Source of the post
Now that we have interstellar planets wouln't interstellar comets and asteroids be an interesting addition?
I know its quite useless (since probable would make the engine slower when loading and since you woulnd't really have the tools to search for them in the first place right now). In principle this is like that, pointless. But let me convince you with some arguments.
Interstellar minor bodies exist (weonly have observed one but we know that they have to exist). According to this research paper there has to be around 1 trillion comets per cubic parsec in interstellar space. And accordign to this one there could be 10 times more density of interstellar comets. So SpaceEngine, having realism as a central idea, should have this feature one day.
The answer is simple: computing power. Handling 1 trillion objects requires octree of a depth 10, per each cubic parsec. As long as luminosity of comets is near zero, it will be traversed mostly by the intersection with a camera (ie max draw distance ~ 0, unlike stars). So CPU will select 10 nested octree nodes, in which camera currently is, and then it must do a linear search of the nearest objects in that nodes. The object count in each node will be ~10,000, so such search is not fast. For stars and galaxies octree, which have similar numbers, SE performs such search not each frame, but each 50-100 frames. This results in not instant appearance of the closer star system or a galaxy (you may notice that Magellanic clouds sometimes disappear and appear on screen after few seconds).
So we have to have a 10-level octree in each cubic parsec, there must also be additional levels to merge it with a galactic star octree, which has smallest node size of 1-10 parsec. An all this just to find a single comet which is within visibility range of the camera! (ie larger than 1 pixel). Occasional encounter with such a tiny thing in a vast space is impossible, the only way to find it is using some search tool.
Note also that octree is not designed for a moving objects, so all these comets will stay at rest.
Planemos are implemented as a zero-luminosity star systems, which are generated in the last star octree nodes, together with the dimmest brown dwarfs and neutron stars. They does not increases load very much, because number of planemos is expected to be 1-2 times the number of stars, so they just doubles count of objects in the last octree nodes (upper octree nodes contain just ~25% stars of the galaxy).
You may see, that although interstellar comets are possible to implement, but using a deterministic generator such as octree is overhead. Maybe some fully random generator can be used instead of this (aka dust "star streaks" in space games), so user will not be able visit the same comet in the next session nor sharing its coordinates. Or maybe some pseudorandom hash to make them deterministic, I dunno.
Source of the post
"They are here" argument:
Not only we now know that these objects can be crossing stellar systems (maybe I could go to Sirius and see an interstellar asteroid traversing between the planets), but also there is now an estimate of the density of objects like the one detected currently in our Solar System. 300.000 interstellar asteroids could be right now inside the orbit of V774104 (~100 AU from the Sun). These objects could be illuminated in Space Engine because they are travelling through the Solar System as any other object but with random and extreme trajectories.
This is way too beyond computational possibilities of SE. Even stationary comets would require some implementation on GPU to calculate illumination (visual magnitude), and still need octree-like structure on CPU to search the closest comet to camera to render its mesh. But we do not want a static objects in the solar system, right?
Source of the post
Historic landmark argument:
Now that we know about an interstellar visitor to the Solar System it would be awesome to have part of it's path represented in SE (maybe some thousand years in the past and future) so as to reenact the historic flyby. Also there are ideas to chase it so why not make people try it with space probes in SE?. As a planetarium we should include things like this for educational purposes.
You can easily add it to SE! SE supports hyperbolic orbits with e > 1, so no problem! The only limitation is that hyperbolic orbits rendering is limited in space (I mean the orbital path line), and that object will follow this trajectory forever, as if there are no other stars and galaxies (this must be obviously, SE is not an N-body gravity simulator, and such computations will be overhead for this purpose).