Talk:Mir (software)

From Wikipedia, the free encyclopedia
Jump to: navigation, search
          This article is of interest to the following WikiProjects:
WikiProject Linux (Rated C-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Linux, a collaborative effort to improve the coverage of Linux on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
 
WikiProject Free Software / Software / Computing  (Rated C-class, Mid-importance)
WikiProject icon This article is within the scope of WikiProject Free Software, a collaborative effort to improve the coverage of free software on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by WikiProject Software (marked as Mid-importance).
Taskforce icon
This article is supported by WikiProject Computing (marked as Mid-importance).
 

When did Canonical decide to switch to Mir?[edit]

I'm looking for a solid reference. I've seen a few people say six months ago, but without solid references. The launchpad project and first source commit were created on 2012-06-20 (nine months ago), but that is not solid evidence that they decided to switch at that time: https://launchpad.net/mir http://bazaar.launchpad.net/~mir-team/mir/trunk/changes/1 Trying real hard to stay NPOV —Darxus (talk) 18:17, 14 March 2013 (UTC)

  • They have only announced plans to move to it. Likely first use will only be to host x on top for pretty transitions. In March (after the April release) they will merge in Unity Next and begin developing it apparently. Gnepets (talk) 09:20, 26 March 2013 (UTC)

At the time of writing, mir and xmir remains an option on the desktop. Mir is in use by default on Ubuntu Touch images since 10 October 2013. https://lists.launchpad.net/ubuntu-phone/msg04555.html Kdub432 (talk) 23:43, 12 November 2013 (UTC)

This article does not conform to Wikipedia NPOV standards[edit]

Under "Controversy", only points of view hostile to the project are quoted. That's not controversy, it's adverse advocacy. Sayitclearly (talk) 09:28, 17 August 2013 (UTC)

Many projects not directly affiliated with Canonical are very skeptic about Mir and do not plan to support Mir and criticize Canonical for how they approach Mir – even some projects loosely affiliated do (ie. Ubuntu derivatives using another DE than Unity).
AFAIK Xbmc is the only project unaffiliated with Canonical that supports Mir optionally. I planed to add Mir support by Xbmc to the article but I hadn't the time yet to do this properly (I don't add infos in a sloppy way to articles).
I am not aware of any Linux distribution that's not an Ubuntu derivative to even consider packaging Mir at all.
As far as I can see the Controvery section covers facts and if a project is very controversial, it's WP's duty to tell so.
If my knowledge about that topic is wrong or at least limited: Please enlighten me and everybody else.
Is the criticism about Mir's licensing not valid? Which DE projects aside from Canonical's own Unity 8 plan to adopt Mir? Is there any GPU vendor that announced to support Mir? --KAMiKAZOW (talk) 10:51, 17 August 2013 (UTC)
I agree with User:KAMiKAZOW, if there are developers or reviewers who are supportive then this can certainly be added, I just haven't found any in print so far. That may certainly change once Mir is fielded in a stable Ubuntu release as reviewers will make comments on how it works or not at that point. Right now the lack of balance in criticism in the article reflects the lack of support out there for the project at this point in time and so, while unbalanced, is accurate. Keep in mind Unity (user interface) went through a similar process, with early reviews almost all negative and then, as it improved with each release, better and better reviews were added. - Ahunt (talk) 12:43, 17 August 2013 (UTC)
But Unity is no display server, its "just" a DE which is not so critical as display server.Mir has no defined spec and its currently unknown how other DE developers could use it.--Thaodan (talk) 01:49, 20 August 2013 (UTC)
What has the genre of this software anything to do with it whether it gets positive or negative feedback? As it stands now, reception is not immensely positive. Once Mir gets officially released with an Ubuntu version and reviewers have something to say about it, we'll add it. --KAMiKAZOW (talk) 10:19, 22 August 2013 (UTC)
The problem is not the software itself its the spec and than the software, because the software needs to support another display server.--Thaodan (talk) 10:44, 22 August 2013 (UTC)
And so far only Xbmc announced support for Mir, as I already wrote. I don't get your point wrt this article. Do you propose anything concrete for it? --KAMiKAZOW (talk) 11:00, 22 August 2013 (UTC)
Why did Canonical claim they want Mir and not Wayland? This goes unmentioned, so it's hard to tell whether the controversy is justified - do they have sound technical reasons? 79.182.17.73 (talk) 22:42, 8 September 2013 (UTC)
That is explained here, but really needs to be summarized and added to the article. - Ahunt (talk) 01:16, 9 September 2013 (UTC)
The main arguments were input stack (claim was proven as untrue) and server allocated buffers (claimed to be a requirement for ARM hardware). I've added both last night. --KAMiKAZOW (talk) 13:26, 9 September 2013 (UTC)
The ref cited above indicates that the reasons were, as Canonical says: "we are aiming for a more extensible input event handling that takes future developments like 3D input devices (e.g. Leap Motion) into account" and "With respect to mobile use-cases, we think that the handling of input methods should be reflected in the display server protocol, too. As another example, we consider the shell integration parts of the protocol as privileged and we'd rather avoid having any sort of shell behavior defined in the client facing protocol" as well as "we decided to go for the following architecture w.r.t. to protocol-integration: A protocol-agnostic inner core that is extremely well-defined, well-tested and portable. An outer-shell together with a frontend-firewall that allow us to port our display server to arbitrary graphics stacks and bind it to multiple protocols." So there is a bit more to it than that. This should probably be added into the article. - Ahunt (talk) 14:21, 9 September 2013 (UTC)

