop is not "just another kind of phone," and if you treat your primary users as second-class citizens, they'll all jump ship.
How can we continue to believe in a just universe and freedom to eat crackers if we have no ale?
Who is going to run X, Mir, Wayland, or fucking SurfaceFlinger on a web server or router? You don't run any desktop environment on those systems.
What do you do when you use GUI tools for configuration?
There's really not that much wrong with Canonical or Ubuntu.
Think for yourself, in light of Snowden and others, why everyone is throwing dirt at it.
If you think the "damned new UI things" sucks, try those yourself. I'd guess 75% of you who mock it have not used it for a week.
I consider myself a power user and was horrified of the Dash and other things. AFter using those for awhile, now they make sense. I actually work much, much faster with those. YEs sometimes I have to dip into the console to fix things, but I'm not exactly mom and pop surf-and-pay-bills-and-mail kind of user. I do work in the console basically all the time.
AND the whole UI layer looks nice and consistent, relatively minimal (but not as minimal as I'd like it to be really). It looks like a tool and not like some rip-off of Windows 95.
For the Amazon spyware things, I just turn it off in privacy settings. Or uninstall that lens. Or hand-hack the Python files...
Maybe with Mir Canonical is about to eat a hat full of poop for not going along with Wayland. But they want to capitalize on hardware which runs Android without involving hardware OEMs, this way they can provide a good experience out of the box for any hardware where one can build an own Android with accelerated graphics.
It's still missing features
XMir doesn't support colour profiles. XRandR properties aren't exposed, so there's no way to control TV output encoding or overscan. There's still no hardware cursor support. Switching to XMir now would reduce functionality without providing any user-visible gain.
no hardware cursor support? talk about a dealbreaker!
Anons need not reply. Questions end with a question mark.
We are cycling around a wheel of and racing to reinvention. All of Canonical's efforts could have been paired up with Wayland to make a one-size-fits-all display server and it probably would have been finished and ready to deploy by this point. It's opposition like Mir that is stagnating much needed innovation. I imagine Canonical's thoughts going something like "Oh, since Wayland isn't moving fast enough - even though it's been in development for years - we'll just make our own, from scratch." If they didn't think it'd cost them this much effort, then they're more arrogant than I thought. The one thing about them I've learned over the years is they love biting off more than they can chew - only to spit it right back out onto the plate. If there is one positive thing to come out of this is that it lit a fire under Wayland's butt.
While nerdy cretins think the greater world exists to listen to an endless litany of 'technical' excuses, things that fail to work 'out of the box' will NEVER (never, never never) earn public acceptance. But then again, the nerdy Linux cretins say "good, we don't want ordinary people ever feeling comfortable with Linux".
That the simple act of rendering to the screen is still something billions of dollars of Linux investment still can't get right disgusts me. It is NOT a difficult technical problem to solve.
1) old hardware crap should NOT be supported in hardware mode by new Linux distributions. All older hardware lacking sufficient technical ability should be driven by software solutions alone. This means being grown up, and selecting a base-line standard for video hardware. All video hardware from recent times allow complete software solutions to baseline 2D rendering requirements.
2) having selected a baseline hardware standard for hardware video acceleration (probably GPU parts capable of Open GL ES2.0), the OS use of the hardware should be as clean and minimum as possible. Clever stuff should be left ENTIRELY to the domain of the apps themselves.
'Clever' 'clever' crap in Linux is actually mind-boggling stupid as an idea. Hardcore video/GPU apps simply want clean access to the GPU hardware, rendering through their own libraries, and expecting the OS screen composition engine to display the final result on the screen. IS THIS ROCKET SCIENCE?
It doesn't matter, thank god. Android for the desktop (which arrives when the first mains-powered ARM parts hit the market in 2014 with AMD and Nvidia graphics cores) will be the only 'Linux' that matters for 99.999% of all users.
In reality, Linux is ruined by the idiots that fail to comprehend how computers have changed. An OS should NEVER be about libraries, or power-functions. That thinking became irrelevant when the resources of even the cheapest computer exploded. Today, every app can incorporate first-class libraries to do all the heavy lifting. The OS simply needs to provide a sane shell, proper message handling of VERY low latency, and the usual hardware management. Otherwise, performance apps simply need the OS to "get out of the damned way".
Problem is, the people that work on Linux are 'fiddlers', "control freaks", "hackers", and technical introverts. None of them cares to see (or has the psychological make-up to see) the greater picture. For that, we have to turn to the OS projects from Apple and Google (and now Valve). The pity is that increasing numbers of Windows users would love to see a non-Microsoft future, but will not migrate to a joke of a platform dominated by profoundly dysfunctional types.
PS can there be any excuse for a 'fat' OS in this age? What possible justification can there be for loading the OS with complexity? All a 'fat' OS ensures is massively increased situations for bugs and errors, and a vastly slower computer as almost every piece of software has to run through insane numbers of very inefficient abstraction layers. If people need such abstraction, it should be in the coding environment, not in the underlying structures of the OS.
I understand the desire to capture the mobile market. However, like Microsoft, Canonical is making a mistake trying to merge the mobile platform with the workstation/server platform. They're different devices requiring different interfaces, and what's desirable in one is horrible in the other (see Windows 8). Since they've decided to follow that model I've decided to part ways with Ubuntu. I'm in the process of migrating my systems to Slackware, and as soon as I have a good process down all of my clients will be making the transition as well.
It's clear that XMir has turned into a larger project than Canonical had originally anticipated, but that's hardly surprising.
Isn't "something you didn't anticipate" almost the defintion of "a surprise"?
systemd is Roko's Basilisk.