Announcement

Collapse
No announcement yet.

Mir Developer Pleads The Case "Why Mir"

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • lowflyer
    replied
    Originally posted by AlanGriffiths View Post

    I didn't make the decision, but you presuppose that everyone knew Wayland would succeed in building an ecosystem around it. (It is even /possible/ that the existence of Mir was key to that success. We cannot know.)

    I've had a good five years working with some great developers and line managers, doing some work I found rewarding. I'm cool with that.

    Yes, things could have been done differently. But things could have turned out differently. Hindsight is great.
    Alan, thanks for answering in person. It means a lot to me.

    Leave a comment:


  • DebianXFCE Jr
    replied
    Originally posted by emblemparade View Post
    But even for those enduring product, during the process a large part of coding involves ... throwing code away. Either because your initial approach was wrong, or because it failed a code review, broke a test, or someone came up with something better suggestion, or there's a general refactoring to a different architectural style.
    Damn true.

    Leave a comment:


  • smitty3268
    replied
    Originally posted by carewolf View Post

    Yes, Wayland is actively developed, but Wayland developers have been complaining no one was using their code as far back as 8 years ago. They basically build a road, but no facilities and then wondered why no one moved in. The mantra of it not being the job of Wayland to do Y, was just pointless, all of the tasks that needed to be done, needed to be done, and if Wayland wouldn't do them, something else would have to. Today we are slowly getting all the infrastructure around Wayland that was needed, but it has taken a long long time, and that is mostly the Wayland developers fault.
    I don't think Wayland devs have complained about that. Do you have any citations?

    I think random internet commenters on the Phoronix forums have, but that's completely different.

    The devs have always known that switching away from X was going to be a long, daunting task.

    Leave a comment:


  • nomadewolf
    replied
    Originally posted by TheBlackCat View Post

    It was just a few months away.
    Yes. That's why i don't agree with the creation of MIR, nor i support it, nor i want it to succeed.
    Not until they bring a valid justification, which has yet to come...

    Leave a comment:


  • TheBlackCat
    replied
    Originally posted by carewolf View Post
    Yes, Wayland is actively developed, but Wayland developers have been complaining no one was using their code as far back as 8 years ago.
    Citation needed. Although the project started in 2008, the first Wayland release was in 2012.

    Originally posted by carewolf View Post
    They basically build a road, but no facilities and then wondered why no one moved in. The mantra of it not being the job of Wayland to do Y, was just pointless, all of the tasks that needed to be done, needed to be done, and if Wayland wouldn't do them, something else would have to. Today we are slowly getting all the infrastructure around Wayland that was needed, but it has taken a long long time, and that is mostly the Wayland developers fault.
    No, the plan has been very clear from the beginning. First, the Wayland protocol and libraries would be developed and frozen. Then, and only then, would DE developers begin the difficult task of porting decades of X11-specific code to Wayland. That is difficult not because of limitations of Wayland, but rather because none of this software was written with the expectation that they would ever use anything but X11. That makes the porting difficult no matter what architecture they are porting to.

    The explicit goal from day 1, which everyone involved agreed on, was that Wayland would be there to help, but not get in the way of, the compositors and window managers. That, in practice, is what X11 tries to do today, but fails because there is so much legacy stuff. You are portraying this as a flaw, when instead it is Wayland's major benefit, and the reason why all these projects are using it to begin with. Why should Wayland duplicate all the stuff the window managers already have to do themselves?

    Leave a comment:


  • TheBlackCat
    replied
    Originally posted by AlanGriffiths View Post
    I didn't make the decision, but you presuppose that everyone knew Wayland would succeed in building an ecosystem around it.
    Everyone and their third aunt had committed to using Wayland. Every major stakeholder had a significant role in its development. The ecosystem didn't need to be built, it was there from the very beginning.

