diff --git a/implementations/uvm-ncc/xerblin.c b/implementations/uvm-ncc/xerblin.c
index c76e77b..5fe2386 100644
--- a/implementations/uvm-ncc/xerblin.c
+++ b/implementations/uvm-ncc/xerblin.c
@@ -10,8 +10,11 @@ size_t FRAME_HEIGHT = 800;
 // RGBA pixels: 800 * 600
 u32 frame_buffer[1024_000];
 
+int wid;
 
-void anim_callback()
+
+void
+anim_callback()
 {
 	// Clear the screen
 	memset(frame_buffer, 0, sizeof(frame_buffer));
@@ -25,13 +28,41 @@ void anim_callback()
 			*pix_ptr = (alpha << 24) | (red << 16) | (green << 8) | blue;
 		}
 	}
-	window_draw_frame(0, frame_buffer);
-	time_delay_cb(100, anim_callback);
+	window_draw_frame(wid, frame_buffer);
+	// This rate of refresh of the whole screen caused ~60% CPU use.  :(
+	// time_delay_cb(100, anim_callback);
 }
 
-void main()
+
+//void
+//keydown(u64 window_id, u16 keycode)
+//{
+//	window_draw_frame(window_id, frame_buffer);
+//}
+
+
+void
+mousemove(u64 window_id, u64 new_x, u64 new_y)
 {
-	window_create(FRAME_WIDTH, FRAME_HEIGHT, "Bouncing Ball Example", 0);
+	// Redrawing the window on move seems to repair damage from
+	// dragging portion of the window offscreen then back on as
+	// you drag, without using too much CPU.
+	// No, weird, it doesn't.
+	// It did, once, but maybe that was teh X server somehow doing it for me?
+	// In any event, moving the mouse over the screen after drag damage
+	// works fine, as expected.
+	// (I just really would like to know why it stopped damaging the window
+	// in the previous try?  Ugh!)
+	window_draw_frame(window_id, frame_buffer);
+}
+
+
+void
+main()
+{
+	wid = window_create(FRAME_WIDTH, FRAME_HEIGHT, "Bouncing Ball Example", 0);
 	time_delay_cb(0, anim_callback);
+	//window_on_keydown(wid, keydown);
+	window_on_mousemove(wid, mousemove);
 	enable_event_loop();
 }