Shuttleworth On Mir: "A Fantastic Piece of Engineering"

[Brane215]

I bet Mark's approach will be to garner a user base that will lead to greater uptake, so don't get ahead of yourself. We are talking about Linux here, and Mark has been the only one to get it into the mainstream so far, so put some trust in him to take it further.

Without Mark and Ubuntu, Linux would be still in the dark ages. Ubuntu has brought competition by their sheer innovation, causing other players to be doing catchup.

Really ? What was so innovative technologcal component that Canonical brought to the table with Ubuntu ? Can you name just one SW product or solution that really shook the Linux world ? Or, failing that, at least one really solid tool that external world adopted fro them ?

[alexThunder]

Well then, I guess I should enjoy the remaining time I have with KDE ;P

[gamerk2]

Really ? What was so innovative technologcal component that Canonical brought to the table with Ubuntu?

None

Can you name just one SW product or solution that really shook the Linux world ?

None

Or, failing that, at least one really solid tool that external world adopted fro them ?

Nope.


You know what Canonical/Ubuntu brought Linux? Public Awareness. In the mind of the consumer, Ubuntu = Linux. If Ubuntu never came around, Linux would still be in the same state in was in 2005, holding <.5% market share, nothing more then a toy OS.

So yeah, Canonical can do whatever they want, and the devs will follow, because they, and they alone, drive commercial Linux development.

[om26er]

More clicks

Other than the reference to your old posts this article has nothing in it.

[nerdopolis]

Just admit it you're a bitter person and can't stand for Ubuntu to go big.

Just the opposite. The fact is, by them trying to write an entire display server and protocol from scratch is setting themselves up for failure. Mir disappoints me because I am rooting for the open source desktop stack.

They have such a small timeframe to complete it, and they still have to design and plan how they are going to do many complex things, and design it, never mind how much time it's going to take to code it. If you watch the UDS video ( http://www.youtube.com/watch?v=w6HnJ...layer_embedded ), the developers don't really seem sure on many of the questions they answer.

Keep in mind, that they've been working on this thing together in secret for nine months. It's not like they just started this project this week. In order to fit their narrow timeline, they are going to have to rush decisions, and design, and they don't have much experience in display servers, which unlike Wayland, the developers take their time to make sure they are making a good display server protocol, and many of them have experience in display servers.

If they manage to get Mir to resemble something working in that timeframe, they are probably going to end up with an inferior display server.

And then the Linux userspace stack would have fragmentation at such a core component. Closed developers would have to consider targeting Mir or Wayland distros. And then on every other distro that uses Wayland, won't be able to support Mir apps, unless Weston adds a plugin similar XWayland only for Mir, meaning More work for the Wayland developers to support Ubuntu's protocol! Toolkit developers would have another backend to support and maintain (provided they actually try to go upstream)

The reasons they use for not using Wayland are questionable

• Wayland developers have gotten Weston to run on a Galaxy Nexus.
• Wayland servers don't require a driver API like the X server. It uses EGL.
• And for drivers that don't support Kernel mode setting, they already have Weston working on /dev/fb* framebuffers. This would probably be used on VMs I would assume. The Mir developers have no idea how they plan to support vms.
• The fact that they say that the input system doesn't support things like joysticks, I think was actually discussed this week, could be added through a simple protocol extension. (And I read a good argument against handling all types of input devices like Mir wants, such as pupil followers, because handling and processing such advanced input is too complex to do in a display server that should be responsive)
• Their criticism of input security was completely wrong, and misguided.
• I'm not exactly sure what they don't like about shells in the protocol, but they seem to dislike the fact that the shell isn't built into the protocol? Usually people like when two major components aren't fused together I thought? Wayland supports other shells, while Mir seems to be only Unity from what I can gather.
• In the UDS video, they constantly say they want something that "runs unity", but they give no concrete reason why Wayland won't work...
• They never even talked to the Wayland developers until after this week!

When krh created Wayland in late 2008, we were in dire need of a new display server protocol. Many developers, including krh had laid the groundwork, like kms, to make Wayland possible before 2008. The X developers are even working on Wayland now. Krh and the other Wayland developers spend a good amount of time planning the design before they commit. Until this week, Wayland was the clear path to finally be X's replacement. Now thanks to Mir, it only adds complications.

Mir is only going to end up being inferior to Wayland, but because Ubuntu's the most used distro right now, many people are going to end up with Mir, which is likely going to have issues when it's released because it's going to be rushed out the door, damaging Ubuntu's reputation, and causing a fragmentation mess if people port apps designed for Mir.

Sorry about the length

[erendorn]

You know what Canonical/Ubuntu brought Linux? Public Awareness. In the mind of the consumer, Ubuntu = Linux. If Ubuntu never came around, Linux would still be in the same state in was in 2005, holding <.5% market share, nothing more then a toy OS.

So yeah, Canonical can do whatever they want, and the devs will follow, because they, and they alone, drive commercial Linux development.

