srsa_4c
As for the capture function, I'm thinking it needs a complete implementation similar to shoot(), using action stack.
I'm not sure. That would make it blocking from script POV, like the current implementation. So the only advantage would be that it doesn't tie up physw task.
I'm not sure if you would include the shoot_half stuff in the action stack, but there are probably cases where you wouldn't want to do it every shot.
To access raw data, you'd still need a non-blocking call with something like the hooks (to prevent MenuOut from being called until the script is done with the raw data) or an explicit call.
There could be value to having both a simple, do everything blocking call ala shoot() and an option allowing more control.
To mimic the current logic of capt_seq shooting in relation to kbd_task and spytask, you could put frsp in it's own task, but that doesn't address the MenuOut issue directly.