Shuttleworth On Mir: "A Fantastic Piece of Engineering"

[Luke_Wolf]

They didn't piss all over Wayland. Yeah they made a wrong statement or two out of error but they went back and publicly redacted that.

And that's the crux of it, Did they piss over wayland or was it in an accidental slight? As for me I do believe there was intent there particularly given how said announcement was set up. You might believe it was purely accident but you need to understand that's where we're coming from.

Also a company can redact something but that doesn't mean that they agree with the redaction and aren't going to continue to push the same nonsense along different vectors in a form they'll think the community will accept better.

[Delgarde]

They didn't piss all over Wayland. Yeah they made a wrong statement or two out of error but they went back and publicly redacted that.

It's more than a wrong statement - it's that the reason (officially, at least) for starting their own project instead of contributing to Wayland was because they believed Wayland to be seriously flawed. Their retraction of those criticisms undermines the entire purpose in developing Mir...

Yes, in 2010 they got behind Wayland, but at the time seemingly not believing it was a perfect solution, if I read Mark's statements from back then correctly. But by 2012 they wanted to do their own thing for whatever reason. Things change.

On the contrary - they were the first group to publicly promote Wayland, Mark grandly declaring that within a year or so, Ubuntu would be moving to it. But as you say, things change.

[jayrulez]

Idiots!...

[ninez]

That's only part of the picture. The other part is that Ubuntu has reached more desktop users than all other distros combined. The same is currently happening for servers. EC2 is somewhat ahead of the curve because it is not as inert as SOHO or even enterprise sectors, but clearly shows where we are heading.

So the claim of the previous poster that Ubuntu is where commercial Linux development happens or is going to happen is plausible.

It's hard to speculate this stuff based on statistics and interpretation. For one most of these online statistics only give you numbers of machines and/or %, they tell you absolutely nothing about the economics involved (like paid services, subscriptions, etc), nor do they tell you about servers that are not outward facing / they have no idea about. ie: you may not see the 100's to 1000's of servers that XYZ company is using internally, for whatever application(s), be that virtualization, common infrastructure/networks, render farms, super computers, etc. (which HPC and render farms, RHEL seems to be quite popular... i'm not sure how many people use Ubuntu).

On the economic/commerical front, i would be curious to see Redhat or Oracle vs. Canonical incomes, in terms of paid support / services / subscriptions, annually. ~ and in those terms, the sheer number of servers, really doesn't mean anything. ~ for example, you go look at XYZ statistic about RHEL vs. Ubuntu, you will probably soon realize that most of the RHEL servers probably have paid support ~ but is the same true of Ubuntu? maybe, but i am guessing no where near to the same degree. (likewise with SLES or Oracle, they are making money hand over fist, compared to Canonical).

I remember a couple of years ago, Mark Shuttleworth claimed that people were migrating in droves from RHEL to Ubuntu - but was that really the case? (i doubt it, Canonical would be making tons of money if that were really the case, and RedHat would not be). More likely the numbers were probably misleading, not accounting for some of the examples that i have listed above and not accounting for that fact, that the numbers give no distinction of who is running those servers, nor their 'success' in terms of capital...my guess at the time, was that very few RHEL customers had switched to Ubuntu.

anyway, the point is - statistics don't always give you a clear picture, nor do they account for all factors, involved.

cheerz

[johnc]

There is really no problems with things changing but think about why those incorrect claims about Wayland was made? Those weren't just a "wrong statement or two". It was serious claims about security issues with Wayland etc and after their retraction their entire rationale for not adopting Wayland has vanished. They never talked to Wayland developers at all before announcing Mir who has been developed for an year in private and they admitted they didn't know about Wayland much. Opportunity for colloboration and a unified solution has been lost because of this. Not to mention Mir requires you to sign a canonical CLA before they will accept patches and this makes it unlikely that Mir will be anything other than a single vendor solution.

To me it sounded like the comments about Wayland input mechanisms came from ignorance. Which, yes, is disturbing that they don't have a complete handle on how Wayland works, but that could also point to one of the overarching problems. For whatever reason they weren't interested in Wayland or weren't communicating or didn't know how to get information or just didn't care. Generally these signals can be picked up if people are paying attention.

I suspect the stuff they were doing "in private" came from things learned from their Ubuntu-for-Android and other touch research, but they didn't want to come forward with anything publicly until they knew they could commit to something on their own. A few weeks ago someone at Canonical mentioned that they were considering their own display server solution. I think we all expected something in-house and Google-like for their mobile OS. That they're extending it to the desktop as well is surprising, and where the real potential for failure is IMO.

[ninez]

I suspect the stuff they were doing "in private" came from things learned from their Ubuntu-for-Android and other touch research, but they didn't want to come forward with anything publicly until they knew they could commit to something on their own. A few weeks ago someone at Canonical mentioned that they were considering their own display server solution. I think we all expected something in-house and Google-like for their mobile OS. That they're extending it to the desktop as well is surprising, and where the real potential for failure is IMO.

