[Bf-committers] Blender and Collada support. Does it matter?

[Toni Alatalo]

On Aug 19, 2011, at 4:49 AM, Juan Linietsky wrote:
>  I work close with companies during all the development process, and that
> includes artists. I've always used Collada, first with the now discontinued
> FCollada and now OpenCollada. Collada support for 3DSMax and Maya has always

Our position is the opposite -- we've always used Blender for our games and other realtime gfx projects, targetting mostly Ogre nowadays but sometimes also other engines, but have never before used Collada really :)

But it's changing now, are already depending on Collada exports as use that with the (current) WebGL engine of choice (GLGE). So far haven't had problems with it, actually - get scenes with the geom, textures, materials and skeletal animations from Blender 2.5 to GLGE ok with the builtin Collada support on both sides. In one case even used Collada to get a customer's 2.5 scene to 2.4 for Ogre exporting, when the 2.5 Ogre exporter was lacking still :)

> works, but more complex stuff such as baking IK or constraints is not
> supported and has never been. So It's impossible to use Blender for very
> simple tasks such as animating a character, blend shapes, or export an
> object that follows a curve, thus alienating a large amount of potential

Ok, we haven't gotten there with it yet, thanks for info.

>  Collada support was supposedly going to be completed in this GSOC, but it
(..)
> features such as animating light colors. I mean, it's as if work on Collada
> is not even taken seriously by the community.

I didn't follow that GSoC effort, but i think we should remember that it's just a summer job to introduce a student to open source etc. Certainly those projects are taken seriously and produce cool stuff, but still are limited in scope etc.

Did you follow it then, did you comment that for you e.g. blend shapes would be more relevant than animating light colours? If not, you were the community not being serious! :p

>  I know that Blender is volunteer work, that volunteers mostly work in the

Not all of it is.

> motivating, and I guess the Blender Foundation does not have the resources
> to cost the development of it. But at this point I think it's weird that the

Also BI has been responsible for challenging movie projects, and Collada has been irrelevant there.

But BF and BI are not all of the community, others are free to set their own priorities and goals and dev projects etc. Certainly they want good Collada etc. too, but are also limited like you said. I think BF actually at some point contracted Nathan to work on getting the GSoC stuff over and improved or so.

> against their will to use Blender and this doesn't feel any good. I even
> considered getting donations from companies to add proper Collada support to
> Blender, but after seeing how that worked for the Illusoft exporter, which
> was always broken, unfinished and never really integrated as core, it sounds
> like making clients waste money.

That is where I think your logic goes wrong.

Don't know anything about that Illusoft exporter, but certainly there must be a way to organize funded Collada (and whatever) dev so that there is no waste of money.

>  So, to attempt being more constructuve, shouldn't it be possible for the
> Blender Foundation to work together with Khronos on this?  (read: get
> funding)  Khronos seems to be doing a great job developing and promiting
> OpenGL, GLES and WebGL and OpenCollada already has paid developers. So why
> not extending support to Blender?

I don't think we *necessarily* need Khronos for this -- they already did their part, the OpenCollada lib, which Blender is using. 

Given you, and to some extent also us and our partners here, and certainly many around the world, have the need, and some have the money too etc., we could just do it.

One way would be to fund BI if they want to organize the dev. Other option is for someone else to organize, contract some dev and collect the money and manage etc. Sounds like you could perhaps do it?

Perhaps the idea to talk with Khronos is good, I think Ton talks with them anyways, but I'm afraid that route might be slower. We really shouldn't think that doing funded on Blender is somehow impossible, I'm sure there are lots of examples to the contrary.

> Juan Linietsky

~Toni, Playsign Ltd.