Map size

[Hellhawk666]

I'm afraid you are overrating possibilities of procedural generating algorithms.
The biggest concern with the procedurally generated content is that it almost always look very generic and after visiting a few locations it's easy to see repetitions of the same elements in various permutations. I havent seen a game or program that get rid of this problem. Also key elements such as alien races and important locations propably will be made by artists hand, otherwise they will look very generic.

Procedurally generated terrain can indeed start to look "samey" both on the micro and macro scales, but avoiding this is not a simple matter. On the micro scale you can add a huge palette of textures, models and colours, increase the complexity of the algorithms and try to achieve a balance between regular features and random placement. Rather than trying to build features up, the best systems I've seen use algorithms that attempt to model the effects of plate movement and weather to create natural looking landforms - but the rendering is intensive and I've not seen any such systems that render in real time.

The macro scale is where I see the real problem. The game should be shooting for some degree of reality, which means that exploring deep space will involve visiting system after system of very similar configuration and appearance - gas giants, frozen giants, baked deserts , airless moons, etc. Every now and then I'd hope to see a planet that was capable of supporting life, and occasionally one of those would indeed support some form of simple life. Finding highly evolved lifeforms should be very rare.

This process in itself will be repetitive and perhaps ultimately boring for some players.

 

[Mazhurg]

When you think about it, there is some level of "procedural" generation in just about anything; from trees using similar (bark - trunk - branch - leaves) textures, to people houses using the same layout, different colours. (In NA especially, walk though a newer or though a post WW2 neighbourhood.

Lots of variation, lots of samey yet we (well most of us) never get bored of it.

I guess it all depends on how complex the generation and resources are going to be. Will it be perfect? Not by a long shot but it should be intriguing enough to make one want to see more.

 

[SuperG]

I'm having a hard time wrapping my head around how "100 billion star systems" will be implemented. How will that work in terms of gameplay? Obviously there's no way that large a map can actually be used nor can that much content be created without just automatically assigning random, but no special, characteristics to each system & planet. I guess my question is what purpose will all that space serve in gameplay terms? It seems to me that once you travel a certain distance there's really nothing more to see.

DB explains it somewhere. But the key point is that random function isnt that random, it is more it must seam to be random but it is not it is even predictable. That because the random function uses a seed.

This means the random result is reproducable. This means the galacy is generated consistent the same each time. On each client. Wich means it doesnt need to be synchronised completly. Every client generates the same result. Only does parts that chance due to game events. Like planet novabomb busted. If the go for mass planet destruction. Or artificial set a star to supernova.

For gameplay well depends. How fast FTL is or jump range and fuel efficency.

That means is the whole galacy reachable in reasonable time. Is the galacy full explored in a week or month or year or decade or century or 1kY 1MY etc.

If it would be like startrek with warp drives FTL speed somewhere around 1KC to 1MC
Or farscape or BSG with fuel for 100 jumps of up to 100 LY

You could make. Even our galacy small to make it easy reachable with the use of speed that fit the size to gameplay.

It isn't mind boggling in game sense there is also overkill.

There is not much more emersion for implementing the whole earth in a game of Arma3. Then a arena schooter like cod or a infantery only milsim like america Army 3.0.

Just like BF4 maps are huge on console wenn you are on foot. But in a tank it make sense. In heli big enough in jetfighter you get clausterfobic. ARMA3 full island size is better match for jetfighter combat.
on PC BF4 is larger so heli make snse less crampy for yets.

I got the impression that they won't match the speed to make galacy a full playground. I think they limmit the speed to still get lost in space or have a more exploring in deep space experience.
That okay. But it not a must have its a dev choice and withou it such games aren bad and not incomplete because. It more a matter of game focus and with that cater for specific audience.

Where is the focus? on gamers or nongaming astronomers. Or both.

What this seams to be, something that more then a game more like a dual purpouse.

What celestia is, its not a game but it suports it but the focus is non gaming. It pure explore space on a pure astrinomical way.

