[Bf-committers] Tuhopuu vs the Main Line (was Cheap Tweak Mode)

[Jonathan Merritt]

Robert Wenzlaff wrote:

>Tuhopuu is a test area.  Once the code is tested the developer(s) of that code 
>need to still be proactive and submit a proposal
>  
>
I guess something that confuses many people (myself included!) is the question of what constitutes code that must pass through Tuhopuu and what is "worthy" enough (or maybe simple enough) to be committed directly to bf-blender.  I think many people would be happy to see this outlined...

- Clearly, there is a development path for certain features that are central to future releases (eg: the new animation system), which are likely to make it directly into bf-blender.

- Then there are bug-fixes; also likely to go directly in the main line.

- Python development seems to proceed almost exclusively in the main tree, although my impression is that the Python developers are quite a cohesive and open group.  The task for Python developers also seems to be much clearer: it lies more in exposing the existing Blender functionality to Python, rather than actually deciding much of the functionality to begin with.

- Finally, there are miscellaneous features developed by Blender's guru coders (Ton et al.).

Have I missed any obvious items in this list?  I think that a list like this placed somewhere on the website might make people happier.


However, then there's a "grey area" of code which I often personally suspect will be bound for Tuhopuu when it appears on the list, but then a particular developer jumps on it and it gets pushed straight into the main tree!  ... <voice type="Kenau"> woah! </voice> :-)  It seems like favouritism when this sort of thing isn't explained more fully on the list, and it's later revealed that some private communication (email, IRC chat, or whatever...) took place to initiate things.

There also appears to be a lack of an "official" community input on what features from Tuhopuu will make it into the main tree.  Let's suppose that a developer has spent some time working on a Tuhopuu feature, but it is then rejected by Ton because it doesn't *quite* fit in with his Blender world view (I tend to agree with Ton's vision for Blender, but I can appreciate some of the arguments to the contrary :-).  How can that developer then "appeal" to the wider Blender community in an effort to assert that their feature is worthwhile?  I don't know of a completely fair way to do this, but perhaps a real voting system of some sort (either community-wide or just a '+1/-1' thing from the committers) might give new developers more confidence of receiving fair treatment?

-- 
Jonathan Merritt BE(Mech)/BSc
PhD Student - Equine Biomechanics
The University of Melbourne
Veterinary Clinical Centre, Werribee