Re: [glob2-devel] Thats it, we need a plan

[Stéphane Magnenat]

Hi,

That vision is nice, but unfortunately unrealistic. Glob2's code is complex, 
because the game engine is complex, very complex. Of course, it is worst 
because it is 6 years old, and has been the big project through which both 
Nuage and I learned to program correct C++.

Now we have two people basically knowing the inner system (and even, we need 
to reread big chunk of code if we have to work on them again).  But those two 
people must harvest money to live and thus have a job. We thus have much 
fewer time that during our student period.

Honestly, managing this project, even if you consider this managment a 
failure, takes a lot of energy. But before giving up, I really would like a 
nice result. Even if not perfect, but well finished, usable, and mostly bug 
free.

So here is my proposal :
We put every feature request into wished bugs, with discussion attached, so 
that they are easily trackable. We can link bugs together to create 
dependency graphs. Then we create a roadmap in the Wiki, refering to bugs (we 
copy the bug title in the wiki page).

We split the roadmap between todo for 1.0 and todo for post1.0.
1.0 is not total freeze, but only quality and gameplay related issues will be 
accepted (no wonderfull new feature such as glob 3d). If someone disagree, it 
is always possible to submit a patch with the new feature. If it is good and 
stable, it will be integrated. In the meantime, Nuage and I try to clean as 
most things as possible while doing changes. You'll see, the new 
unit/building matcher will be much more understandable.

Steph