Lack of tutorials and guides...

[Yong Bakos]

Hi Inhabitant,

On Apr 30, 2016, at 1:37 PM, Cestarian Inhabitant <cestarian at gmail.com> wrote:
> 
> I've been wanting for a while now to get started with developing my
> own Wayland Compositor/Window Manager/Shell. And I honestly can't find
> much in the way of "how tos" or "getting started with wayland
> development" or well, anything.

I am basically in the same boat as you. Short story: make positive
contributions to the project.


> The only "official" tutorial or educational material seems to be the
> weston project.
> 
> The problem with the weston project is that it is HUGE and clunky, it
> might be ok as reference materials for someone who is already well
> along his way developing a wayland shell but needs some features that
> the weston shell has and he doesn't, other than that it is a useless
> (and might I add way too poorly commented) material that can never
> serve as anyone's starting point unless that someone has an unhealthy
> amount of patience.
> 
> This simply won't do. There may be some enthusiast or hobbyist created
> smaller compositors (swc), usually very incomplete and often including
> buggy features, that people are often trying to use as a starting
> point, but all in all it seems like people often either progress very
> slow or just give up on developing their wayland compositorss (I mean
> it has to speak volumes that the only projects that are seriously
> making any sort of progress on the wayland front are Gnome and KDE,
> and despite the size of their development teams they are facing major
> hurdles (although that may be to be expected for porting such big
> projects)) there's also of course the (also slowly progressing)
> enlightenment project. These projects are all respected and have
> sizable and active development teams beforehand (we aren't seeing many
> smaller projects that are perhaps going in a direction similar to
> openbox or xmonad succeeding very much, they are there, yes, very
> many, yes, but none of them is 1.0, none of them is usage ready)

I respectfully disagree on all points above. Have you written and
published any significant software, or contributed to large open
source projects? I'm not trolling you - I simply encourage you to
be more like Gandhi or Bill Murray. :)


> None of these are good starting points either, it's not because
> they're impossible to learn from, it's more because they weren't
> intended that way.
> 
> I believe that this may be holding back the progress of wayland quite
> a lot, and slowing down the rate at which people who might want to use
> it are transitioning to it from X11, especially on the developer side.
> 
> I would love it if someone would take it upon themselves to build such
> an educational project which is meant to serve as a starting point for
> developers who want to develop a wayland compositor of their own, or a
> wayland compatible application (that may be using it's own GUI
> framework (e.g. not Qt or GTK or EFL or SDL...))

Wouldn't it be great? :)
I'm personally working on this as I educate myself about this part of
the Linux stack. I should have some significant work published by the
end of summer 2016.


> It would most definitely accelerate the adoption of wayland, increase
> wayland projects, and increase the amount of people who may start
> doing experimental projects on wayland.
> 
> If no one is up for the task, I personally would be willing to do
> this, but I can't because I don't understand wayland at all at this
> point. If someone could teach me the knowledge needed to create the
> most basic, simple wayland compositor possible (a desktop shell that
> can be successfully executed and run from a console, and is capable of
> opening windows; it doesn't need to have decorations or anything to
> begin with), or at least create and share such code, code that is easy
> to understand and well commented with as few lines and few project
> files as is realistically possible for such a compositor, I would be
> very grateful and after testing and understanding it (playing with it)
> I would personally create my own version of such a project along with
> tutorial material (guides) to show other people how to get started
> from an empty project file to a functional and usable wayland shell.
> Hell, I might even make a video tutorial.

You'll need to dig. You can ask me questions on IRC, and I think you'll
find that the entire community is very professional and helpful. Just
be sure to a) be humble (these folks have been working on this for
_years_ and b) ask productive, intelligent questions only after you've
tried finding the answers on your own (use people's time efficiently).


> So please, someone, explain wayland to me. (The ideal programming
> languages for this would be either C or C++ (highly compatible, very
> widely used and very widely recognized by all programmers), graphical
> code can be whatever language (Preferably GLES/EGL for maximum
> compatibility, and easier than vulkan as a starting point)) so that I
> can explain it to everybody else.
> 
> Am I asking too much?

Respectfully, yes. :)
We're going to have to do this ourselves. I really want the 
documentation to be great. Any quality contributions are certainly
welcome.

yong


> Or will somebody help me try to ignite a wave of
> wayland development hobbyists so that Linux can again become as
> diverse and cool as it used to be back in the day with endless choices
> between desktop environments?