Whith that ED is apperently more then a game it is merge of space game and a celestia aplication.

it isn't a necceserity but a part of the gamers just want it all.
I can do without, lot of space games without it and they are also different to each other and some overlap.
Now are there some way to do interesting things with exploring in not so astro way. Like finding artifacts derelic ships salvaging scavanger. Or reach isolated save spaces with different ruleset and ship classes. So there are a few options to bring some kind of gameplay stuf out there in deep space.

 

[[VR] Mad Mike]

it would make no sence to put it at the other side of the galaxy,

Actually I think it WOULD make sense. If it was too close, it would be a bit odd if for instance I as an explorer had been on an obscure empty planet yesterday, then I download an update with some new DLC and all of a sudden I am slap bang in the middle of the thargoid homeworlds.

BUT

I dont think this is an issue. In the 1st 1st game there were 8 (I think) separate galaxies which you needed a special drive to travel too. I think an equivalent thing could be for in a new big DLC, a worm hole (or something similar) is discovered which, with the purchase of a special jump drive, you can go through and takes you to a whole new section of space previously way to far away to get too.

 

[kilvenny]

There is an inevitability that the planets will eventually look similar, there is only so much variation you can make to colour/texture/height etc.

So there has to be some other forms of differentiation. So far we know

1. Space stations - how many/size/location to the planet
2. Faction presence/Thargoid - at war/peace etc
3. Rare items for sale
4. Missions
5. Player activity in the system influencing development of space stations
6. Mining
7. Salvage
8. Buildings

Done properly I think this will make it less samey.

Other possibilities to me.
1) Space station inside clouds so you have to find your way in with sensors/navigation.
2) Some areas allowing player generated planet development
3) Asteriods you can fly through (hide in)
4) Catacombs to fly through on planet. Space stations below ground level on planets.

etc

 

[JeffRyan]

It can't be worse than what's happened in my town in Italy after a certain chain opened up. Now every house I visit has different iterations of the same tables, chairs, book cases and cubes....

 

[Chock]

I'm afraid you are overrating possibilities of procedural generating algorithms.
The biggest concern with the procedurally generated content is that it almost always look very generic and after visiting a few locations it's easy to see repetitions of the same elements in various permutations. I havent seen a game or program that get rid of this problem. Also key elements such as alien races and important locations propably will be made by artists hand, otherwise they will look very generic.

It's like Minecraft. Every mountain, forest and river is different, but at the same time they are all pretty similar because they are all made by the same algortithms and exploration soon became boring. I agree with Exmortis, it's a big chalange for devs not to make exploration monotonous.

I'm aware that using procedural generation is not a magic bullet to create a dazzling array of brilliantly unique new things. FD's system for this will after all, be just a more complex variation of what MS did with Flight Simulator to populate places with autogen scenery. But that said, like with MSFS, it does the donkey work and allows the further addition of more hand-crafted stuff.

Hand-crafting stuff takes time, but, as I noted on another similar thread, this opens up the possibility of FD releasing an SDK and allowing third parties to create more detailed content; a system which, for MS, created an entire industry where Flight Simulator is concerned and was a breeding ground for fledgling content creators to develop into some of the industry's most creative and respected companies, just look at PMDG or FS2Crew for examples of this.

All it then needs, is for FD to have some kind of approval system for content to be included and 'Bob's Yer Uncle', suddenly FD has an army of content creators working for them, and all at no cost in either wages or time. This is really no different from what script-writers do for sci-fi TV programs; they are handed a set of 'lore' guidelines for the show, and so long as they stick within those, creative freedom becomes their playground, this being a system which has garnered some of the finest and most ingenious sci-fi we have. If you look at some of the stuff third parties have made for FS, it becomes apparent that this kind of thing can really push the boundaries of creativity to stellar levels of brilliance, and there is no reason why this system, with some minor modification, cannot be repeated by FD. And presumably, such a new revenue stream source would be welcomed by them, as it appears FD are hoping Elite Dangerous will be a game which is here for the long run and one that will evolve.

