search

minetest / minetest_game

Minetest Game - A lightweight and well-maintained base for modding [https://github.com/minetest/minetest/]
http://minetest.net/
Other
1.42k stars 571 forks source link

[Discussion] Development direction #515

Closed Ekdohibs closed 4 years ago

Ekdohibs commented 9 years ago

Minetest Game is as of now the only game that is bundled with Minetest. Thus, it should be made attractive to newcomers (actually, I would say that Minetest Game needs to be developped more, and that making a basic game is for minimal). Now, looking at what most newcomers want:

paramat commented 9 years ago

Mgv5/v7 will soon have 8 biomes including a sandstone biome and dunes where sand and grasses overlap. I could happily add more biomes but these require many new nodes and i encounter resistence from the MTGame team to any new nodes (for example see the freshwater/riverwater discussion).

We could add savanna with Acacia trees and the red wood items (see watershed/paragenv7/riverdev) I was actually trying to keep the new biome system fairly simple and 'Minetest classic' as i got the impression that is what is wanted (i was wrong).

est31 commented 9 years ago

About mobs: Yes, please, they have to be of higher quality though than current mobs. What I've seen with current mobs is that we need better way to highlight them (the current frames look bad, mobs are no blocks), they need better models (they look ugly), and better behaviour. Some are engine issues, some can be done by the game.

About quick travel: I'd like to have carts too, because teleporters can affect gameplay too much, but currently they are subject to lag, which makes all current carts horrible. Carts would therefore either require client side scripting or native support by clients. I also don't like with carts that there is no way to accelerate/decelerate them once you're in them. It would be cool to have a break and some manual acceleration mechanism. Also it would be great to be abled to decide which way you want to pass a switch from inside the cart (I know this doesn't match reality, but it would give at least some control over carts).

About goals: goals aren't what minetest is about, are they?

paramat commented 9 years ago

Minimal game is too minimal for a basic game, so MTGame is the basic game, it is is developed too far we would just need to create a new basic game to take it's place. Some more development of MTGame is good but we really need a variety of subgames.

C++ mob infrastructure may be needed to reduce their load and lagginess on lua. Mobs should be optional (but perhaps on by default) as many don't want combat.

I feel Minetest Game should not have a goal, that can be left to more specialised subgames. It's more about open-ended exploration and building, and if mobs are present, survival. The 'End' of MC was a mistake that seems out of place and seems a decision made under the mainstream pressure and assumption that a goal is needed.

C1ffisme commented 9 years ago

/me applauds.

About number 4, I agree that carts are a good option. One of the things about minecraft is, you don't have to build everything in one building. You can, but it's fun to build lots of things like a brewing station and enchanting table/library, In minetest the only method of crafting is the furnace. Even with a crafting table, you can still build special places to contain them in minecraft. In minetest you really don't have any reason to build. So by adding minecarts, you can build a whole new line of buildings, train stations, train manufacturing stations, T-junctions, etc.

About mobs, we just need a C++ API!!!! Everything would go about 20 times faster.

Minetest should have a Goal, that way there is a REASON to build buildings and kill mobs. Without the thought of having bragging rights for having killing the over-powered ender dragon. Otherwise, people just parkour around and such. You know that server feeling you get when you feel like you have nothing to do? You just parkour around spawn? Yeah, that happens without goals.

Lymkwi commented 9 years ago

Minetest should have a Goal, that way there is a REASON to build buildings and kill mobs.

I don't think Minetest Game should be a game with a goal. Not only people would stop playing when they reach it, but Minetest Game should be a game designed to make players discover the nodes, the biomes, and the environment of the world. If you want a specific goal, then I think it must be included in a second subgame. However, if you talk about some specific goals, they might be included with some specific settings and modifications in the actual subgame : enabling damages could trigger a timer and a list about players that have stood alive the longest, creative would enable some building achivements, an so on.

C1ffisme commented 9 years ago

@LeMagnesium Haven't you ever played Minecraft before? I defeated the ender dragon and I still play it. I built a security room for my dragon egg.

Maybe we could add goals.

Lymkwi commented 9 years ago

@C1ffisme , I never played MineCraft before. And if I reach the goal, I assume I have finished the game.

paramat commented 9 years ago

Concerning mgv6 biomes, there are now 5 with snowbiomes enabled: tundra, taiga, normal, desert and jungle. The original addition of hot biomes split into hot dry and hot wet is now mirrored by the addition of cold biomes split into cold dry and cold wet. The system is nicely balanced and complete and i feel should not be expnded due to the simple classic character of mgv6. Because mgv6 is complex and hardcoded adding more biomes is complex work.

The biome API of mgv5/v7/v8/fractal is however ideal for adding more biomes. I will soon be making biomes larger by increasing the noise 'spreads' of heat and humidity noises, the noise 'spreads' of the flora are also twice as big as in mgv6, both these will encourage more travelling.

I support adding carts as default, perhaps after some further optimisation work.

twoelk commented 9 years ago

After having played with Minetest for some two years now I am convinced that the default game of Minetest should indeed not be based on reaching a big hardcoded goal. Rather it should be made easy for a subgame to define such a goal. A model use case might be the idea of providing Minetest as a tool for education. As such it should be made easy for a teacher to quickly define a mission for the class and as any good tool, with little effort, adjusted to an unlimited number of new goals. I guess basic for this would be some easily configured achievement system. Indeed many things useful for teachers could be of benefit for other server admins and game designers.

