But I've only ever really heard him talking about planetary generation, but TBH, I'm not very interested in looking at planets from up in orbit. I'm sure they'll look beautiful and varied, but I still suspect after a few dozen, you'll feel a sense of "been there done it".
It's the asteroid with strange markings, or a tunnel going into it I want to find/visit. It's the star with its atmosphere being ripped away by its twin I want so find/visit. It's the moon with geysers shooting water vapour up into its atmosphere I want find/visit. It's the jovian giant with huge lightening storms I want to find/visit![]()
I have great hopes for the potential of procedural generation in this game - David Braben and FD have been doing this sort of thing for a long time and they are certainly aware of the issues involved.
However I agree with your concerns. On the micro level it is simply beyond the capacity of current personal computers to render highly detailed and convincingly realistic terrain in real time, and that is without also considering the problem of AI injected "unique" features.
On the macro level it is simple reality that 99% of systems are going to look very similar, and after the 10th typical Type M system there is a chance that the experience will become rather dull. So what wins? Interesting gameplay by handcrafting features and injecting events or simulated "reality"?
Procedural generation can be used to achieve amazing results, but it's great strength is no so much the ability to create boundless semi-random worlds - it is the fact that they can be generated identically for every player as required. It's great weakness is that algorithms can sometimes actually produce repetition instead of variety (fractal geometry is after all a form of procedural generation).
I'm hoping for a great experience that will surprise me with its capture of reality plus interesting features and events, but I'm also prepared to face the fact that it just may not be possible to provide as rich an experience as I would like.