Why do so few Summer of Code projects make it into the trunk..And can they be revived

Looking at another post (and an older one of mine) about the Particle Nodes build… I got to wondering…

What ever became of some of the really cool projects that have been started, and, if not finished, much progress made, particularly through the Google Summer(s) of Code?

In particular, this: http://wiki.blender.org/index.php/User:Jwilkins/GSoC2011/Google_Summer_of_Code_2011_Preliminary_Documentation

That is from years ago, and there is a now obsolete branch build (Onion) on GraphicAll… but, it is such a useful project. One of the main functions was to project the sculpt brush ONTO the geometry, much like ZBrush does, in order to see where your next stroke will be (without guessing and undo-ing constantly). This is absolutely a feature that should have been trunked long ago and should be so commonplace that we don’t even notice it.

But…someone put in the work to make it happen. Where is it? What happened?

Can these things be revived? It is maddening sometimes, to see all these branched builds with really cool features, like Particle Nodes or Sculpt Brush Projection…but to have no access to them except as novelties in some antiquated version of Blender.

I understand that alot of the features may not be FINISHED, but why are they not included as add-ons? Even if they aren’t written in Python and go further into the guts of the program, is there really no way to keep them on-board as Experimental functions? Molecular add-on goes beyond regular python and functions… so why do we lose all of these other really really cool functions?

One of the main functions was to project the sculpt brush ONTO the geometry, much like ZBrush does
Except ZBrush doesn’t project the sculpt brush ONTO the geometry as in the image on that page.

Aren’t a lot of those things in that page now in the paint/sculpt tools ?

In general, Blender’s code-base isn’t structured in a way we can partially include features.

Typically GSOC projects touch the UI, file format… internal API’s. So they are either included or not…
(Addon’s use a limited API, so we can’t simply convert various features into add-ons … in most cases)

Also, I don’t think its great to push half working features into Blender anyway.
If nobody is prepared to finish the project, its not a great sign going forward - since anything thats included in a release should be supported, get bugfixes, be documented… etc.

Wouldn’t it be nice if it was, though? :slight_smile:

How hard would implementing something like that be? I imagine it would take refactoring of the whole code base.

Nicholas Bishop had a blog post a couple years back that explained it.

I can sort of understand the title of the thread because the are a lot of GSOC were the features developed never make it into Blender. But then you went and picked the one summer of code where 80%-90% of the features developed made it into trunk.

more sushilikeness

The projects are too ambitious, the code ends up crappy, nobody wants to stick around after they’ve been paid to get things into master, and existing devs don’t have time to fix the bad GSoC results.