NPOV, a fair chance for mir[edit]

I User:ScotXW have done some work on a couple of the articles here, and though I "like" Wayland (display server protocol), I do not mind giving Mir the same chance. What Canonical could have or should have done is not as important, as the

  1. design of the software architecture behind the Mir display server protocol,
  2. its implementation (libmir-client, libmir-server, etc),
  3. the license of these implementations,

This LWN.net-article: http://lwn.net/Articles/541336/ from March 6, 2013 displays the stand-off between Mir and Wayland quite nicely. I would like to know very much more about Canonical's multi-touch and gesture input stack (released with Ubuntu 10.04). Instead of spreading more half-truths about Wayland and Mir, let's instead look at (and document since the developers are obviously too lazy to do so) at the technical stuff. I already created this File:Wayland display server protocol.svg (and I hope you all agree it is somewhat better then this) and I would not mind creating ones for Mir and SurfaceFlinger as well. ATM, I do not have the necessary comprehension to do so. Politics aside, the implementation of mir is free and open-source software, that means it deserves a fair chance! ScotXW (talk) 12:58, 4 September 2013 (UTC)


Wayland is not Build on EGL and utilizes "MOST" of the infrastructure originally developed for Wayland[edit]

Wayland is not Build on EGL who in the hell is doing the these Wiki's and Mir utilizes "MOST"(over 90% of it) of the infrastructure originally developed for Wayland not Some of it this Wiki is off this need fixed — Preceding unsigned comment added by Wtfisthisbullshitnewbs (talkcontribs) 16:18, 3 September 2013 (UTC)

Care to elaborate, dear WTF is this Bullshit Newbs? --KAMiKAZOW (talk) 16:27, 3 September 2013 (UTC)
When you look at the driver situation for Linux kernel (and others) quite nicely described here: Free and open-source device drivers: graphics, you should realize it could be better. Instead of GLX Wayland (and many many others as well) will use EGL. So EGL is very important. When you read a bit about Wayland, you will notice, that Wayland cannot use plain EGL, but requires modifications/extenions to it! Look here: File:Wayland display server protocol.svg or upstream. Those have been implemented in libWayland-EGL and have already been accepted into Mesa and are already present in Debian: http://packages.debian.org/sid/amd64/libegl1-mesa-drivers/filelist
The other crucial part is the INPUT. At the moment (in Linux) this is handled via evdev, and I do not know where to put Canonical's open input framework for multi-touch. Is this simply an extension to evdev or is it an alternative or what? As long as it is (and remains) free and open-source and (not exclusively) related to the Linux kernel, I am interested in it. So should anybody interested in Ubuntu and other Linux-based operating systems. ScotXW (talk) 13:13, 4 September 2013 (UTC)

