loving the game so far, but lets get real about the PL promises

[bassman]

Procedural generation is easier (not easy just easier) to do for planets as planets are always changing and so they don't have to be exactly the same every time someone visits and could be done in an instance just for that person, whereas stations have to be more static and the same for everyone sharing the universe. So I can see how they could use some sort of procedural thing to create atmosphere, land masses etc for planets and not have to store billions of bits of data to make sure it's exactly the same for everyone, whereas if they had loads of variety on stations that are fixed things it would mean much more data to be sent back and forth or larger game installs.

Procedural generation is not random and so everyone seeing the same thing is not an issue. Procedural generation is just much harder to do with things that need to look like they were designed as a whole with a specific purpose e.g. ships and stations rather than things that can have arbitrary variations.

 

[EDbubba]

Hang on, the stations are different. They use the same modules but the configurations are different, e.g. one, two or three wheels, different 'cores', some have solar panels, some have industrial modules. Some 'classic' stations have extended arms etc. The same goes for outposts.

True, they took the time to make the exterior of the stations vary a little, but the worst part is that once you enter a certain type of station the objects (boxes, doors, etc) near the pad(s) are identical.

This is by far the biggest problem with the station models at the moment, incredibly obvious and jarring. This particular oversight was inexcusable since it's a pretty trivial fix (randomize small elements while generating/building the station's interior).

As far as outposts go, how often have you seen the same "crashed" ship on one of the landing pads of any given station?

These kinds of copy/paste iterations are just horrible, and smack of shoddy, careless development.

 

[xkjacob]

From what I understand procedural gen is already there. If you are asking whether the guy who basically came up with it can do it on a modern system I would say yes.
Planet surfaces would remove an element of orbital placement , which I bet is a pain to design.
I wouldn't call this super complicated to do either, as the system level generator most likely has the key criteria necessary to populate the surface square footage with templates based off...
1. Surface sq meters
2. Habitable zones
3. Area temperature (would be cool if it took into account seasons!)
4. Population
5. Tech
6. Faction
7. System trend
8. Specific land marks
Based off just a few randomly or system dictated attributes you could boil down to a set of templates to express as you are in the entry sequence.

Tl;Dr.
With the system generated data available they can apply custom templates to each planet while you go through the entry sequence.

 

[Nagual]

Hang on, the stations are different. They use the same modules but the configurations are different, e.g. one, two or three wheels, different 'cores', some have solar panels, some have industrial modules. Some 'classic' stations have extended arms etc. The same goes for outposts.
:
You could argue that the number of base modules is too low or that there should be greater variety between Imperial and Federation assets, but not that there is no variety.
:
Regarding planetary landings. I can very well see FD pulling it off. Here's how they might do it.
:
When you get close to a planet at the moment you 'drop' into normal space. With PL you will 'drop' into 'planetary space'. From a code point of view this is similar to SC but instead of a solar system with planets, moons, stars etc you have a planetary atmosphere with land, water, ice forest etc. Instead of USS, stations etc you have cities, PoI, landing fields.

This pretty much sums up what a lot of folks have said. Although I'll add my own 2credits worth, ever since we saw the concept art for stations back at the start of Kickstarter, we imagined much more perceived variety. For the core systems we were given visions of futuristic cities in space, that made just about anything else you may have saw in games or films look tame by comparison. Right now, what we have is a very basic shadow of what I had hoped for. Nice palm trees. Great sports field. Ahh, this one is in red.


EDIT:

I blame this man:



https://forums.frontier.co.uk/showthread.php?t=7534

 

[xkjacob]

True, they took the time to make the exterior of the stations vary a little, but the worst part is that once you enter a certain type of station the objects (boxes, doors, etc) near the pad(s) are identical.

This is by far the biggest problem with the station models at the moment, incredibly obvious and jarring. This particular oversight was inexcusable since it's a pretty trivial fix (randomize small elements while generating/building the station's interior).

As far as outposts go, how often have you seen the same "crashed" ship on one of the landing pads of any given station?

These kinds of copy/paste iterations are just horrible, and smack of shoddy, careless development.

This is kind of a pain. Human brains are designed to pattern find. It should not be difficult for ED to add more templates to give more variety, but let's say there are 100000000000 templates but you can only distinguish 4 due to pattern finding bias.

The interior of the station I felt was a compromise that maybe they'll fix. I men, not like we can walk around.

 

[CMDR MAIN SEQUENCE]

FDEV has chosen not to do this with stations. It was their decision, not what's technically possible or not, that made this feature what it is today.

Was this part of the DDF/DDA? Do you have a link to the source? I'm not doubting you I'd just like to understand the reasoning behind that decision.

 

[kilvenny]

I think sometimes we forget that this game was built with a limited budget with no certainty of success. Even now as a result of around 500k sales they have only about 20 million Sterling to play with. Not much when you look at games like Star Citizen/Destiny. The good thing for us is this is a marvellous show case for their Cobra development tool, so building some sort of FPS into ED is vital if FD want to license their engine which is what planet landing/walking about is all about.

If I am building a game with a limited budget and I need to prioritise development, where do I compromise? In ED the easy choices are ships, stations and planets. The key thing is to build a design that allows new variants to be "easily" added. Over the last 12 months we have watched variety being added. The thing is its not a priority because the amount of variants needed before we wouldn't think them the same is huge.

What we can see is variation of colour, the clear use of elements of specific sizes which are interchangeable. That indicates to me some thoughtful design catering for additions down the line. Also elements used in the outposts could easily be adapted to create on planet bases.

In my view the big issue for the planet landings is what you can do once you are on the ground. Its worth looking at trailers for the Outsider a game FD made that was never published which indicates what their engine was capable of a few years ago.

