1) V3.0 would still be despecializing the BD, only using the 35mm scan as a source for the unaltered footage - here's why:

I realize that there's more compression here than there would be in the final release but it still shows pretty clearly that since the BD was scanned from the original negative, it has more fine detail than the LPP print even though the '04 scan is outdated, so a 1080p restoration using a 35mm scan to despecialize the BD could yield similar results to what a professional restoration from maybe an IP could achieve and that's something probably still worth the effort.

2) I will definitely finish ESB and ROTJ v2 before starting any work on a (again, still just potential) v3.0 of SW.