Here are some thoughts not directly about *which* system is better, but might help you think
* I'd like to be able to quickly replace the content of a layer with one of the same-named layers in a different frame,
so maybe, if you rightclick on the thumbnail, you could get a list of those (with the default choice being the frame before this one, and the list being arranged by time proximity). I think this begins to address the 'too many layers available' problem.
* Failing to name your layers meaningfully is a recipe for failure when the animation is complex.
It's probably therefore better to optimize for mostly-meaningful names (for example, treat the source layer name as much more important.. say you drag a layer from frame 2 onto the same named layer in the current frame -- then the position and other parameters of the layer should reset to match the source layer.)
* Using SHA1 hashing to distinguish unique layer content can also help in cutting down the list of source layers.
instead of listing duplicate entries
[001] foo
[002] foo
[004] foo
[006] foo
, list a single entry [001,~2,~4, ~6] foo and take the content from frame 001
This can also prevent you from keeping duplicate layers in memory
* you might want to keep an eye on how Grafx2 ends up doing it: Layers are currently being worked on (basic functionality with a maximum of 8 layers currently). Once they're done it will open the way for animation.
One of the angles they seem to be taking to simplify things: all layers are the same size. I think this is a good simplification that reduces the amount of commands needed and does not impact significantly on performance for typical pixel works.
http://code.google.com/p/grafx2/issues/detail?id=7&colspec=ID%20Type%20Status%20Priority%20Milestone%20Component%20Owner%20Summary%20Level* I expect any pixel app to keep all frames in memory for quick flipping -- there should never be any wait loading from disk (or if it does, it should be because virtual memory is getting used -- this is what it's for, the OS does it better than any immature application can.)
* I'd like to see 'ghost' layers in the layer list..
Like, say your previous frame has 4 layers that aren't in this one.. they should show up in red as a single line.
layers from the next frame should show up in green as a single line.
clicking on their lines should allow you to 'realize' a copy of them in the current frame.
This does not address all reuse situations, but I believe it addresses quite a lot.
* I think (complimenting the above) if you could quickly browse through thumbnails+layer names of layers on nearby frames and copy one complete with name into the current frame, that would address the remainder of the problem of easy reuse of nearby frames.
* finally, I say: 'new frame -> copy of all layers from the frame you were previously on (and no others)'
Hope that helps
