Shared publicly  - 
 
....blarambleblabla... "When lots of competitors attack a project on purely political grounds, you have to wonder what THEIR agenda is. At least we know now who belongs to the Open Source Tea Party And to put all the hue and cry into context: Mir is relevant for approximately 1% of all developers, just those who think about shell development. Every app developer will consume Mir through their toolkit. By contrast, those same outraged individuals have NIH’d just about every important piece of the stack they can get their hands on… most notably SystemD, which is hugely invasive and hardly justified." ...blablarambleblabla... -- Mark Shuttleworth

So yeah I am not entirely sure what he is saying there, but I think some "outraged inviduals" from "most notably systemd" just got called the "Open Source Tea Party" by Mark.

I really wonder though how we deserve such high praise. I haven't publically commented on my blog or anything on Ubuntu and Canonical in years. And I don't think I commented much about Mir either. Yes, I can tell you that Upstart is misdesigned, but that's a technical opinion, and heck, other than that I don't really care...

(Also, I think +Matthew Garrett probably posted the most relevant reviews of the Mir situation on his blog, but I find it cute suggesting he would torture the english language. I mean, knock yourself out complaining about my english, but heck, Matthew is Cambridge and shit...)
99
10
Cristian Rodriguez's profile photoLeandro Inácio's profile photoDavid Dušanić's profile photoÁlvaro Villalba Navarro's profile photo
29 comments
Koen Kooi
+
1
0
1
0
 
and he still can't get the caps right on SyStEmD
 
I like this quote: 'What closely to see how competitors to Canonical torture the English language in their efforts to justify how those toolkits should support Windows but not Mir.' He who throws the first stone...
 
Ehm, the authors of Mir calling systemd suffering from NIH syndrome seems to win the award for the most hilarious thing I've heard this week.

I must admit I still think systemd has a long road to maturity before it, but hey, it ships and is working most of the time. Which I can't really say about Mir (and Wayland, for that matter, too).
 
I'm sure he's correct because no apps other than window managers use any X calls right?
 
So if we are the Tea Party of Open Source now, then I want to be our Sarah Palin! Who wants to be our Ted Cruz?
 
Is the Open Source Tea Party a closed society, or can anyone join? ;-)
 
I am preparing a rally plus a flag.. who is with me ? :-D
 
"...SystemD, which is hugely invasive and hardly justified."
Hmmm, I might be mistaken, but Canonical is trying to use logind in Ubuntu, right ?
 
I have read Shuttleworth blog post completely once and the cited paragraph a couple of times and still do not quite get its sense: I think that Shuttleworth is actually attacking KDE  mentioning systemd is just to make sure that really no one is not upset.
 
+Matthew Williams Why would he? I mean, Jono works for Mark, I am pretty sure his job is primarily being a loudspeaker repeating and pushing whatever Mark says, not contradicting him or apologizing for him.

I mean, some years ago Jono called me up on phone about things I said about Canonical, but this time the roles are reversed...

Also, not everybody can say that he's among the favourite enemies of a space billionaire (even if I am not sure I deserve that praiseworthy role he chose for me). I think that makes me special, no? ;-)
 
I think it is an indication of the general quality of Mark's character that he resorts to ad hominems, calling people out of their use of english, while he butchers the english language with all his talk of first class citizens and beautiful interfaces.

The fact that there are those in open source projects that do not have english as a first language, and sometimes struggle to speak the Queen's best is no reason to criticise them. Especially when he can't spell the name of the thing he was talking about correctly, (SystemD vs systemd), which shows how little he actually knows about it.

"13.10 is a very special release for me because I think we are leading the GNU/Linux world into a very important arena, which is mobile personal computing."

"Earlier this year however, I discovered that a well-known company had taken the code - disappeared underground with it for several months, improved upon it, utilized the capability in their advertisements and demos and in the end posted the code utilizing their own source control system, detached from any state of that of the upstream project's. Even to the extent some posters around the web thought libhybris was done by that company itself."

(Taken from http://mer-project.blogspot.co.nz/2013/04/wayland-utilizing-android-gpu-drivers.html)

Yeah, leading the world, indeed.

I am trying to think of another company that was open source, but had a CLA.... and what they did with it in the end. Oh, yeah, thats right: Fork Yeah! The Rise and Development of illumos they used their CLA to take other's contributions and close it up. I am not saying they are going to do this, but I would not be surprised at all. This is the ONLY reason I can see why Mir exists, to give Canonical the ability to do this. The other possible reason is Server side buffers, and even Intel wont accept patches to enable that in their drivers.
 
