I am seeing all of these cool interfaces and diagrams on are.na but I should be very careful not to fall into the trap of optimizing for them alone. They are the mechanical objects and are not the goal. The goal is completing a task.
Most people only care about the means to an end, they don't want to rotate a mechanical object in order to complete a task. Mechanical metaphors are extra noise.
A friend said:
If something much simpler does it better, then the room is just another pretty mechanical object haha.
See Tabs.
Signalling
A lot of sci fi interfaces are meant to communicate "this is a futuristic (complicated) interface" to show how complicated their technology is. While they may look cool, I doubt a lot of them will be truly useful.
A lot of these agent monitoring projects people make are mental masturbation/showing off/signalling rather than being actually useful systems. It's a way for people to show off: look at my technical prowess/look how many resources I have access to, fear me/reproduce with me.
See Human in the loop.
See exactly what is happening, or nothing
For animations or visual processes, you either want to see exactly what is happening, or see nothing at all. (Everything should be deliberate.)
Can I have a visual for commands? Is there a right balance between that and just plain text execution? I want to see what's happening. (Or I want to see nothing.)
Open questions
I am thinking of things as too 2-Dimensional.
Hmm, do I want the interface to feel analogue, or extremely digital? Hyper sci-fi poet engineer, or like you are tuning physical dials and interacting with a real object.