Canonical's Open Input Framework and multi-touch input stack[edit]

This article here: http://lwn.net/Articles/541336/ mentions Canonical's Open Input Framework and also their multi-touch input stack. Since Mir relys on it for the INPUT as it relys on EGL for the OUPUT, I would like to read more about it in this article (and with enough material create distinct articles, like there is one for evdev.) Anybody else interested? ScotXW (talk) 13:21, 4 September 2013 (UTC)

I found an elder article: LWN.net 2012-05-22 A uTouch architecture introduction Does anybody know what they mean by "window server"? A window manager or a display manager? ScotXW (talk) 20:49, 6 September 2013 (UTC)

Display server protocol?[edit]

Since when does Mir specify a protocol? --KAMiKAZOW (talk) 00:17, 4 September 2013 (UTC)

PS https://wiki.ubuntu.com/Mir/Spec states:

  • A protocol-agnostic inner core that is extremely well-defined, well-tested and portable.
  • An outer-shell together with a frontend-firewall that allow us to port our display server to arbitrary graphics stacks and bind it to multiple protocols.

So the inner core is “protocol-agnostic” and the outer shell can be bound to multiple protocols. Therefore display server protocols may be implemented using Mir technology but Mir itself if no protocol. --KAMiKAZOW (talk) 11:22, 4 September 2013 (UTC)

Naming - Mir (display server) or Mir (display server protocol)[edit]

Mir is reported to actually be the name of the display server written by canonical. But when looking at the names (of the implementation of the "nameless" display server protocol) libmir-server and libmir-client and also libwayland-server and libwayland-client, I simply conclude that the display server protocol is also called mir. If canonical would like to call it bob, canonical should officially do so. I think it is possible, that there will be more implementations, even commercial ones, so I'd rather name this article Mir (display server protocol) analog to Wayland (display server protocol). This is my argument, you can additionally have a look at my contributions regarding this. I welcome anybody's logic arguments and appreciate constructive work (as opposed to specious arguments and (clever) vandalism) ScotXW (talk) 12:58, 4 September 2013 (UTC)