+Tim JP In that case, I am already a member :-) I am simply not generous because I develop FLOSS software without CLA :P

Seriously: I am very disappointed by Mark... He should not need ad hominem attacks. He basically is doing now what he claimed other people would be doing against Canonical. (Which makes the whole thing look hypocritical, and makes any discussion based on facts impossible)
 
+Lennart Poettering  Actually.. my first reaction to reading his post was a deep anger at you personally at taking over my role as spaceman's archvillian.  I feel like you've stolen something precious from me.  I'm absolutely lost without the thought of him fuming in rage at me to buoy me up every morning and put a spring in my step.
 
What bothers me most about Mark Shuttleworth and Canonical in general is that, outwardly they try to lead you on to think that they are leading the open source community into an era of glory and success, when in fact they merely use the open source community as a stepping stool, that they've summarily crushed under their feet now that they feel they no longer need it.
 
+Lennart Poettering my job is not to act as a loud-speaker to Mark. My job is to build a fun, productive, and motivating Ubuntu community.

Sure, I generally agree with much of Mark's perspective because I think we are philosophically generally pretty aligned, but there is not a direct connection between his brain and my mouth; I do have least a little sentient consciousness evaluating things. ;-)

There have been many cases when I have disagreed with Mark about something, and I don't agree with every decision Canonical makes, but as I am sure is the same case for you and Red Hat, when this happens I generally raise those issues directly with the individuals concerned rather than on social media or blogs.
 
The difference, Lennart, is that I respect your right to spend your time and energy making something like systemd. Go ahead, open source gets better for the perspective and the competition. Thank you for doing it.

I don't like that you misrepresent Upstart in the process, I think it's unhealthy to subsume so much into pid 1, I dislike that you sledge the competition for existing, and I think you're copying all the parts of Apple's init that seem dodgy to me. But you did that before and it worked out just fine with Pulse, years later. So I don't go around trashing you for writing open source code.

At the end of the day, if the engineering leads on ubuntu want to switch to systemd that will be absolutely fine by me, because it doesn't matter one bit. But so far they tell me upstart is cleaner and simpler and preferable. And that's fine by me. Cleaner and simpler is better.

By contrast, the commentary that I find so stupid is the suggestion that it is somehow wrong for Mir to exist in the first place. It's pure hypocrisy to criticise the Mir team for making something crisp and clean that solves a real problem in a new way, when you do the same all the time.
 
+Shaun Reich  I'd actually like to know what Mir solves that SurfaceFlinger doesn't.  I mean pragmatically speaking Mir replaces SurfaceFlinger in the existing Utouch technology stack that was originally demo'd and MWC and in subsequent testable Utouch images for the DIY install onto existing phone hardware.

So instead of focusing on why exactly Canonical didnt go Wayland. It would be I think more instructive to understand the bigger picture if we knew why Canonical didn't just stick with SurfaceFlinger and build an XsurfaceFlinger extension instead of Xmir.  My understanding is the original Utouch demo used SurfaceFlinger.

Its also my naive understanding that much of the pre-supposed technical requirements for Mir, like the server/client buffer  allocation  is a direct result of needing to work as a replacement to Surfaceflinger. I surmise here that much of what Canonical is trying to accomplish is driven by there need to be fully Android compatible as a forward looking strategy so they can prepare for a future where their mobile offering can leverage the existing Android ecoystem.

But again... it begs the question... why build Mir when SurfaceFlinger already exists, works very nearly in the way they want Mir to?
 
Mark looks like a person playing Canonical and Ubuntu like a toy. Because utilizing of existing OS technologies with solid community support seems more practical for me. Several recent and past decisions of Canonical seem coming from love to write a lot of code.
 
+Mark Shuttleworth, I will give everybody who asks for it my technical opinion on Upstart, and that's my right. Other than that I don't really care about Upstart, and don't spend much thoughts about it.

systemd is a suite of things, very little of which is "subsumed" by PID 1, as you claim. And you guys know that, since you actually include at least two major components of systemd in Ubuntu by default (systemd-logind, and systemd-udevd). Also, we never "copied parts" of launchd, that's just an allegation you made up and which I will just ignore for now.

I try to keep the Upstart folks up-to-date about what we do, for example at Debconf, where we talked for an hour so, just to make sure they know what we are working on in the context of kdbus and things. We did this, because we thought it might be useful to you. Also, at conferences like the Plumbers Conference a few weeks ago in New Orleans we invited the Upstart guys to talk, since we want to make sure everybody can hear their side of the story too. All in all, we try to be fair to them, we try to keep them updated, we try to get them out of their little Canonical bubble, so that they communicate with the outside on where things are going, and they understand what other people do.

