Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Beyondrepair]

I'm sure several people had the thought of making some space colony builder with similar concepts to DF fortress mode -> did anyone actually get started?

Searching this forum didn't turn anything useful, but perhaps someone (if any such project still is alive) has a clue.

[somebody]

Theres Maia but you have to buy it.

Its basically a DF Mars colony game but it's not even in alpha, although it looks promising from what the developer has shown us.

Bay12 thread
http://www.bay12forums.com/smf/index.php?topic=118473.msg3746005#msg3746005
Website
www.maiagame.com

[Beyondrepair]

^ Seems more like Dungeon Keeper in space (not that it's a bad thing, will definitely keep an eye at it!).

My original question was more in the vein of "you see several cool projects being presented in this forum every so often but little seems to make it past early implementation stage".

I do have reasons for asking, primarily as I don't enjoy repeating work done by others.

[Randos]

I am working on something primarily inspired by the randomization and simulation aspects of DF which takes place in an "alien environment", although I think it will become relatively abstract in comparison and I am not intending to do so much z-level traversal as I am interested in traversing planes.  I have actually had this idea for years before even playing DF, but I think after playing DF I was really inspired to get into production because I was basically looking at a working model of something I had always wanted in a game.

This project is already in development phase but just slightly beneath presentation so I won't be showing any screenshots yet.  I do hope to release an abbreviated version by the end of this year.  The basic scenario is:  Randomly generated creatures, you are controlling a certain race.  Start off with nothing and grow into building shelter, maintain renewable food source, and etc while defending yourself from opposing creatures. 

The reason I associate it with DF (in theory) is because each game play should be completely randomly generated terrains, creatures, objects, and basically everything that I can apply randomization too - as opposed to the other simulation games which tend to have static character types and terrain maps, I don't find that very interesting.  And because of these things I will be focusing less on graphics and more on a mental conceptualization of the activities in the environment which also seems to be a DF-like quality.

[Beyondrepair]

Randos: That sounds cool, while not overlapping my idea completely (I was imaging something more along the line of humans trying to settle an unknown planet), it's definitely close. Especially so since aliens are pretty much given adversaries in this kind of game.

Wouldn't mind more info on it. Don't worry to much about screenshots and such, as long as you keep working on it rather than (solely) talking about it, it will happen.

[Randos]

Beyondrepair:  Thanks!  I like the idea of settling alien planets too, but I decided not to focus on that for the sake of simplifying things on my first game of this type.  I have made quite a few Flash games and such before but these types of strategy games take a lot more work.  I am a year into it already, and don't worry it will get done if even in a less than ideal completed state, but hopefully I can bring it all the way home. 

I was thinking there are a lot of cool things to take into consideration with the foreign planet / alien environments such as:

- time length of day/night and different seasons
- maybe critical weather during certain seasons
- random natural disasters possibly occurring more often than Earth
- and of course those off world critters !

And so the survival elements become much more interesting.  I actually am not including many of these things in my game, I am focusing more on survival, exploration, and the interaction of the creatures.  I am trying to stray away from "good vs evil" and simply allowing things to occur naturally.  So when or IF different races ever start fighting each other it will likely be for a reason such as food shortages, squabbles gone out of hand, greed, or other naturally occurring events.  This obviously takes a rather sophisticated AI which is why I am planning on spending another year at it.

I know what you mean about people talking a lot about making things and never really doing them.  Although you can't blame them for getting excited with their ideas and then realizing how difficult of a challenge they are facing afterwards In my situation I have actually been a programmer for years already and so it's really just a matter of putting the time in.  I keep a log in a private wiki and update it every time I work on the game and/or have any new ideas I want to write down.  I think little work ethics like that help to get the job done.  It really is an overwhelming task for one to take on by themselves, especially if you are busy working for a living at the same time.

[Beyondrepair]

True that.

It will be very interesting to see what you end up with. If you managed to keep going for a year already you've probably outlasted 95%+ of the projects started here.

[zwei]

I have already written worldgen for galaxy/starsystems :-)

[Beyondrepair]

^ Cool, though it's a lot more work to have a full game.

Does anybody have opinion on how 2D version of DF was? Is going "3D" (world) worth it or does it over-complicate matters? I'm still debating whether it would be worth it for a game of aforementioned type. Heck, even in DF only a few levels are really used for building stuff with the exception of crazed mega projects...

[zwei]

^ Cool, though it's a lot more work to have a full game.