What ever your forbidden original research concluded, the available and official source material (“Mir is a next generation display server”) contradicts that.
Considering that Mir’s source code is covered by Canonical’s CLA, the possibility is certainly there that others license the source code for commercial implementations. That, however, is not the point. Your unjustified claim that Mir is a protocol is.
I get why in the light of Wayland you think that Mir is a protocol and your work was well meaning, but sadly it is incorrect. --KAMiKAZOW (talk) 13:11, 4 September 2013 (UTC)
This http://lwn.net/Articles/541336/ LWN.net article, that was referenced and removed by your ham-fisted edits. ScotXW (talk) 13:23, 4 September 2013 (UTC)
And where does it say that Mir is just a protocol?
You also wrote: '''Mir''' is a computer display server [[communications protocol|protocol]]<ref>{{cite web |url=https://wiki.ubuntu.com/Mir/Spec?action=show&redirect=MirSpec |title=Specifications of the Mir display server protocol |publisher=[[Canonical Ltd.]]}}</ref>
You mean well. I get that. However no original research means no original research. No source says that Mir is the name for a protocol. Follow WP’s rules, please. --KAMiKAZOW (talk) 13:30, 4 September 2013 (UTC)
I see that you are currently active editing various articles here, yet you refuse to answer why your original research, which is forbidden on Wikipedia should stay here.
Mir is no protocol and you have not posted a single reference that says that otherwise. No, the LWN reference does neither. --KAMiKAZOW (talk) 10:25, 5 September 2013 (UTC)
I have reviewed the refs and most are pretty clear, like this Canonical one that while Wayland is a protocol, Mir is, as they say "protocol-agnostic" and is not a protocol. Canonical is pretty clear in calling is a "display server". None of the refs cited say that Mir is a protocol. While I can see it makes things nicely symmetrical to call both Wayland and Mir protocols, that seem to be not the case. - Ahunt (talk) 13:12, 5 September 2013 (UTC)
Well, as I have already written I concluded from the analogues naming and also from the fact, that Canonical was going to replace X with Wayland (both protocols...) and then abandoned Wayland. Well if they abandoned Wayland, then they obviously created their own display server protocol. Hmm, because a display server talks to its clients via a display server protocol... If this is too much for you to fathom, hint, read http://lwn.net/Articles/541336/. Canonical (or just you two?) don't want to official talk much about their own display server protocol that stays in direct concurrency to Wayland, but rather play around by a) not giving the protocol a name (but it does exist...) and also stating the the Mir display server could support Wayland as well. I do not understand the statement agnostic. It has to support a protocol to talk to its clients, and the implementation is called: libmir-server respectively libmir-client. This libraries do not implement Wayland, and not X. But the Mir (display server protocol) (or rather the Bob (display server protocol). ScotXW (talk) 20:43, 6 September 2013 (UTC)
Concluding something by yourself is original research and not allowed here.
Fact is, your original research is the only thing that calls Mir a protocol. No reliable document ever called Mir that way. --KAMiKAZOW (talk) 23:53, 6 September 2013 (UTC)
I agree that the policies by which we are allowed to build articles, namely WP:OR and WP:SYNTHESIS, prohibits adding the kind of conclusions that you have come to. Just because you sell your car and buy a new vehicle, it doesn't follow that you bought another car. We are required to go with what reliable sources state and none that you have presented have stated that Mir is a protocol, in fact the "protocol agnostic" statement seems to indicate that it is exactly not a protocol. - Ahunt (talk) 00:44, 7 September 2013 (UTC)
In addition to that I also think that Scot really lacks comprehension what a protocol actually is. Just the lead of the WP article explains that a protocol must be well-defined. Without that it's just some programs talking to each other via IPC with calls that are constantly prone to change for whatever reason Canonical thinks is necessary.
The facts are: Random IPC calls to not constitute a protocol, Canonical does not refer to Mir as a protocol, and even http://lwn.net/Articles/541336/ only uses the word protocol when referring to Wayland.
I think we can all agree that Scot meant well and had absolutely no intentions to spread false info but I cannot see how we can keep his recent protocol additions in the article. --KAMiKAZOW (talk) 10:50, 7 September 2013 (UTC)
Why Not Wayland / Weston?, especially: "A protocol-agnostic inner core", this means the stuff that does the compositing will be protocol agnostic. The display server as a whole of course cannot be protocol agnostic. I think you two should have an appearance in a cabaret ScotXW (talk) 11:32, 7 September 2013 (UTC)
User:ScotXW: regarding your comment "I think you two should have an appearance in a cabaret", please see WP:NPA. Also you haven't made your case yet, as per WP:V, we need at least one reliable reference that says that Mir is a protocol. - Ahunt (talk) 12:24, 7 September 2013 (UTC)
I’ll restore the older article version and I state this waring right away: If you, ScotXW, continue to re-add your original research and synthesis and continue to insult us, I will not hesitate to report you.
I gave you the benefit of the doubt but with your insult it’s now gone.
As someone who’s active on WP for just 3 weeks, you may want to consider taking advise from long time WP contributors with ~10 years of experience each. --KAMiKAZOW (talk) 13:23, 7 September 2013 (UTC)
There was a really good performance by some CCC-Members concerning "Tschunk". You should definitely watch it and learn whom not to piss off. ScotXW (talk) 19:14, 8 September 2013 (UTC)
I cite the article: "In March 2013, Mir was announced by Canonical Ltd. as the replacement display server for the X.Org Server in Ubuntu.[6] Previously, in 2010, it was announced that Wayland would be used.[23] There were several posts made in objection or clarification, by people leading other similar or affected projects." Why does the sentence compare mir, which is claimed to be only another display server and not introduce a new display server protocol as well, with Wayland, which is a display server protocol only? I would not waste my time to insult you! It is you, who are bullshiting around here. ScotXW (talk) 18:52, 19 October 2013 (UTC)
I think we can final and definitively put this one to bed. This very authoritative article explains that while Wayland is a protocol Mir is not. - Ahunt (talk) 01:29, 24 October 2013 (UTC)
Mir is a display server, yes, but Cannonical additionally wrote libmir-server and libmir-client. And pretty much everybody criticized them for this protocol, and not for their display server. Nobody would care about another display server, there are a couple. But now they do not want to know about this protocol, and you bullshit here around. AFAIK they won't use the protocol for the moment, and Mir will concentrate on X11 and Wayland, however, Cannonical keeps its option to add additional protocols to Mir. Which is all fine. Just do not deny, the Cannonical did write libmir-server and libmir-client, the implementation of a protocol as in concurrence to Wayland. ScotXW (talk) 20:33, 5 March 2014 (UTC)
BTW, the Controversy section of the article of course refers to the Mir display server protocol and not to the display server/software. Support for the Mir display server protocol was added to GDK: https://git.gnome.org/browse/gtk+/commit/?id=0f26b26f1d93f002ccd5fe6b6fd636172bfbb24b User:ScotXWt@lk 10:51, 28 March 2015 (UTC)