All in all I'd call that a maybe cool, and distanced, but somewhat fair relationship with them. It might have been more heated years ago, but today, it's really just a boring, maybe uninterested and certainly uneventful relationship with them.

Regarding the topic of Mir: I had to google today to figure out whether I ever said anything publicly about Mir at all. Turns out, I did, on the day it was first announced, where I made a joke about the name on Google+ and drew a comparison to some other Canonical projects, and that's it. I never fought it, I never wrote any other public line about it, I really don't care about Mir a tiny bit. I never "critisized the Mir team" as you say. Display servers are not my area of expertise, and I couldn't care less. I don't publically comment on things I have little knowledge of, because, well, I don't like looking like an idiot, you know...

So, I am really not sure what makes you so angry at me. I was kinda happy with our relationship to the Canonicals, but now you pick me out for a fight, but I am pretty sure I am the wrong person, since our relationship to the Upstart guys is actually boring, and regarding Mir I am really the wrong guy to complain about, because it's not my personal playground.

Also, let's note for a fact that you just did go around, and thrashed me, in contrast to what you write in the same comment of yours.

Anyway, with behaviour like yours, and calling me names, all the goodwill from my side that might be left (or well, has regrown in the years) is really at zero now (again). You did an awesome service to your "upstream relations".
 
+Jef Spaleta Most likely, the reason Canonical pushes Mir is pure business: they need a core component in their Ubuntu phone stack which defines their open source platform (and hence cannot be easily replaced) but which only they control. The controlling they do via licensing it GPL3, and making sure they are the sole copyright owners via CLA. That way they can license exceptions to the mobile industry who is normally allergic to GPL3, and nobody else can. It's a way how to make sure that the Ubuntu phone stack cannot really be forked and made successful with phone vendors/carriers without Canonical getting a share of it.

The combination of GPL3 with copyright assignment has been Canonical's business idea since a while, and I figure that's why they tried to make it more widely acceptable via Project Harmony.

I think it's even totally legitimate what they do with this, after all, the stuff is true Open Source in the end, but then again, it's a bit shabby and unfair to contributors, it really doesn't show any respect for the community. I personally don't want to be a part of these kinds of shenanigans... For some reasons though Mark picked me out as one of his favourite foes now though, even though I don't think I really deserve that position. What's a low rank developer at Red Hat against a Space Billionaire after all?

Not sure why Canonical doesn't simply say that GPL3+ClA is the core of their business model. Maybe in order not to piss off the community? But I am pretty sure they already did that comprehensively, so they could just be honest, and maybe win a bit of respect back.
 
+Jef Spaleta & +Lennart Poettering Canonical has never, ever, done a commercial dual-licensing deal. To describe this as "Canonical's business idea" is nonsense - perhaps that's why we don't describe it as the "core of our business model" ;)
 
+Mark Shuttleworth  I'm going to refrain from going another round with you concerning the CLA requirement on Canonical projects....

Still very interested in that contrast  between Mir and SurfaceFlinger and some insight into why sticking with SurfaceFlinger wasn't adequqate...on technical merits..even after Utouch was successfully demo'd running ontop of the SFlinger implementation. 

Honestly it looks like you cost yourself 6 months to a year by taking on the effort to implement a replacement to SurfaceFlinger after you had a mobile UI stack up and running ontop of flinger. I don't understand the merit in the expense. Even covergence, it seems to me you could have created an Xsurfaceflinger equivalent and been ahead of where you are right now towards a consumer ready offering.
 
+Mark Shuttleworth never having done something doesn't really mean that you don't want to, does it?

But if the CLA stuff is not relevant for for your business model, why don't you just give the CLA up? If you couldn't so far and won't monetize it, and it only gives you negative responses from the broader community, why have it at all?
 
... because I think it's the right model for small codebases. Dealing with inevitable changes in license (GPLv4?), for example, is much more legitimate when it can be done directly rather than polling a list and hoping for the best.

For companies like Mongo, dual-licensing is a viable model and I think the free software community should be wise about that. Yes, it means things might happen that you don't like. But it definitely also results in more innovation over time. Big companies - Intel, Google, Red Hat etc tend to beat up on that model because they would rather not negotiate with those smaller companies, but I think that's because they are not very wise about where innovation actually comes from.

In my case, since I think it would be good to have more companies like Mongo and MySQL driving innovation, and I think the dual-licensing model is the best way we know how to make them sustainable, it would be cowardly to not adopt the same CLA requirement just because it's not popular.
 
Maria is a great example of how to do things right - put IP under a community organization and follow one license. That code can never be relicensed and the community can thus be sure of trusting it. This is in stark contrast to Mir.
Add a comment...