Well, first because I work with TIFF files in 16 bits, and I can use either RGB or RGB float on them, and on the vectorscope it makes a hell of a difference compared to 8 bits. Second because I use them as a digital intermediate for 16 mm scans. Once I have done my color corrections/ grading then I export in 8 bits.
Cheers E Now I wonder why a 10 bits 4:4:4 file is needed. If it's a matter of image quality, you're going to lose it in the internal engine, which works in 8-bit space (I tried to edit Canon 500D footage under RGBA-FLOAT and got a black screen).