Right to relicense[edit]

Longtime Linux kernel developer Matthew Garrett criticized choice of licensing for Canonical's software projects, particularly Mir. Mir is licensed under GPLv3 – "an odd one" for "GPLv3-hostile markets" – but contributors are required to sign an agreement that "grants Canonical the right to relicense your contribution under their choice of license. This means that, despite not being the sole copyright holder, Canonical are free to relicense your code under a proprietary license". He concludes that this creates asymmetry where "you end up with a situation that looks awfully like Canonical wanting to squash competition by making it impossible for anyone else to sell modified versions of Canonical's software in the same market".[27]

To the best of my knowledge this is a very common practice for open source projects. I believe most, if not all big OS projects, do the same. Here an Example for MariaDB: https://mariadb.com/kb/en/mca/ which is considered a very good citizen in the FOSS world.

More: http://en.wikipedia.org/wiki/Contributor_License_Agreement

I don't understand why this is presented as an issue in the Ubuntu case, or indeed why this personal opinion of Mr. Garrett is part of the Wikipedia presentation. — Preceding unsigned comment added by 62.218.164.126 (talk) 13:23, 19 September 2013 (UTC)

No, it’s a very uncommon practice to give a single party the exclusive rights to relicense free software under a proprietary license. A similar CLA was one of the factors OpenOffice.org was forked as LibreOffice.
However, this is not the point. The reason why it’s included is that A) Matthew Garrett is an influential voice within the FOSS community (if you feel that other media picking up his criticism should be added, I’d do it) and B) the competing solutions – X.Org Server, Wayland, and Android SurfaceFlinger – do not use GPLv3+CLA licensing. Therefore Mir’s licensing is a unique feature and one that draws criticism. --KAMiKAZOW (talk) 00:56, 20 September 2013 (UTC)
The point is, that Canonical's mir is bashed pretty hard. E.g. this article: http://lwn.net/Articles/541336/ depicts the larges picture and critiques both, the Wayland and the mir developers for given reasons. It does not bash. Some sever comprehension problems around Wayland and what it actually is are still present in the heads of many users, and I think the same is true for mir. This article comprises quite some strong critique without making it crystal clear to the reader what mir actually is. My attempts to point out, that mir does come with its very own display server protocol, that e.g. in contrary to Wayland draws things server side instead of client side, was deleted.
AFAIK OpenOffice.org was being developed under CLA for a long time! It wasn't the CLA but the change of ownership (from Sun to Oracle) that ultimately cased the fork.
A better argument for pointing out the CLA so harshly could be, that the display server is a key component (and, as you mentioned) that the alternatives X.Org Server and Weston/KWin/Mutter are not developed under CLA.
Before we delve into the motivations of Intel to invest into Direct Rendering Infrastructure or Wayland and Canonical into Mir, uTouch, Upstart, or LightDM, we should thoroughly look at the technical advantages respectively the disadvantages of the software packages respectively protocols we would like to talk about. ScotXW (talk) 09:00, 4 October 2013 (UTC)

