Note: This thread has been split off from a previous one here.



The tables have been released today
http://www.terathon.com/voxels/

Excellent, it's really good that he makes this publically available, especially as the voxel terrain is one of the features that sets C4 apart from other popular engines such as Unity. Presumably he hopes that it will generate some good publicity for the C4 engine.

I would certainly be interested in seeing this algorithm implemented in PolyVox, and it seems that this will be possible. I'll probably implement it myself (if no one else does) but it won't happen for a while because I'm currently working on the improved volume compression, and need to work on my game after that. I also don't have a personal need for this system as I'm curently working with Minecraft style terrain, but it would certainly be interesting to see if the two styles can be combined in a single environment (transvoxels for the terrain and Minecraft-style for the buildings?).

If anyone does experiment with it I'd like to kow what you find...

I'll be interrested in implementing these later too, after finishing a publication on something else.

It'd be great to see it in PolyVox after you've improved the volume compression: it could have a lot of applications for research projects in realistic graphics rendering.

Yay, looking forward to making use of this when it gets implemented.

I was actually considering doing exactly what you mentioned for my own game, David, with the block based system for buildings/walls on top of smooth voxel terrain.

I just had a look at the license behind the released tables, and it might be necessary to check some details with Eric when it comes to implementing this in PolyVox. In particular, the license says:

Which sounds quite reasonable, but may actually be incompatible with the zlib license which PolyVox currently uses (emphasis mine):

Technically you can use PolyVox without ever telling your end users you are doing so, though of course it's nice to receive credit. But end users of the PolyVox library aren't going to know that Eric's code is being used internally and so can't be expected to credit him.

It might sound pedantic, and maybe he is happy to simply be listed in the PolyVox credits/copyright (which might not get passed on to the end user), but it's worth checking with him when the time comes.

I'll come up with something more specific to put in the data file. It wasn't my intention to require that a credit actually be displayed in an end product or anything like that. I'd be happy if there was just a comment in the source code of any implementation using this data that indicated where the data came from. And, of course, there shouldn't be any false claims as to the source of the data or the algorithm itself.

That sounds entirely reasonable. We would of course credit you as much as possible - i.e. a mention in the code and documentation with a full reference to your paper/dissertation and website. The only thing we can't do - as Dave says - is force users of PolyVox to mention you in the credits of their project.

_________________
Matt Williams
Linux/CMake guy

Thanks for the clarification, that's what I imagined your would say. It's really just a technicality but still important to keep things straight. We'll be sure to include credit in our own implementation when the time comes.

Thank you so much for releasing the source of your algorithm. I intend to sublicense the source code as zlib license (http://en.wikipedia.org/wiki/Zlib_License), which I think does not infringe upon your terms. Is this ok?

_________________
irc://irc.freenode.net/#polyvox

I think your probably better off emailing Eric directly, I don't think he checks these forums very regularly.