 

[Herik]

Folks...100 billion. Star systems. There seems to be some expectation that by exploring you will find continue to find cool and interesting stuff. There's not enough time left in the developers' lifetimes to even sparsely place interesting stuff among 100 billion star systems, much less players' time to find it. It seems inevitable it'll be a huge amount of nothing. Maybe that's the plan and it will mostly be there for aesthetics more than anything with a small fraction of those systems being more customized with content (e.g. Federation, Empire space etc.). If that's the case I guess that would be fine - I'm just kinda curious what the plan is.

I suppose they could do something like generate one of some pre-set number of special characteristics per 1000 systems or something - like ancient artifacts for example. Still though - at some point one would have seen all these things and exploring further may not be very compelling. Once you've seen Rich worlds, Industrial worlds, worlds with ancient artifacts, etc. etc., there's no point going further. There must be limits of course so I'm not suggesting that would necessarily be a bad thing. I really am just curious if there are any plans for all that space or if it's just there to have a realistic Milky Way and won't really be used.

Exmortis; idealy the plan is to have many more expansions (up to?) a ten year game life span.... Imagine what they could add over those years. The 'Milky Way' just isn't a one expansion deal. This gives the game a large future for new purchasers as well as the current players. I don't see any limits with this plan. Remember, we're biting the edges of the possibilities using Procedural programming. I don't mean to drift us off topic but 'This' knowledge base will grow over the years. What seems improbable now, (or impossible) will change; look at "Limit Theory" to see what may be coming down the pike.

 

[nicomak]

Actually I think it WOULD make sense. If it was too close, it would be a bit odd if for instance I as an explorer had been on an obscure empty planet yesterday, then I download an update with some new DLC and all of a sudden I am slap bang in the middle of the thargoid homeworlds.

BUT

I dont think this is an issue. In the 1st 1st game there were 8 (I think) separate galaxies which you needed a special drive to travel too. I think an equivalent thing could be for in a new big DLC, a worm hole (or something similar) is discovered which, with the purchase of a special jump drive, you can go through and takes you to a whole new section of space previously way to far away to get too.

no you wouldnt be able to enter that system in the first place so no "thargoid" surprise(see development plan video), but what i meant is it has to be reachable(almost, or in the "near" future), there's no use to start adding content in the other side of the galaxy if there's no possibility to get there right away, now yes if they decide to add some kind of mhole, then it's different.

 

[Herik]

That's the whole point of procedurally generating stuff though, you don't need to do anywhere near as much by hand, you just need to set up some well thought-out parameters and then let it rip.

Even if they went for something as simple as only four possible features on a planet (let's say desert, ocean, grassy and mountainous) and ten basic presets for each of those four feature types (so let's say full surface of the planet, one ninth coverage, one eighth coverage etc). Then even something as basic as that system is going to give you a possible ten thousand different planet configurations at it's most basic level, many more permutations if you randomise coastlines and terrain heights.

And we know that FD have a vastly more complex and feature-rich planetary generator than that because we've seen a preview of it, which is based on all kinds of stuff such as type and proximity of stars, orbit and tilt of the planet etc, etc, etc.

When you bear that in mind, the chances of running across two planets even vaguely the same as one another is probably going to be minuscule. They will all look and feel different, so you are always going to see something new.

There is no reason why FD cannot apply this methodology to things such as creatures which might populate a planet, or derelict vessels, or bits of old technology you might find that you could add to your ship. All it needs is a dice roll when you jump into the system to see whether it gets plonked in the systems somewhere, or not.

Granted, it wouldn't be the work of five minutes to set it up, but neither would it be a lifetime's work, so long as it was fairly well thought out and tested.

Very good Chock.:smilie:

 

[Kulin]

I've never seen a game that used procedural generation really well though. Like using procedural generation to create mission and story as well as pretty landscapes. And have a simulation that feeds these story and missions. Sure you have every right to be skeptical, but I wouldn't write the idea off.

