The rumor that Collada will be dropped from Blender is circulating. See: SL Post: Blender dropping COLLADA support and Blender Lists. It is undecided whether the Blender development team will drop it or not. Everyone strive for No Panic Mode.
Don't Panic
It looks like Collada is the ideal format for use with SL. Several other pay big money modeling programs have Collada export. So, it is unlikely it is going to die out. I have a hard time believing the Blender Team will give up on it. But, those playing with the code say it is a bit of a mess and likely will be rewritten. That could mean we will see it removed and come back latter. But, even that seems unlikely to me… the ‘remove’ part that is.
It looks like Domino Marama is going to make a Python Collada exporter for Blender. For those of us that have purchased Primstar or Avastar it will be free. Or it may end up free to all. Only time will tell us what is finally decided.
I’ve heard someone at the Lab say they had some people working with Blender Org on Collada. So, may be they can or will LL chime in at Blender Org or otherwise help to keep Collada in Blender. We’ll ask but don’t expect a definitive answer. It is obviously in LL’s interest to have Collada export in Blender.
With the Deformer out and asking for a new flag in the mesh format it may be a good time to consider whether any changes to mesh are needed. Content made from mesh is and has been coming to the grid. But, it might not be too painful to make a change now. Only the Lab can make that call. They will take heat either way, but I’ll go with their call since they are the heatee.
That Blender Collada support is undecided and a new Python Plug-in is likely to be created if support is dropped and may be created in any event just to provide a better SL specific export, this may be an opportunity to get some Collada problems fixed, publish some wish list exporter features, or make any changes to the mesh format. We may hear some LL feedback at the next user group meeting.
To understand the Blender side of things read Ton’s blog post: COLLADA momentum.
To understand that this is not just open source contributor grumpiness about Collada, consider that Ton is Ton Roosendaal, Chairman Blender Foundation. His points are well made and it may sway the direction Blender takes.
What is Likely to Happen?
Reading the mailing list thread, I would say Collada WILL BE DROPPED from the main development branch of Blender. That does not mean they stop work on it. Actually the problem is no one is working on it. So, the Blender people want to get a buggy part of the code out of the main branch until someone is actively working on it. I would say they see it as an embarrassment. So, Collada Export/Import will become a user rather than Blender supported effort and plug-in, which it is and has been. But, now they will make that VERY apparent.
The result is we will add Collada as another plug-in via a separate download. It is unlikely to come with the standard Blender install.
The result we would like is a Collada export/import plug-in written in Python as suggested by Domino Marama. It might not be free. But, it may not be possible… or perhaps more accurate efficient or practical. The Blender peeps say considerable work would still have to be done on the Blender side to provide the hooks for Collada in Python to use. There really isn’t anyone to do that work.
But, Juan Linietsky reduzio at gmail dot com, has made a Collada Python exporter and is offering it to whoever wants to work with it. So, it may not be all that much work. Without being into Blender and Python code it is hard to know. But, a working program is convincing. The code is in this zip file. Gaia Clary has played with the script in Blender 2.61 and had problems.
Erwin Coumans has provided another exporter. This one is said to operate from outside Blender and reads the .blend file to get the export.
The Blender plan is to move the Collada code out of the trunk and allow Collada developers to work on the Collada branch. It looks like this will happen, but that is just my opinion.
SL is Not Alone
SL is not the only 3D modeling system that uses Collada. There are lots of users, loads. But, until we have some programmers that want to work on Collada, it does not matter how many people or games use it. Whether Linden Lab or Google (Google Earth/Sketch Up use Collada) will step up and help out is unknown.
Dropping Collada from Blender may get the problem on the radar. At that point people may step up and fix the problems.
Why Drop Collada?
It seems the Collada project is slow moving, slow to answer questions, slow to respond to bug reports, and dreams of creating the ultimate 3D exchange format… I’m overstating to make a point. The Blender development pace and the Collada development pace seem mismatched and that is creating friction. As Collada is open source it is an unfunded project, AFAIK. Perhaps a corporate sponsor could get things moving faster. Or even crowd funding. Whatever the solution it involves someone doing something or offering to pay someone to do something. Volunteers cannot be yelled at or forced to do things. One must do the work they want done or find a way to motivate or inspire someone else to do the work.