Lightspeed

[Arian]

I recently exploded a few stars around my planetary system just to see how it would look. I noticed that the novae became visible instantly as the stars went kaboom, even if they were 50, 100 or thousands of ly away. Well, the flash was visible, the material cloud showed up only very much later when it was big enough. (btw, since that last update, nova clouds are flickering for me)

What I mean is: Normally it would take as much years as the distance in ly to the nova to see that flash. I realize that US2 would look quite different if the visuals would actually simulate lightspeed, but I was wondering how different it would look, in what way it would change.
Since the camera can move at speeds far beyond that of light, how would the vision be distorted? Stretched and squeezed objects most likely.
I was also wondering whether that could be implemented as an option. Maybe only for simulations with a locked camera position or the camera itself only moving as fast as light, but I think it could be rather interesting to watch.
On a similar topic I would like to see the maximum velocity of any object being capped at the speed of light, because that would change some simulations quite a bit. This also should be an option rather than a general setting, but it should be there. In case of an acceleration beyond the speed of light, the exceeding energy should heat up the body instead. But actually such an acceleration should not be possible to begin with.

[Jenn]

Yes, Universe Sandbox 2 would be a very different game if we treated the camera as a physical observer such as a telescope. There has been talk of a view setting to allow us to show red-shift and blue-shift, but showing events happening in observer time relative to camera position would be quite frustrating from a UI perspective. Anything you did in the game you would have to wait a delayed amount of time depending on your zoom distance to see take effect. Would you really want your view to disappear when you zoom out too fast?
There are some relativity relevant fixes that we would like to possibly add in the future:
1) It would be nice it gravity moves at light speed. That would mean that if you delete a host star it would take time for the orbits of each planet to change depending on that planet's distance from the deleted host.
2) Yes, to a simulation setting that puts a cap on speeds. If I could reasonably add in post-newtonian terms to our n-body calculations that would also mean that orbits would lose energy according to the prediction of relativity and you would see inspirals of close binaries at more appropriate rates.
3) Yes, we have also discussed having relativistically accelerating bodies radiating off energy. Universe Sandbox 2 is a newtonian gravity simulator. We try to conserve energy and momentum. If we arbitrarily apply a speed limit without accounting for accurate predictions for energy loss, than we lose more accuracy than we gain.
Full general relativistic simulation requires the simulation of the spacetime itself (by decomposing it into a 4D grid of discrete points/regions usually) which is far more computationally intensive than treating the bodies as point particles of a certain singular mass and than checking for the attraction to all the other point particles as we do now. A general relativity simulation of many bodies would not run on a consumer's desktop computer. There is however something called Post-Newtonian, and I would like to look into the computational cost it would take to more accurately show, for instance, precession of mercury or compact binary inspiral and spin interaction. It would also help enforce the speed of light as a limit in an more accurate way. This might however require a mode where you would set up your system, turn on PPN, and wait a minute or two for the initial data to calculate before letting it run on with the corrective terms. And this may simply not be a practical or sensible feature for a sandbox game.

All that said, relativistic ray tracing has been done in the past, and might be a possibility for us: http://www.spacetimetravel.org/tuebingen/tuebingen5.html
It can be nauseating, though. :-)

[Arian]

You are so dam good at this 
I considered all of that (well, not on that level of detail, but in general I did).

1) Gravity being communicated at lightspeed should actually require only slightly more calculations, but I'm aware that it would have to be updated quite frequently to retain accuracy. Maybe it would be an idea to only apply that on sims with distances greater than a lightyear, since most people will likely run "smaller" sims at timesteps that wouldn't make a big difference.

2) As long as it is optional I am all for realism. However, to really incorporate relativity, the newtonian physics would probably have to be abandoned and we would face the start of US3. A simple switch like "hyperlightspeed on/off" might just be enough for now.

3) Capping the maximum velocity at lightspeed would just be a crutch but it could work as an error correction in newtonian physics. Since this is a computer simulation based on newtonian physics rather than on relativistic terms, energy doesn't have to be conserved. It may not be correct to simply drop exceeding energy but neither is newtonian gravity. Newtonian gravity theory is missing some things, so why not have said energy "go missing" as well? You could see it as some kind of "Newton's Cosmological Constant".

As much as I would love a relativistic simulation, I am aware that it would be anything but easy to handle for the average user like me. I am pretty content with Newton because that is already something I have wanted for decades but never found in sufficient quality (on my very subjective scale). Also A relativistic simulation would make people call for superstrings and M-theory to be incorporated. We would soon have people want for 10 dimensional branes and such, which would have nothing to do with a simulation of the observable universe anymore.

BTW: I didn't have the chance to visit the “Sonderausstellung Albert Einstein” in Munich back in 2005. But I live in Stuttgart, so I'm kinda familiar with the city of Tübingen but never rode a bike that fast there. However, I had a similar view two or three times after a long evening with friends although I'm pretty sure it had nothing to do with speed 

I really appreciate your reply, since it provides a picture of how much work and considerations go into US2 without users ever noticing. It also shows that US2 is a very dynamic project (no pun intended but if you crack a smile that's fine), and it shows that you appreciate the users' thoughts on it.
Thank you.