I'm not sure why culling or other tricks would be incompatible with a 'persistent, real time, fully-physicalised-loot' game. I'm operating under the assumption they'd be mandatory and also a given. Why track the specific details of a hundred of the same kind of object when only their differences are relevant? Why show any given client more than it needs to see at that moment? Why wouldn't the same tricks that apply to things like rendering or compression work in these scenarios? All the tricks have gotten better at reducing the amount of work that needs to be done to create convincing real-time detail and the underlying hardware has simultaneously become vastly more powerful.
I can accept that it's probably not going to happen in this game, but I don't believe for a single second that it's some kind of categorical impossibility, and when a project has the resources devoted to it that this one does, anything less than a technical marvel is going to be a disappointment.
Imo, You're asking the right questions. And the outcome is of course disappointment. SC is like coitus interruptus. Or the girl friend who always got a headache like SaltyMikes wife once remarked. She had it nailed a 100%. My impression was that her hubby must now see clear, too. But sime just don't see the forest when they're looking at the trees.
They can't make the prettiest-looking single-player game ever, which is also the bestest twitch-action MMO ever, which is also the largest bespoke open world ever. They just cannot. They don't have a roadmap that gets them there, and there comes a point where telling smart people to do impossible things reaches its limits...
The fact that SC is still stuck in one solar system (of 100), on placeholder networking capped at 50 (not 1000s per location, or a single shard etc), in its 9th year of production speaks volumes on that front. And the missed deadlines along the way only amplify that message further....
Second Solar System:
Initially aimed for 2015 with the Nyx system. (Delayed by the introduction of full proc gen planets. And then the shift to bespoke planets using proc gen art tools.)
The current plan is to get a basic form of 'Static Server Meshing' (essentially just familiar Destiny style networking) for Q3/Q4 2022. But first they need to get the new version of the item persistence database in...
The bizarre tragedy of SC is: That they're actually still trying to do it all. Which sounds daft, for a 'dream game'. But if they gave ground on even one of the key areas, they could actually make a genuinely decent fist of a game, even at this point.
But they won't. Because Chris Roberts. Crazy pitches got them here. Crazy pitches will keep them here too...
Whatever we end up with is going to be writ janky by the many years of attempting 3 impossible things before breakfast. It ain't gonna be commensurate with either their grandest claims or the epic dev lavished upon it. (CR kind of specialises in such deliveries these days, so it shouldn't come as a major surprise )
On a related tangent the implementation of the canister drop limit in ED was a travesty! Being able to carry one's own sensor overwhelming sensor clutter in a stealth Clipper...that was gameplay.
On a related tangent the implementation of the canister drop limit in ED was a travesty! Being able to carry one's own sensor overwhelming sensor clutter in a stealth Clipper...that was gameplay.
Maybe it was gameplay, but it was a killer for performance, so FD did the sensible thing.
You may remember, CIG some years ago did a similar thing, ships exploding and dropping loads of parts and containers. But they acknowledged it tanked performance. It was never released.
I know why they did, but I would have preferred a more considered fix, rather than the expedient one.
What FDev takes they rarely give back, even if the 'fix' is no longer necessary. It's one of the reasons why ED has had such an inflationary set of balance issues and why we've had various graphical downgrades that persisted far longer than they should have. Terrain in Horizons 2.0 was better than in 2.1 or any version of Horizons since, despite the reduction in geometry quality being pointless after four generation of hardware improvements since it was implemented. Fire in EDO is even worse...in many places it looks comically bad, far worse than it did upon release, like something straight out of 1995...and that nerf didn't even improve performance, they were just in a panic to trim everything they could to get the game playable.
You may remember, CIG some years ago did a similar thing, ships exploding and dropping loads of parts and containers. But they acknowledged it tanked performance. It was never released.
I suspect there was no real "considered" fix. Tons of objects require lots of processing and network bandwidth. Sometimes you just have to say no to stuff. Something CIG seem to struggle with.
Better to not release it than to tease people only to retract it after the fact.
Instead they just teased it by hyping it for months before they knew they could do it, even showing videos off of it in action, only to fail to deliver. Now that is something CIG excels at!
Here is my unrequested opinion of the river video:
The light reflections on objects are lush. Like that.
The walking under water 'feature' is clip-tier terribad.
Seasonal vegetation is probably the last thing the over-burdened game needs right now, and would probably be stealth-removed 2 years after any addition.
That goes double for the proposed 'trash floats downstream' physics.
TLDR: It’s pretty. But it’s pretty flotsam. Floating along on top of a river of dreams.
Maybe all of the flotsam will form a dam? And if we stand on it we can see Pyro?
It would be...however. Remember this is a game where CR talked about fully physicalised NPC's, that would go to work, go home and sleep and return to work the next day even if there was no-one there to see them, these NPC's would be using processing power and server bandwidth even if no-one else was even on the same planet. So while yes you can use culling and tricks, I think that's something that hasn't so far been considered. I mean really in the above scenario you would just use some spare cycles to caclulate where the NPC should be at that time of the day and only render them when actual players get close enough to see them, but that's not what CR wants to do, he wants the cities to be living and breathing cities with physicalised NPC's undertaking day to day activities even when there is no one else there.