Does anybody have opinion on how 2D version of DF was? Is going "3D" (world) worth it or does it over-complicate matters? I'm still debating whether it would be worth it for a game of aforementioned type. Heck, even in DF only a few levels are really used for building stuff with the exception of crazed mega projects...

2D is definitelly much simpler to implement and resource friendly for computers.

However, in sci-fi settings you could to cool stuff with 3D:
 * Per planet geology
 * Make alien structures like this: http://www.rockymountainmagazine.com/national-parks/arches-national-park-utah/ very common.
 * Embark on asteroid or comet. Or building space habiat / space station.
 * Floating colony embark on world with thick atmoshere with elevator to bottom.
 * Domed colony on airless or toxic world
 * Floating colony in planet completelly composed of liquid.

Those would be possible, but lackluster in 2D.

[Beyondrepair]

Exploiting some of those 3D scenarios you mention would be problematic due to how one should handle gravity:

Embarking on a small Asteroid would be that handling gravity direction wouldn't exactly be straightforward because the it would need to vary, and in a 3D grid you only have six directions. OTOH on maps where surface is so small it cannot be approximated by a plane gravity could probably be approximated to 0 with good results. Still, with no defined "down direction" you would have enormous implementation concessions to make (consider our perception of a "floor" for instance).

Floating colony - heavily assumes you're simulating gravity for blocks (note that DF has nearly no simulation of gravity on blocks except for the "connected to ground flood fill"). If your building blocks were something else than cells it would easier to do it justice, but still at a high computational cost.

Airless/toxic world can be done in 2D as well, though environment is definitely more interesting in 3D (just an inevitable PITA viewing it, even in a 3D viewer due to underground features). But from a gameplay perspective... is it worth it...

Appreciate your thoughts.

[DarkWolfXV]

That would be incredibly awesome, space ships and shit.

[Beyondrepair]

TBH I'm not sure how well space ships would work, space colony seems to work better simply for the fact the environment can be static which is an enormously simplifying factor. Perhaps you could construct a static spaceship then have an implicit launch while changing the frame of reference, ending up with managing the spaceship in space. Still, once launched, a spaceship wouldn't carry much materials for building stuff, a significant factor to what makes DF fun.

Therefore (unless you wanted to go on a tangent and make a FTL like game) I think planetary colony is a better focus.

Started experimenting a bit myself last weekend:

Spoiler (click to show/hide)

Can build a pressure chamber to make sure no air escapes my 3x3 main space when my lone astronaut needs to go outside. I.e. I let it pressurize, then pump air back into main space. Not sure how accurate it is. Not that there's anything worth seeing outside as of yet, not that I will prioritize terrain generation any time soon before having some gameplay worth generating terrains for.

[DreamThorn]

Looking cool so far.

I think you should stick with 2D, so that you can have a more dynamic world; cellular automata aren't really practical in 3D on modern CPUs.  (I assume using the GPU is out of the question.)

[Beyondrepair]

I've considered using OpenCL but... the issue is basically boils down to: How large part of the world will have physical simulation running at any given moment.

After some thinking I chose a sparse representation where the only cells that participate in physics simulation are on a list of active cells. Unfortunately, to access neighbors in such a scheme one has to use a map which is rather unfriendly to GPU (or otherwise introduces a data preparation stage on CPU). (Just using a hash table at the moment but an octree is likely to be in the final implementation.)

On the flip side, this approach I have now means I can't really simulate a significant part of the outdoor environment. But since I want a game, not a physics simulation, I still believe it's the right approach to not do physics simulation everywhere. This means I can actually have a conceptual grid of 1000x1000x400 cells if I want to without exceeding memory bounds so very badly.

The 2D vs 3D question is still very relevant and I haven't made the final decision on that (only did 2D yet). It's true that if 2D is chosen it makes doing full physics simulation a whole lot practical, but again, the question is whether there's a point in doing full physics simulation in 2D at all. So I'm fairly sure the real question is: Is 2D more fun in this context, or does 3D add enough to motivate the added confusion multiple Z levels cause?

My plan for this is basically 1) implement a tiny subset of the game that exhibits all the main performance constraints this type of game has, 2) try to optimize it for multiple CPU cores (Intel TBB), 3) make the rest of the game (the latter stage of which I would open it up - have a Github repo already but I don't see any point of sharing it until I've nailed down the architecture). Another thing I'm trying for this game is using unit tests everywhere - in my earlier games I haven't used automatic testing much at all, so it's an interesting departure from my normal workflow.