Ah. I get it. You read four points at a time, then twiddle them around. 
Yes. I guess this would work. 
  
But I still think it would be slow. Every pixel on the part of the 
screen that is to be rotated has to be POINTed and PSET. On SCREENs 11 
and 12, this area can be as large as 480 x 480 pixels, so more than 
220,000 POINTs and PSETs would have to be done, even if the original 
image consists only of a dot on an otherwise blank screen! My routine, 
on the other hand, is able to skip over blank areas very quickly. If 
the screen is largely blank, which is often the case in realistic 
situations, it's very fast. 
  
                                 dow 
--- Platinum Xpress/Win/WINServer v3.0pr5