I know a few examples. The best one, i think, is Starbound. Even if it's 2D, you can see the possibilities growing from patch to patch. Yes, it's just bringing together a set of pre-defined pieces. Pieces like: procedural trees(different fruits, colors, blossoms, loot after harvesting), procedural monsters(shape, color, behavior, attack-styles), procedural planets(surface-color, size, materials/material-color, resources, asteroids in atmosphere, inhabitants, underground & caves, background), procedural POIs(cities, prisons, laboratories, ancient ruins, pirate bases and tons of more things to explore). And so on. Yes, there are single pieces you might remember. Like the shape of a flower, but since color and loot is different, its not the same as before. There are so much variables that the overall picture is never the same. Its not just about random generated planets with different surfaces, size and terrain color.

Space Engine is also nice. Even if there is no random generated content at all beside the planets, roids and suns, the planets look all different.

Now imagine random generated space stations(inhabited, abandoned, relic, non-human), floating ships, mining outposts, black holes, pulsars, neutron stars, comets, nebulae with different effects on your ship, pirates, faction interactions, trade-routes, laws, missions. Tons and tons of variables.

I don't think that we have to be afraid about star systems looking all the same. I think there is more potential to do something in the wrong way if the things you really DO are too similar. Think about exploration. Its not only about finding new star systems. Its also about finding new POIs in existing space. If this POIs are all the same, even if they look slightly different, exploration will be boring very soon.

I think there are three types of relevant( things you can actually interact with) POIs when you enter a new system. Mining-POIs(roids, comets, planetoids). Station-POIs(Trade, Pirates, Relics). Factions & Inhabitants(Missions and Trading).

• If there are only two types of asteroid fields, categorized by yourself in valuable and trash it does not matter if they look all different. It gets boring after you got the money you need, if the mining and the scanning is too dull and if all the mining is just about making money. There could be much more to it. For example the search for real rare materials that help you to build something special.
• The same with stations. If there are only inhabited(trading + random missions), abandoned(with some loot) and pirate stations they are very soon getting boring, too. This does not depend on the visuals.
• The same thing is happening when interacting with factions/inhabitants for trading and random missions. If there are only 5 types of random missions(bounty hunting, seek and destroy, base defence, ambush, base attack) you won't be doing them for a real long time. And if there is only trading without production chains, base building or something like this, it will be over when you have enough money, too.

So the main concern shouldn't be: Are the systems looking different enough to fill a few billions of star systems? It should be more about: are the things we can do actually deep enough to entertain us a long time when everything is procedurally generated?

 

[JeffRyan]

I think one problem might be player mindset. We have to bring the story, care about our ships and NPCs and missions. We have the responsibility to bring our imaginations, Frontier must gives us the tools to allow that to flourish. FD don't really have to give us the story. This is very different from most modern games.

 

[Crayfish]

I guess my question is what purpose will all that space serve in gameplay terms? It seems to me that once you travel a certain distance there's really nothing more to see.

Although most systems will probably be quite generic (as they are in the real galaxy), that doesn't mean there won't be any point to exploring deep space. The procedural generation algorithm can be tuned so that "gems" can still be discovered in deep space.

For example if we have three desirable attributes for star systems that each have a one in one thousand chance of appearing, only one in one billion star systems will have all three features. The discovery of a system with even just two of those features would be newsworthy and a planet with all three could be game changing.

 

[Exmortis]

Although most systems will probably be quite generic (as they are in the real galaxy), that doesn't mean there won't be any point to exploring deep space. The procedural generation algorithm can be tuned so that "gems" can still be discovered in deep space.

For example if we have three desirable attributes for star systems that each have a one in one thousand chance of appearing, only one in one billion star systems will have all three features. The discovery of a system with even just two of those features would be newsworthy and a planet with all three could be game changing.

Yeah I get it. It'll be tough getting the right balance though. After visiting 200 generic spots one might get discouraged about finding that proverbial needle in a haystack. Aesthetic variation is one thing (and not really all that important to me), but achieving the proper mix of gameplay affecting variation will be tough.