|
From: Ethan M. <merritt@u.washington.edu> - 2010-08-25 20:27:40
|
On Monday 23 August 2010 01:35:08 pm Ethan Merritt wrote: > That's not to say there is no path forward, of course. Just that it > would require re-thinking the API and creating new code for a whole lot > more than just one new terminal driver. Before plunging in to any > particular 3D terminal driver, I'd like to see some discussion about > how to revise the API and the flow through the core plotting routines. > There would need to be a clean alternative path through the code that > avoids calculating a 2D projection, thereby losing the full 3D > coordinates. More thoughts: If we can work out a reasonable code path for passing 3D coordinates to the terminal drivers, this would also be of great use for the HTML canvas terminal. Currently it manages to support most mouse/keyboard interactions for 2D plots, but none at all for 3D plots. It would be nice to also support rotation of 3D plots, although hidden surface removal is probably too much to ask. Ethan |