 

[emacinnes]

Dont get me wrong Im loving the game, but with only 20h or so played, I can see that the game has a very limited amount of variety on the stations and systems, all almost look the same: the clouds in the planets, the stars, all the hangars are exactly the same in all the outposts and stations, the same. And after you fly into the widescreen, for example, you cant tell the difference between one station to another.

So... reality hit me like a big icy moon... frontier wont be able to do this, ever (from the wiki):

"Will it be possible to manually land on planets?

Yes, just like the last two Elite games you will be able to seamlessly freeform (manual pitch/yaw/roll) enter and fly through entire 1:1 scale populated planet atmospheres with living cities and wildlife enabling extra-vehicular activities as part of expansions."

Reality check: I dont think so.

How in this galaxy they can promise that if in the core space realm you dont have almost any variety? (in stations/outposts for example)

How?

There will be an inherent element of "sameness" in this as well, but with more overall variety, as the planet generation is just 2D and doesn't need to account for rotation etc... i.e. imagine it to be like SimCity 4, where any city can basically be generated, they'll all end up being some variation of a grid structure (or spheres, etc..), but with a different spread pattern. Individual building types will repeat-ish, as shapes can be auto-generated, but individual textures will be finite (i.e. stone buildings, brick buildings, wooden, etc..).

So there will be some kind of "blur your eyes and all cities are the same", which is as true in real life as it would be in the game, one hotel is much the same as another, and from 10,000 feet up most cities just look like a splodge on the landscape, but the general overall structure will be much more varied than stations, overall colour schemes and surrounding terrain, except landing ports which will have to follow a much more rigid set of rules.

 

[Senzafine]

I'm concerned whether procedural generation can deliver realistic planets on a level with Outerra. Because that uses over 10 Gigs of SRTM at 90m resolution. That elevation detail reflects hundreds of millions of years of very complex geological changes. So far Space Engine and NMS are the best we've seen and they don't get even close to looking truly realistic imho.
I'm also concerned that we're going to have to wait for PL dlc before planets look significantly better from orbit or if we can reach low orbit - ISS level. Much as planetary flight excites me the view from low orbit would really blow me away if it managed to look anything like as good as pictures we see from the Space Shuttles etc. Or even as good as Outerra would look if its team ever manage to implement rivers, lakes and cloud/weather systems.

 

[Holven]

I think procedural generation of stations with variations is a possibility, but not before stations and all their functions are fully developed. It is impossible to foresee what you need the procedural generation to handle, until first person walk about with features have been added in. Then and only then would it be possible to figure out what sort of variations one would like a procedural station generator to handle and figure out exactly to which extent it is possible.

That being said, it would be a monumental task to do and I think such a task would be something that could be considered much later in the game's lifespan.

If it were to be done at all, I'd think it would be around halfway through the game's projected lifespan. If version 1.x is centered around adding functions and content within the current model, I'd expect version 2.x (paid expansion) to focus on early planetary landing and first person. If so, I'd expect 3.x (also paid expansion) could be the absolutely earliest possible point for any complete overhauls, possibly not until later, as I'd fully expect planetary landings and first person to be split into 2.x and 3.x.

 

[jcrg99]

I think procedural generation of stations with variations is a possibility, but not before stations and all their functions are fully developed. It is impossible to foresee what you need the procedural generation to handle, until first person walk about with features have been added in. Then and only then would it be possible to figure out what sort of variations one would like a procedural station generator to handle and figure out exactly to which extent it is possible.

That being said, it would be a monumental task to do and I think such a task would be something that could be considered much later in the game's lifespan.

If it were to be done at all, I'd think it would be around halfway through the game's projected lifespan. If version 1.x is centered around adding functions and content within the current model, I'd expect version 2.x (paid expansion) to focus on early planetary landing and first person. If so, I'd expect 3.x (also paid expansion) could be the absolutely earliest possible point for any complete overhauls, possibly not until later.

I agree.

While everything have to be improved and with time it will be, I think that the entire focus of Frontier from a competitive perspective is NOT to put visuals as their priority, but meaningful features and gameplay modes instead.

Star Citizen is coming in the future (maybe) with visual fidelity appeal as their entire motto, and in practice (except by dreams) everything that they really will be able to deliver is basically that, limiting a lot in many promised features due that, and basically, in practice (not in dreams and theories) with serious limitation to advance fast to add meaningful features and updates ,as Frontier currently and always will be if focusing/prioritizing the right thing.

While for single-player the path of SC is a good one, for MMO (meaning that who is faster to deliver meaningful and continuous updates, wins more fidelity) is a terrible one, and Frontier is doing the right thing at this case.

Higher level of fidelity is great, but MMO (long-term experience and fidelity, not short-term adventure as Single-Player games delivers) a balance must to exist so the things advance in what is more priority for a majority of people in the world... and that puts meaningful features far ahead of graphical fidelity (who actually can limit the implementation of more meaningful features in TIMELY manner in an unique and seamless experience).

So, I think that Frontier has been smarter... and will continue to be.

 

[Koen]

Minecraft in space! With a fixed seed, it's possible to generate the same planet every time you visit.

Hey, I remember a 1984 game which did something like that! Well, not for planetary landings, but it did generate 8 galaxies from a fixed seed...

 

[jcrg99]

Hey, I remember a 1984 game which did something like that! Well, not for planet landings, but they did generate 8 galaxies from a seed...

Yes... Tell me how the person who has been working on this kind of thing since 1984 would know if its possible to add such thing. No way... Braben just told that out of his mind, without evaluate, just to hype people so they bought ship jpegs for hundreds of dollars each, regardless if such idea could be done or not in any time before 10 years... oh wait... wrong game...