Image optimization

I was thinking about a automatic image size optimizer for non power of 2 images that would cut the images into 2 or more polygons with 1 ore more texture

For example a 1280 by720 would be broken into 2 polygons with with one 1024 texture

While a 1280 by 1024 would be 3 polygons with one 1024 texture and one 512 texture

http://files.getdropbox.com/u/1027218/exmp.png

http://files.getdropbox.com/u/1027218/exmp2.png

I know it can be done manually but was wondering if it can be done as a feature(also is should be optional) especially if some don't know about it and waste memory

We talked about this; it's pretty complicated but a possibility for 2.0.

It's probably not possible with render targets, however - but some newer graphics cards support non-power-of-two textures anyway. There are other problems with splitting textures, such as slowing down rendering, and also image quality: due to floating point positioning, it's difficult to stitch the two sections together without a visible seam on the join appearing in many circumstances.