I'm not sure that is really why they were doing things 'in private'. I think this was a preemptive sort of plan. For example, Smspillaz (ex-canonical / compiz-dev) very recently posted an article called "delivering free software to the masses", which had this interesting bit;

Trade secrets and competition can’t also be transparent and open: One of the most effective ways to compete with someone is to take them by surprise. Do lots of R&D to figure out what their shortcomings are, come up with something and then right when they don’t expect it, boom, sweep the market from underneath their feet. This is how every company ever is successful. If you want to reach a mass market and unseat the people who already have the lion’s share of the pie, this is how you do it. But then how do you involve your community? Well, you don’t.

I could be wrong, but this sort of sounds like what Canonical was attempting to do with Mir. (minus, the R&D part, since obviously, Canonical made a bunch of remarks about Wayland that were simply incorrect / FUD ~ but they are aiming for Android drivers/compatibility and Wayland is lacking that, even if attempts have been made in the past). I kinda feel like this is what Canonical intended to do, but it seems this approach could easily backfire on them, being as the people they are trying to 'sweep the market from' is the larger FOSS/linux community, which they so heavily rely on...

but obviously, this is speculation

article: http://smspillaz.wordpress.com/2013/...to-the-masses/

[garegin]

when ubuntu ships with mir i would purposely boycott it and not use ubuntu.

[johnc]

I'm not sure that is really why they were doing things 'in private'. I think this was a preemptive sort of plan. For example, Smspillaz (ex-canonical / compiz-dev) very recently posted an article called "delivering free software to the masses", which had this interesting bit;

I could be wrong, but this sort of sounds like what Canonical was attempting to do with Mir. (minus, the R&D part, since obviously, Canonical made a bunch of remarks about Wayland that were simply incorrect / FUD ~ but they are aiming for Android drivers/compatibility and Wayland is lacking that, even if attempts have been made in the past). I kinda feel like this is what Canonical intended to do, but it seems this approach could easily backfire on them, being as the people they are trying to 'sweep the market from' is the larger FOSS/linux community, which they so heavily rely on...

but obviously, this is speculation

article: http://smspillaz.wordpress.com/2013/...to-the-masses/

Definitely.

But Sam's article also made another important point that I think explains the motivation for an in-house project: depending on the community doesn't work for a business with business-oriented goals. The decisions have to be in-house and much of the work and direction has to be in-house. The writing was on the wall when GNOME went batsh!t crazy. To not have control over your final product is a scary proposition for a business to accept. And it can't accept it.

Though Sam also paints a somewhat depressing view of what can be realistically accomplished by the open source "community" and talks about the struggles of spreading FOSS. Frankly, I just don't get it... and it seems like he's come to accept some of the realistic limitations as well. But what is the deal with FOSS that makes it such an ethos or almost religion for many? And don't get me wrong, I love religion... but I go to a church for my religious experiences. FOSS is an extremely poor substitute for a holistic belief system. I, personally, only really believe in "open source" in the sense that I think code should be visible to verify no invasions of privacy and the like. I don't have any particular fondness to one open source license over another.

But otherwise, yeah... community-driven projects = failure.

[smitty3268]

He still doesn?t own KWin. He should be supporting the best solutions, and if that's Mir, do it. If he plays politics then he should be booted by the community.

KDE has a lot to gain with Ubuntu's move into QT.

Martin sort of does own KWin. That's his job, as the official maintainer. And I think the community has been pretty unanimous in it's panning of Mir. If someone wants to fork it, they are welcome to do so, that's one of the benefits of open source.

Anyway, you are sort of missing the point. Ubuntu is still going to ship an X server, for backwards compatibility. KWin already supports that, so it's not like they are going to stop working on Ubuntu all of a sudden. It will just keep working in X, while other distros move on to Wayland.

[GreatEmerald]

Nonsense. In the mind of the consumer, "what's Linux?", and "what's Ubuntu?". Seriously, the average computer user still has no idea what either of them are...

I wouldn't bet on it. Linux is not unheard of by regular users. Most have never used it, and may not realise what a distribution is, but they do hear about it.

That reminds me of one YouTuber I know, in his view Linux is a magic box that can do anything, but is difficult to set up and adapt to. Whenever he complains about something on Windows and I remark that the problem is not there on Linux, his reaction is typically "*sigh*... Of course it would be, wouldn't it."

Last time i checked, everybody is entitled to develop their own opensource solution.

Yes, they are. From what I can see, though, the negative effects of that could be the proliferation of X11 and less/slower/fragmented driver support from vendors. Though we will have to wait and see how this develops to tell if these will be an issue. If there is any development to be had, anyway, since to me the schedule for Mir sounds like a plan to build Rome in a day.