In the past we've had this aversion to doing double work, but what you just described is doing both [writing a brand-new transit system for server meshing while another group works on a short-term self-healing fix to the current system]?

Yes. Well, we want it to function the way it should function, which is all the time, reliable, right? Now, it might mean, because we are still going to have the issue of the streaming out and some error handling, that it might take longer for a lift to arrive in certain cases, but it will arrive.

Because it'll have that error checking, it'll have that self-healing. So the idea is that it won't just not show up. You won't just fall through it. You may wait twice as long as you were used to in those cases, but then it will show up.

Yes. Because we need to fix it right now.