[compiz] [ANNOUNCE] Compiz feature branch compiz++

[Kristian Lyngstol]

On Wed, Dec 24, 2008 at 01:48:17PM +0100, Dennis Kasprzyk wrote:
> Hi,
> 
> I've currently pushed a new branch called "compiz++" to the freedesktop 
> repository, with some features I've been working on during last months. 
> Because most of the features also require (BIG) changes to the plugins, I've 
> decided to put them all together.

I've thought long and hard about Compiz++ and Compiz. This is a really hard
choice. I've talked around to try to get the general consensus, and I
believe the essence of what I'm about to say mirror the feelings of several
other developers. Since I can only speak for myself, I'll write "I", though
in many cases, it's not just my own thoughts that's gone into this.

Essentially, I want to go ahead and develop Compiz based on master, not
Compiz++, but I can't deny that many of the features you've added in
Compiz++ should be implemented in Compiz. I believe that a switch to C++
was premature at best, and looking at core, I believe it will put us back
years with regards to a stable release.

What I want to do, is release Compiz 0.8.0, then focus on getting Compiz
1.0 out the door. This should be a matter of cleanups, documentation and
generally boring but necessary work. The goal should be to maintain Compiz
1.x as a binary compatible branch that is easy and comfortable to work
with both for Compiz developers, plugin developers and package maintainers. 

When Compiz 1.0 is out, we can start doing what you've already done,
except I still want to do it in C, because I do not believe C++ is the
answer to all our problems. In other words, I want to move 3D rendering
into a plugin, I want to investigate reparenting and tiled textures, and so
on.

Some of the features that you've introduced could probably go into 0.9
and thereby end up in 1.0. Removal of multi-screen support _could_ be one
of those features, as could the WRAP/UNWRAP replacement. Neither of these
features really require C++, and both have very clear advantages without
introducing too many new potential bugs.

The main reason that I do not want to go ahead with Compiz++, is that it's
simply too much in one bite. I'm really saddened by the way Compiz++
evolved. If it had come at an other time, and been developed openly and in
a community that had an open minded and interactive project management, 
I have no doubt that the end result would've been different. I also
understand that much of the reason that it was kept secret was to avoid the
risk of being overwritten by other branches (of and nomad), and that's a
sad testament to the state of Compiz as a community.

My only hope now is that Compiz can move forward, and that we can gain new,
active developers without loosing any old ones. My biggest fear is that you
(Dennis), will detach yourself from the project, which would be a huge
setback for Compiz, as your ideas, your understanding of Compiz and the
environment around it and your friendship has meant much both to Compiz and
to me personally. 

And I do not believe that Compiz++ was done in vain even if it is left
unmerged. The experiences gained in it's development can benefit Compiz, if
you let it.

- Kristian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.freedesktop.org/archives/compiz/attachments/20090105/db6a8b82/attachment.pgp