Page 2 of 3

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 6:46 pm
by Titan21
Aww, I was so hoping for some feedback / acknowledgements about Factorissimo :)

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 6:58 pm
by Twinsen
Titan21 wrote:Aww, I was so hoping for some feedback / acknowledgements about Factorissimo :)
Yeah, we did have a few talks about it. We probably forgot to put it in the Community Spotlight :)

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 7:07 pm
by EstebanLB
Hey, what about using different sizes for the terrain patches? That will give more diversity to the terrain while still using a small set of designs

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 7:38 pm
by alan2here
Yes, embrace the fractal :) or the pollock if you prefer, I think that artist said that strokes of different scales make it into his work, presumably for similar reasons.

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 8:05 pm
by RobertTerwilliger
New biome promises some new fresh variety (only aesthetic though, however aesthetics is also quite important thing)
This will probably cause some rewriting of map generator, 'cause snow will look VERY odd close to desert. Not sure if we can have forest up to desert now, but I almost positive we can, and this will be not really acceptable in snow+desert example.

Those tile patches deffinitely need some rotations at least))
#2 looks the best, followed by #6. I know this is very early WIP, but still, these two are quite good already.

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 8:13 pm
by Jump
Where is doodads #14? :)

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 8:29 pm
by ssilk
Interesting read and some things are now more clear to me.

I remember my days, when I programmed on C64 and I made a small crappy game for me I used for the timing of the main-game-loop the internal interrupts. It was not a loop, but an interrupt-driven small program with start and end. The start was triggered by the interrupt (a signal on the CPU, generated by a special timer chip). The CPU then saved it's content to the stack and jumped to the begin of the main game loop, did some screen updates, things that were needed to make the game run fluid and regular and ended. The CPU restored the the state from before the interrupt (read from stack) and continued it's job; in that case it calculated things like landscape (just a up and down), enemies, the movement of the vehicle (from the formerly read input) and then finished also, which means: It waited.

The timer chip was like a super, super exact clock.

The big advantage with the interrupt was: I used a second interrupt for the controls: Checking the keys, joysticks, looked for the bouncing of the contacts, etc. The big advantage of this was the super-fluid reaction to the input of the user. Cause the joystick was polled in that exact intervals the control was really exact.

When I worked the first time with a UNIX-like system (I think I was 15 or so) I asked, how they made an exact timing. The answer: Sleep. Then I asked: Ok, sleep sleeps for minimum a second. And how can I do timing in microseconds? Nanoseconds? Why I needed that, they asked. I answered: For music for example.
They looked at me like an alien.
Today's music industry would be unthinkable without that. :roll:

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 9:10 pm
by sabriath
An idea for the terrain so that it doesn't look tiled at different zoom levels is to use a similar technique as infinite map bumpiness. You take one of those terrain gfx and blow it out to 10x or even 100x its size but reduce its opacity to 20% and overlay it. This would make the overall terrain in local areas look different even while placing a few smaller gfx in its place. Map generation can easily randomize and store location, size and opacity levels in a list for each as chunks are created without wasting much memory....and if position isn't normalized to the block, it would be even more realistic.

Re: Friday Facts #150 - New Terrain Experiments

Posted: Fri Aug 05, 2016 11:12 pm
by MeduSalem
Interesting Friday Facts...


Funny coincidence that I'm currently writing a little Mod for Prison Architect and there I currently have similar problems like Factorio's main game loop (altough I don't need one that is 800 lines long, instead one with a couple dozen of lines). In Prison Architect the API is polling Lua scripts a couple hundreds of times each second and for each instance of a script. On top of that the game is really badly optimized and if your script is quite long or requesting other timeintensive stuff from the API you start noticing the lag because the script doesn't run asynchronous from the game engine. It waits until the script finishes, however long that takes.
So if you start modding on that game you have to implement a despamming structure to spread code over multiple frames or delay stuff for seconds to unclog your update loop. :D


On the new Terrain features/biomes... I hope you guys really consider giving the Biomes not only aesthetical properties, but also make it interesting for gameplay, like resources/enemies or some other things depending on it. Now would be a good time to think about that and there have been threads in the suggestions section about it. The Terrain generation has been broken anyways for some time now or at least there have been several people complaining about it.

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 5:16 am
by y.petremann
About biomes, firstly I assume that factorio use a system to generate temperature and humidity of the terrain correctly so it would choose the terrain depending of these parametters.

Tundra seems a great idea and I would love that, also for the red desert
Based on this image :
Image

I really would like some Ice/snow plains, and lava biomes (gray/black floor, lava lakes, poor with water lakes)

For now the game don't allow other transitions between water than with grass, I would be fond of seeing beachs or to see swamp with the greenish water, and finally, simple water would not be constructable but would be walkable

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 8:55 am
by miturion
I really like the red desert and the snow biome. They look really cool already

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 11:02 am
by alan2here
As I thought, somtimes desert, as well as various visually similar sorts of biomes, are adjacent to or even covered by snow :) But isn't that red Lava block on the diagram a bit long left/right wise?

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 12:14 pm
by malecord
Cool. Now that you work on biomes maybe you'll start consider adding weather and/or environmental conditions to the game that require the player some building maintenance (and automated maintenance of course) instead of current "build once and forget".

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 1:16 pm
by Dark_star
can doodads be rotated (0,90,180,270) when placed for terrain variety ?

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 2:06 pm
by FlyHigh
Very interesting, the new terrain. However I feel there needs to be a mechanism to put decals back into the terrain once a tile is put and then lifted.

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 2:18 pm
by EurypteriD
Great MS paint numbering job :P

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 2:41 pm
by MrDnmGr
New biomes? Awesome! I believe, however, that all proposed ground sprites (not the doodads) are oriented towards the South East (they all appear to have their definitive feature, AKA a corner or a curve, pointed to the SE), which might be the reason they create a pattern when zoomed out. Keep on squishin' those bugs!

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 3:43 pm
by hitzu
New_Dune_biome

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 5:04 pm
by Engimage
Always nice to see performance improvements.

While new biomes are nice to see, they are more of a "bells and whistles" category in my head. Especially looking at the current map generator code with all that biome soup. It was already mentioned multiple times that "terrain fragmentation" behaves really bad now.

Still waiting very much for that fluid handling...

Re: Friday Facts #150 - New Terrain Experiments

Posted: Sat Aug 06, 2016 6:38 pm
by RobertTerwilliger
hitzu wrote:
New_Dune_biome
That one's nice, especially considering we already have hostile giant worms : D