Next step: find more reliable sources[edit]

This article has too many citations of unreliable sources such as blogs and wikis. This is wrong. I propose:

  1. We don't allow any additional information without reliable sources. If anyone adds information to this article without a reliable source, we will quickly try to find a good source for it, and if we fail, then we revert it.
  2. Gradually and slowly, as time permits, we should replace each unreliable citation with a better one.

If you agree that this is important, I will put this goal on a new Wikipedia:To-do_list -- Jorge (talk) 16:53, 3 November 2013 (UTC)

This is already covered at WP:RS and WP:V and is uniform across Wikipedia. - Ahunt (talk) 16:56, 3 November 2013 (UTC)
Well, for this article I am more concerned about the reliability of sources than with any other problem. Anyway, even if you think it is redundant to state this on a Wikipedia:To-do_list, do you at least agree with proposals 1 and 2 above? -- Jorge (talk) 18:55, 3 November 2013 (UTC)
Well of course, but then we already disagree as to which sources are reliable ones. Keep in mind that blogs by established experts and non-public wikis are acceptable sources under Wikipedia policy. - Ahunt (talk) 00:39, 4 November 2013 (UTC)
WP:SPS does not clearly recommend self-published sources, even by experts. It only says they may be acceptable. And it cautions that they should be used with care. I have not really analyzed this article yet, but based on the sheer number of blogs and wikis that it cites, I fear that many are inadequate. Remember: even if the author is an expert, he is likely to write much more carelessly on his blog than if he was publishing an article. I prefer references to books, or at least high-quality publications such as Ars Technica and maybe LWN.net. -- Jorge (talk) 01:27, 4 November 2013 (UTC)
Policy is policy. If sources comply then they can be used, but let's discuss before any are discarded. - Ahunt (talk) 01:36, 4 November 2013 (UTC)
I understand that policy is policy, but I subtly disagree with your interpretation of policy. I am not a senior Wikipedian, though. I would like to hear expert opinion. I believe the place for that is the reliable sources noticeboard. It is a pity that only one other user replied. Anyway don't worry, I do not plan to remove content without due discussion. -- Jorge (talk) 01:52, 4 November 2013 (UTC)
Well what that response on the RS noticeboard means, as per WP:SILENCE, is that there is no consensus to avoid those two sources. So, as you note, we should discuss here. - Ahunt (talk) 01:54, 4 November 2013 (UTC)

Mir 0.16.0 released but article outdated[edit]

Hello, according to the article it is still at 0.8.x could it be updated to 0.16.0 as stated in softpedia's article here http://news.softpedia.com/news/ubuntu-s-mir-display-server-reaches-version-0-16-with-mir-on-x-refinements-more-493152.shtml DarkstarCommand (talk) 16:38, 13 October 2015 (UTC)

Fixed - Ahunt (talk) 21:25, 13 October 2015 (UTC)

Toolkits[edit]

First is better to clarify if those toolkit support mir by upstream or by tree-party/canonical/not-uptreamed patches, this will help konw what is supported and what need patch. 13:53, 22 November 2015 (UTC) — Preceding unsigned comment added by 186.78.208.79 (talk)