(1) More ascetically pleasing things like ship paint jobs, ability to change your hud and weapon colours, space legs and atmospheric landings (more on those two in a sec), more ships, new outfits and accessories for holo-me
I think that customization is the spice of life. Wherever a game allows players to be unique in how they look and what they project, players WILL dive straight into that. A game that does not allow advanced, what I call Ego Enhancement Attribution (EEA) and that, as such, being part of a larger immersion scheme is not of this time. Incorporating such mechanism should be considered core game mechanical and regarded, for what it costs to implement, as a type of overhead - unavoidable costs that arise out of creating a marketable product.
For the same reason I deem it necessary to be able to walk around your boat and interact with it. This is no longer a matter of computing power. Or limitations of a GPU. I have felt 'being the boat' from the beginning of the game. And unless you define the keys in the ever-loathsome options menu to switch to eternal camera's and gawk at your Holo-Me created avatar, usually representing your self-image or the girl you want to marry (if she only existed and loved gamers (which is probably a myth - and even if she did, how likely would she create YOU in
her Hole-Me)) you will not often look at yourself as you play.
I remember my fascination with myself playing one of the Duke Nuke'Em games where you could see yourself in a mirror. Unreal had this as well, or maybe the first UT. I forgot. Anyway, the time of 'being the boat' should surely be abolished in this day and age. 20 years it wasn't that important, you see. Because you would
play a game whereas with today's' technologies combined with ICT you
are part of the game and its extended community. We no longer play games, we are the game. We see how when ED was under development some people climbed aboard with such vigor and enthusiasm that they became entwined with FD. People such as Kerrash and Psykokow who created all sort of fan-based follies

Bless them for it too!
So to me, if you create a game with the notion in mind to ever expand it where community building is part of the whole extended game structure - that is what ICT
does, create communities - you cannot really do without advanced avatar creation and editing as part of EEA, being a three dimensional character in a game world and being able to walk around any assets you collect. I believe in the future it will be increasingly normal to walk around and inside your ship, interact with it, I suppose, as you would already do in Star Citizen, simply because developers will recognize the need for the go to interface with their preferred escapism.
It will not be considered a 'game' if the experience of
being there is not satisfied.
In that sense, what you write about atmospheric landings looking more real is an immersion argument. And immersion of our ego in a game world is a function of duality: if you see clouds whiz about you and parts of your boat glowing from entry, it is this eternal environment that outlines the ego. That is to say, you feel that
you are there to contrast the perception of the game world. You, as a player can only be immersed if there is a background against to which you exist. I doubt most developers truly understand this, after all, psychology is not usually their forte.
So your question about 'how often' one would walk around the ship is not really a question if developers make the experience interactive. That is to say, you click on a button on the hull, out comes a keypad on which you enter your pin code so the door opens. In the cargo bay you point at some chain, use the mousewheel to 'scroll' a hook down to a cargo canister, scroll up to lift it and then by some game interface mechanic move it around to a spot where you drop the canister. The only issue here is, too much realism kills the game. because people game to have an easy life in contrast to their normal real life employment and the drag that that can be. This is where Star Citizen gets it wrong, by showing you how your avatar can pick up a box. How amazing. That is not gaming, that is chores.
If walking around the boat is just to feel proud of your new toy you just bought, then it is an expensive gimmick to implement. If it is part of the game experience, it is worth it. Is FD not promising to make mining more a proper engagement? It should be more than lasering an asteroid, they say. I have talked about this on here years ago. It never seemed to be complex to me, to allow a game mechanic to have you scan an object, depending on the scanner module and its abilities, to detect its content, then to have to move your ship to target the deposits just right to get maximum yield. I would allow a player to make it a mini-game: the better you position the laser, the higher the yield. Others less so inclined will still get good results on some deposit without playing the percentages, but the yield will be less, yet still worthwhile. That would probable me. It is then that the game allows for precision but concedes extensive demands on the player, see?
The same applies to your remarks on atmospheric landings. It is not
just eye candy, it is about immersion. But secondly on worlds with these circumstances landing should be gameplay, that is, your success at landing depends on some flight skill to reach the surface.
I would like detailed plants. Procedural programming allows for diverse environments. You can check Infinity: The Quest for Earth to see a method of implementation. I haven't looked at it for a while, but I have been involved with it for three seconds when they were looking for mission material, before they decided to put the game on the backburner and release some sort of combat demo first. And that brings me to your points about missions.
I wrote missions that are modular, in that parts can be exchanged between missions. The goal here was to avoid repetition and within a type of mission, sameness. Say you have an escort mission, to escort another boat to some destination. I imagined a group of clowns on a theater tour. Yes, that is right, a merry band of clowns, who, and this is where you could modulate the mission, stole a bottle of rare whiskey from their last employer, who wants it back and sends people after this group, or, wants his daughter back that fell in love with one of the clowns.
You can combine all sort of modular sub-targets or goals into chained missions and give the player choices as to whether or not help the father, steal the whiskey for yourself or go off with the girl...and depending on whatever you choose, you will be rewarded as you sell the whiskey, sell the girl to a slaver or massacre the clowns (my favorite).
So thinking about underlying mission structures is about non-linearity. Unfortunately, developers fail at thinking about what a 'mission;' is within the broader game world context. That is why we see linear missions with only a few possible outcomes. I don't think in terms of 'missions', mission to me are stories. The stories can be fun - who doesn't wanna massacre a ship full of clowns - but always variable and as long as you remain within the believably, you should be okay.
In principle missions are in future games hopefully, part of Your Story. In that there is no 'mission board' apart from them being simply stations or opportunity altering offers that a trader or Wandering Mage might find in their respective game worlds. Your gaming IS a long chained mission where because of the modular approach, you can, like in many games, pause to pursue, skip, abandon or decline any sub-part of it. In any more or less defined mission phase there will be entry points where the game will offer random logical deviations, such as the mechanic of selling the rare whiskey, drinking it, selling it etc. So chains may be completed insofar as the whiskey is concerned, but it will trigger a logical response from the owner. A chain is then never truly closed. It just means a contract might be put on you.
A game will forever be interesting because of its randomness and ability to surprise. As long as good imaginative writers provide fun and interesting stories, that can be cut into logical modules and be interchangeable without becoming unbelievable.