    Originally posted by AlanGriffiths View Post
    (It is even /possible/ that the existence of Mir was key to that success. We cannot know.)
    No, sorry, there is absolutely no reason to think this is the case. There was no increase in the pace of development of anything remotely related to Wayland after the Mir announcement. At best, it had zero impact on Wayland development. At worst (and much more likely) is it took resources away from Wayland development.

    Originally posted by AlanGriffiths View Post
    Yes, things could have been done differently. But things could have turned out differently. Hindsight is great.
    I think it has been clear from day one to everyone besides Canonical and its fans that this was the only possible outcome. Even if Canonical had the expertise to do something like Mir, which their early statements showed they didn't, it is simply infeasible in practice to develop and manage a complete display framework on their own. Everyone outside of Canonical and its fans have been saying this from the beginning. And it happened exactly as predicted.

    The only way this outcome could have been at all different is that if Canonical went with Wayland, then Unity may have survived. Since Wayland has always been several years ahead of Mir in development (at least), then Unity developers wouldn't have had to wait as long.

    Leave a comment:


  • TheBlackCat
    replied
    Originally posted by nomadewolf View Post

    I know. Read my post. I was talking in past tense. At the time MIR was started, Wayland wasn't (past tense) ready. That's all i said.
    It was just a few months away.

    Leave a comment:


  • AlanGriffiths
    replied
    Originally posted by lowflyer View Post
    To me it seems to be as simple as Alan Griffiths is now paying for the sins of Canonical. Poor guy indeed. But he should have known better. The facts have been on the table from day one of the Mir adventure.

    If Canonical would have invested that money into Wayland, these features would also have progressed faster and more complete.
    I didn't make the decision, but you presuppose that everyone knew Wayland would succeed in building an ecosystem around it. (It is even /possible/ that the existence of Mir was key to that success. We cannot know.)

    I've had a good five years working with some great developers and line managers, doing some work I found rewarding. I'm cool with that.

    Originally posted by lowflyer View Post
    Alan confirms that Mir wasn't "in the state" either. Again, Alan could have implemented that for Wayland instead, making it "into the state".

    I can relate to Alan's hurt feelings. But from my shoulders, he will not get more than a cold shrug: "You should have known better". That's the very essence of software development. I'd rather say: "Alan, get over it quickly - we need you for Wayland development."
    Yes, things could have been done differently. But things could have turned out differently. Hindsight is great.

    Leave a comment:


  • nomadewolf
    replied
    Originally posted by ua=42 View Post

    Wayland is finished. All that is left is clients for Wayland (Gnome: Done, KDE: InProgress, etc, etc)
    I know. Read my post. I was talking in past tense. At the time MIR was started, Wayland wasn't (past tense) ready. That's all i said.

    Leave a comment:


  • carewolf
    replied
    Originally posted by pininety View Post

    WTF are you talking about? Wayland support is actively developed and not a road leading nowhere (in contrast to Mir). Weston has the ability to do full screen recordings out of the box, so if you need that, good for you, just press Super+R and you get a screen recording without any third party needed.
    If you are not using Weston, write a feature request for your compositor!

    What you do not seem to grasp is the problems you get with third party screen recording software and why it is hard to define a standard way to do this without exposing your system to all the security problems we had with X and got right off with wayland. And at the moment, there is apparently to little interest in this area to solve this problem right now or then (or people have no clue yet how to solve it but I think it is the former).

    Most of the problems you see in wayland is due to the tighter security it supplies. Doing things right takes some more time and afford then just giving all information and accesses to all clients and hope non is malicious.
    Yes, Wayland is actively developed, but Wayland developers have been complaining no one was using their code as far back as 8 years ago. They basically build a road, but no facilities and then wondered why no one moved in. The mantra of it not being the job of Wayland to do Y, was just pointless, all of the tasks that needed to be done, needed to be done, and if Wayland wouldn't do them, something else would have to. Today we are slowly getting all the infrastructure around Wayland that was needed, but it has taken a long long time, and that is mostly the Wayland developers fault.
    Last edited by carewolf; 13 April 2017, 04:45 AM.

    Leave a comment:

Working...
X