These lists do not represent the full variation I'd like to show (and some hypotheticals, like the Toroidal Planet, will likely be omitted). For instance, habitable worlds could be further subdivided by whether or not life is present, the complexity and development of that life, (if civilisation exists there) its civil and technological advancement (see also: https://en.wikipedia.org/wiki/Kardashev_scale), and whether or not life still exists there or if it is an "extinct planet" (still interesting to study, as there will be tell tale signs that a planet once supported life, perhaps even fossilised remains).
We also want stellar variety, with differently positioned and sized habitable zones, and the life present in some systems may not be indigenous to a single planet. Think "space whales", interstellar immigration and "artificial planets" (large space stations?).
To generate this variety, invoke deterministic procedural generation (#10) and probably work our way down a generation tree that begins...
Kind of star
Planetary positions (any in habitable zone?)
Nature of planet
(If habitable) kind of life
(If developed) advancement of civilisation and technology
etc.
Note that extinct worlds may be extinct for a number of reasons, either caused naturally or by the inhabitants; in some cases, life may have developed while the planet was in a habitable zone but become extinct later after changes to the parent star left the planet outside of that zone ("Heat Extinction Planet", "Frozen Extinction Planet"). Determining whether or not this is possible for a given planet will depend on the kind of star, so...
Given a kind of star... we might consider adding some weight to the later factors in the procedural generation in order to inflate the probability of finding some more interesting worlds.
When it comes to life, also remember the Drake Equation: https://en.wikipedia.org/wiki/Drake_equation Determining certain factors of this equation and applying those in how we weight our generation methods will create varied frequencies of life. We don't necessarily want it to be too abundant, but we don't want to limit variety too much by making it too rare - it's a careful balancing act that will affect gameplay and story enormously.
See...
These lists do not represent the full variation I'd like to show (and some hypotheticals, like the Toroidal Planet, will likely be omitted). For instance, habitable worlds could be further subdivided by whether or not life is present, the complexity and development of that life, (if civilisation exists there) its civil and technological advancement (see also: https://en.wikipedia.org/wiki/Kardashev_scale), and whether or not life still exists there or if it is an "extinct planet" (still interesting to study, as there will be tell tale signs that a planet once supported life, perhaps even fossilised remains).
We also want stellar variety, with differently positioned and sized habitable zones, and the life present in some systems may not be indigenous to a single planet. Think "space whales", interstellar immigration and "artificial planets" (large space stations?).
To generate this variety, invoke deterministic procedural generation (#10) and probably work our way down a generation tree that begins...
Note that extinct worlds may be extinct for a number of reasons, either caused naturally or by the inhabitants; in some cases, life may have developed while the planet was in a habitable zone but become extinct later after changes to the parent star left the planet outside of that zone ("Heat Extinction Planet", "Frozen Extinction Planet"). Determining whether or not this is possible for a given planet will depend on the kind of star, so...
Given a kind of star... we might consider adding some weight to the later factors in the procedural generation in order to inflate the probability of finding some more interesting worlds.
When it comes to life, also remember the Drake Equation: https://en.wikipedia.org/wiki/Drake_equation Determining certain factors of this equation and applying those in how we weight our generation methods will create varied frequencies of life. We don't necessarily want it to be too abundant, but we don't want to limit variety too much by making it too rare - it's a careful balancing act that will affect gameplay and story enormously.