Suggest for what? As file format, qb isn’t too bad, although I’m not sure if the engine actually uses the pre-encoded alpha byte (which could be used to improve rendering/mesh generation I guess?). Most of the header could be removed and the compression enabled, although even that doesn’t work too nicely (the gain from enabling encoding/qb’s compression isn’t as much as one would hope for).
So if we’re not using the alpha byte, file size could be reduced by 1/4th by omitting it, which would be a major gain. Enabling qb’s encoding thingy might help too but I haven’t really been convinced from the short tests I’ve conducted back then.
The engine likely just has an abstract voxel model (something like a three-dimensional array of color information) and a function that loads a qb into one of these structs. It would be possible to load other things too; there’s nothing Stonehearth currently uses that would be exclusive to .qb or Qubicle Constructor. The things I’ve produced for my own stuff have all been made with Minecraft/Panicle/my own stuff, which uses .qb as exporting format, but isn’t using the Constructor and all in all just sets default values for most of the stuff.
As I’ve said, when we get down to it, qb is really just a bunch of RGBA values. We could even go so far as to say that when palettes get introduced, each model could be limited to 255 colors, therefore reducing filesize to 1/4th of the current one easily - but that would require custom editors/compilers to bake the models, which would make things more… difficult.
As far as editors go, I have to say that I don’t like Qubicle because it’s so counter intuitive to use. What I want in a voxel editor, basically, is Minecraft’s creative mode plus a few tools maybe. Add voxels, remove voxels, recolor voxels - that’s it. But then again, I’m really not a reference when it comes to visual stuff.