The "public" linux desktop is *nothing*. Money in Linux is in servers, 3D workstations, developer workstations, embedded devices and embedded graphic devices. It's not public awareness that counts. Its corporate awareness. The people that pay for Linux are the one that fear control of other corps on the tech they use. Intel, Samsung, car, phones, set top boxes and TV manufacturers. These will not use Mir, for the very same reasons that they use Linux. Because wayland is MIT licensed and supported by multiple players, while Mir is GPL with extra Canonical CLA.
Linux is not driven by hobbyist, but by corporations. Good for Ubuntu to bring awareness to the desktop. Good for them if they bring mass market devs to their platform, and to Linux in general. But while Wayland makes much less noise, it has much more weight. And that's for the best.

[mendieta]

Right idea, wrong approach

So, I think the push from Canonical to have a simple open source offering is great. Consolidating all their codebase, great. Using a simple light interface, great.

That's all good. But.

You do want to reuse what you can, and build what you can. Android is really popular, build on that (the plan for Ubuntu on Andoid is genious). ChromeOS is picking up. Build on that (no plans for that). Make a one-click install of Ubuntu on top of the ChromeOS kernel, based on this:

https://github.com/dnschneid/crouton/

At the end of the day, it seems like they are spreading too thin. There own thin desktop? not bad. Their own display server? Maybe too much. What's next, their own kernel?

Long story short, I think they are overall on the right path, but they should pick their battles. In particular, ChromeOS and Android are not the "enemy". All of that stuff is open source, and hey, very highly successful!

[Brane215]

You know what Canonical/Ubuntu brought Linux? Public Awareness. In the mind of the consumer, Ubuntu = Linux. If Ubuntu never came around, Linux would still be in the same state in was in 2005, holding <.5% market share, nothing more then a toy OS.

Sure. In the same manner as Bugatti Veyron or Mclaren F1 are "toy cars" due to their miniscule market presence. Or as McDonalds sells surely the best food since it sells so much of it.

And all that public awareness per se did not help them to make superior product. Which is BTW for what they are aiming now. Some smart and certainly experienced dudes are making first big qualitative step away from old X11. For that one needs tech smarts, not public awareness.

It seems to me Canonical is misusing all that public awareness to get to some very personal goals. They stand on foundations of open source, but seem to be giving back very little, if anything.

So yeah, Canonical can do whatever they want, and the devs will follow, because they, and they alone, drive commercial Linux development.

We will see about that. It seems to me they are not driving but mooching off the open source development...

Compared to RedHat, they are not even a midget. Everywhere you look, you can see contribution of their people, be it gcc, kernel, system infrastruture, libraries etc.
There are numerous chunks of their brainwork in practically any and every distro.

And yet I don't see RedHat making such unsubstantiated claims and wild turns...

[arioch]

fanboys

Other than the reference to your old posts this article has nothing in it.

Agreed.

Man, and they call Ubuntu fanboys....don't these Wayland maniacs, Weston cult followers have mirrors in their places? Never seen so many bitter people ranting at once. Many consider themselves Leet...so they feel entitled to insult other people because they respectfuly disagree. Well let them. Ubuntu has come a long way, and wheater they like it or not, the name of linux is in the mainstream media very often, nation wide newspapers, frontline tech specialized websites, etc....

Out of 100 random people i would say 5-10 would have some notion of what Ubuntu is....I seriously doubt anyone would know what Wayland is otherwise, nor they would care. Just give them a FOSS OS that works. You don't need to be an engine engineer to buy a car, or a circuitboard expert to buy a blender.

Call me fanboy, call me names...I don't care, Ubuntu has changed my life many years ago, so at least they get the benefit of the doubt from me.

You all have a good day.

[Pajn]

What are peoples problems? Gosh.
This is like a school yard.

Everyone is defending Wayland with there lives and most seems to think
attack is the best defence. But, there is none attacking it.
Canonical have started a own project that they believe in. They haven?t
forced anyone to use it, and still people are vetoing patches that haven't
even been submitted. Other hope the project dies.
Why?
Wayland isn't done yet. No one is using it and in the near future no one will.
So why is Wayland the holy graal?

We haven't seen yet what Mir looks like, maybe it will become great?
We cant say yet.

Both project strives against the goal to improve Gnu/Linux graphics, isn't that
a good thing? Why pick a winner before anyone of the candidates even have
taken up the battle? Why pick a winner at all?

Lets just wait and see. If Wayland succeeds in bringing glory to Linux that is
a good thing. If Mir succeeds in bringing glory to Linux that is an equally good
thing. Why spend half of the money betting on one horse and the other half
destroying for the other one when you could bet on both and become a winner
no matter what happens?

Calm down people. We don't need a fight, we need to get together and create
good software. Even if Mir fails, Canonical may succeed in having Nvidia provide
a driver for it, a driver that will most likely also work on Wayland.

Canonical develops Mir with there money, not yours. No one is telling you where
you should put your money so stop telling Canonical.

Thanks. Now I'm going to enjoy a good nights sleep. Hope you will too
At least those of you who live in a timezone that allows them to go to bed
right now.