Beryl, where do we want to go?
by cyberorg,
Tuesday, February 13th, 2007 @ 2:29 pm
Here are jotting down some of the ideas about the role that Beryl should be looking to play in future of Linux Desktop development.
Leave the core to Compiz project to develop and maintain.
I don’t believe core is the identity of Beryl or any major work by the project has gone into it to say we cannot give up its maintenance. Compiz project maintaining the core and Beryl developers helping improve it would serve several purposes that I will list down shortly.
Compiz core should be treated as infrastructure like x.org that we are using to meet our aims. Obvious question that would arise out of this is what is the identity of Beryl if we don’t duplicate core work?
I suggest redefining ourselves and creating another niche for Beryl than to compete and duplicate efforts of Compiz project. There are more important things we can put our efforts into and make Linux Desktop pleasure to use for everybody.
Some time back some developers floated idea of creating another Desktop Environment around composited WM, I think we can aim for something close to it, but not quite full DE.
Same as GNU applications and Linux kernel makes up GNU/Linux OS, we could aim for being something like GNU. Beryl should focus on the following to accomplish that goal and define its place.
1. Create repository of all sorts of plugins, usability as well as eyecandy ones. Let them be known as Beryl-plugins with dependency on Compiz-core. Let’s not just stop at that.
2. Brings together various projects such as kiba-dock, gnome-dock, Awn, Screenlets etc creating a project with such components that has compositing as fundamental requirement.
3. There are some fantastic ideas here on Gnome 3 and KDE 4 goal pages, we can help achieve those.
These are just personal ideas, please don’t take it as official Beryl road map, I just think we should move on from where we started.
I am sure many of you would have other ideas of what we can/should do, do share (no flame wars please ;)).
Have a lot of fun!
: cedric, February 13, 2007 @ 4:09 pm
KDE4 will never use compiz or beryl, it already have it’s on opengl composition manager…
: felipe, February 13, 2007 @ 4:34 pm
Thank you!
That would be a very brave thing to do and could imply some great advance in the next generation 3D Free Desktop
: scott, February 13, 2007 @ 5:51 pm
How about something akin to Apple’s forthcoming Core Animation API? I’m working right now on an IM client and I was thinking how cool it would be if I could have the ole horizontal “iPod swipe” animation when the login widget switches to the contact list. Right now the UI just changes when I swap one widget for the other, but if I could pass both Gtk widgets and an animation param to a changeWithAnimation function, Gtk could render both widgets, Beryl could take the textures and then do the animation. A new frontier for bling: in-app UI effects.
: DanieleDM, February 13, 2007 @ 6:02 pm
Hi, i think it’s good idea, i always believe that forks are not the right way of work in open source projects, and i really think that the beryl project could be something similar to http://www.gnome-look.org that permits to everyone to have the eyecandy plugin that better fits his desires!
: Bobbyjoe, February 13, 2007 @ 9:03 pm
This is the best solution for beryl in my opinion, it always has been the best place for it but unfortunatly certain people have always insisted that they somehow own beryl and they are the ones who will do something radical (or whatever)
Even though you are not a programmer, you can see the real dilema they face. Personally I think they will take a punt with beryl and hope that someday they come up with something that people cannot live without but DavidR refuses to include in compiz. Good luck with that one
Why do we never see quinn blogging about these issues?
: John, February 13, 2007 @ 10:38 pm
I could not agree more. I left a very long comment on Kristians post suggesting simiar things.
It takes a very big man to do what you have done, and to question the decisions made, to admit some mistakes. I commend you for that and wish you well that the other beryl developers will take onboard what you have said
John
: Agony, February 14, 2007 @ 12:04 am
With that I must agree. Beryl and Compiz seem to chase the same goal and even go down the same path to achieve it. Merge the two projects into one big project. It would certainly be nice to have a DE that was built around OGL accelerated desktops.
: iznogood, February 14, 2007 @ 11:34 am
This is the best thing coming from a beryl person. Why not merge the 2 communities back?? Right now the core is at a pretty good state and there are proposals of funcitonality implemented in separate libraries, some of them already exist. This way if people want to add functionality they should not need to hack the core but use other libraries and hack on them. The compiz community would need all the help it can get because after input transformation + a few more things there would be a great need for effects and usability enhansments on the desktop environments and not just dock-bars or theme engines. We should not try to create a new DE but provide the code and patches, even experimental ones, that people can take and use them on the existing DE. This way we can accelerate the adoption of new effects on these DE’s before they are merged in their repos.
What i propose is very simple: use compiz community-site for compiz related stuff (plugins - themes etc) and provide another site-community for all extra stuff i said before, with direct links to and from compiz
What do you say??
: jigish.gohil, February 14, 2007 @ 12:23 pm
This musings are in no way suggesting reversing the fork, what I would like is the the beryl-project be much much more than compiz by another name.
Ideas like creating Animation/Effects/OGL APIs, utilizing compiz core in a way that ‘pushes the envelope’ and doing stuff other than what is being done/will be done on compiz shortly is the central theme.
Beryl developers can choose to do whatever, including ‘improve’ the core, but after thorough introspection, after all I have seen days of work gone wasted, like someone pointed out blur/blurfx, aquamarine, heliodor, many settings frontends. Just chasing our own tail.
One of the thing that bugs me is, why are we waiting for Keith to finish input transformation, why didn’t we start working on it like David did. If that peice of technology is so critical, none of beryl developers have any suggestions/contribution to make in that direction? That is just one example. Lets help get our infrastructure fixed.
Hey guys just treat this as a friendly bashing, it is meant to get our bearings right and head somewhere. We have achieved quite a lot, let us now consolidate our achievements(0.2) and build even better beryl-project.
: Bobbyjoe, February 14, 2007 @ 4:15 pm
I am sorry to say but the reason that nobody did input redirection first is the simple fact that you so not have the skills and experience necessary to do it. Sorry but its a fact of life, most of us couldn’t add input redirection even if we wanted to.
: Jeff Schroeder, February 18, 2007 @ 12:29 am
Re-inventing the wheel is by far the largest problem in the open source community as a whole. Using the same core as compiz and building beryl ontop of that would likely help *BOTH* projects.
Merging back together likely will never happen due to differing directions, but this is a very good idea. Cyberorg, it takes a very big man to swallow pride and write a post like this… kudos to you.
: gilir, February 18, 2007 @ 6:51 pm
With the futur work of David Reveman on compiz core and X, I think your vision is the best for all, beryl and compiz. Please try to continue on this way