What the default game could or maybe even should provide for a singleplayer is some sort of crafting guide and achievement tree that leads the player through all aspects of the game. This could include producing materials that include several levels of crafting, coaxing the player to search and visit every kind of biome and introducing the player to farming. So maybe the default game could act as an extension of the tutorial focused more on the gameplay once the player knows how to jump and punch. Of course if the player just wishes to "live" in his world and do whatever pleases him he should not be molested by some forced gameplay focused on achieving certain goals. I played mc quite intensive for over a year before switching to mt and I pretty soon turned of the hostile mobs and never searched for any dragons on pillars and guess what, I still very much enjoyed playing the game my way - just exploring, farming and above all building. There are many ways to enjoy a game.

What I once said as ironic almost sarcastic comment doesn't appear to me as such an unrealistic goal anymore: IRC-2013-07-03

"so you want some random gameplay gain by generating a random map with a random danger to suit every random player to get some random game depth? ... we need some random genius then."

Minetest boasts to be just an engine that anybody can easily use to build whatever game suits him. If this is indeed the major goal then work should be put into making this easier and games that show good ideas should be featured more prominently.

Calinou commented 9 years ago

Some way to travel quickly.

Just increase movement speed. We can set it to 4.5 (and 1.5 when sneaking) just fine.

I, however, definitely agree that we need lightweight mobs. Simple Mobs is quite suited to this, it might even be added as-is with some fixes.

Wuzzy2 commented 9 years ago

Please be very careful with mobs, because currently, most mob mobs are a horrible mess:

This mess is mostly to be explained by total clumsiness and total disregard for compability from mob mod makers. I can't think of another explanation.

So, if you include any mob mod, you need to be aware that the subgame is now de-facto barred from being combined by the player with another mob mod from the community. A player can not simply load another mob mod, because most mob mobs are incompatible. Unless the modding community somehow decides to clean up this mess. But you could also decide to make Minetest Game a “complete” game but then drop the goal of extensibility then, so you accept the incompability with other mob mods.

But another problem would be server operators. I bet they would instantly hate if such a change is done, because such a change would almost certainly break current heavily modded Minetest Game servers.

To the “goal” thing I want to add: What I would like to see are more hazards to the game, not neccessarily mobs. Something like traps or whatever. But hazards would not be enough. Survival is also way too easy. You can easily keep your health high by just stockpiling 100s of apples and breads. This is possible because they instantly and directly heal you and those items are also relatively cheap.

So at the end of the day, the main question boils down to: Do you want to throw away compability and stability in favor of gameplay, or do you want to be more “conservative”?

est31 commented 9 years ago

We'll should chose (or write) one API with good design, and then all mobs mods have to use that. And forking is bad behaviour, yes. I think this isn't a signal to stop and let things solve themselves, if they ever will, but to start making a solution. If there is a good mob api, perhaps mobs mods will use that. You see a catastrophy where none is. Heavily modded servers usually have very skilled modders, who will know how to adapt their infrastructure to the new API.

Wuzzy2 commented 9 years ago

Est31, you obviously have never attempted to create a subgame with multiple mob mobs. You will quickly run into many problems, some obvious, a few of them subtle (of the sort “why the hell does mob XYZ disappear when I include mod A?”). Anyways, finding a API with good design sounds challenging. I have already mentioned Casimir's Mob API. It's nice and simple, but clearly not the most feature-rich API out there. Other APIs are much more powerful, but sadly also much more cluttered with mobs, so it should be decided which mobs to keep and which to throw away. Seperating mobs from the API should also be considered, if this is needed.

To be clear: I am not against mobs in Minetest Game, I am actually in favor of them. I just said such a step should be done with care. Unless you want to screw (backwards) compability, in which case (almost) everything is allowed. ;-)

est31 commented 9 years ago

Yes, of course, doing things carefully is very important. I guess one can chose an API that doesn't negatively influence compatibility issues.

C1ffisme commented 9 years ago

@Wuzzy2 About mobs, most are created by one person, and get forked and model creativity gets mixed. However, if one or two people did the modeling, some did the coding and others did the artwork, than we might have better mobs. And we should have rules, like no using polygons, just cubes.

kilbith commented 9 years ago

Just increase movement speed. We can set it to 4.5 (and 1.5 when sneaking) just fine.

No, we should remain at 4 nodes/s.

1 node = 1 m. 4 nodes/s corresponds to 14.4 km/h. 4.5 nodes/s corresponds to 16.2 km/h (!). The human walking speed is averagely at 4-5 km/h.

Wuzzy2 commented 9 years ago

@kilbith: Minetest Game does not have one goal for sure, and that is realism. Therefore, your argument is invalid. You should more argue from a gameplay perspective. Also, the “1 node = 1 m” rule seems to be pretty nonsense and arbitrary anyways, Minetest Game apple trees are only 4 or 5 m high, leaves start a 2 m above the ground, Sam is 2 m high, etc., etc., etc. But as I said, realism is not a goal.

kilbith commented 9 years ago

Sam is 1m75 (average human male height) compared next to 2 nodes stacked (considering 1 node = 1m). Apple's trees are culminating at 6-7 m, so yeah it's realistic and not arbitrary.

As for a pure gameplay standpoint (although realism has a role in there), better keep 4 nodes/s for walking and grant "fast" priv by default (with a reduced speed) and the addition of a stroke movement like in MC. 4.5 nodes/s with the current "fast" is just unplayable.

C1ffisme commented 9 years ago

One of the main things about a game is that it has a puzzle, hidden by a metaphor or story. Every game is just a puzzle that you need to solve. Think of a strategy game. It's basically just a puzzle you have to solve, but covered up by the story of nations battling for control. And usually, but not always, the puzzle is what makes the game fun.

In Minetest, there isn't a puzzle. There isn't anything to solve.

Let's ask some "Why?"s.

Why should I build a house? There aren't any mobs!

Why should I build an animal pen? There isn't any animals.

Why should I build a city? There are no NPCs.

Why should I build at all? There aren't any goals to achieve or uses for my buildings.

Why should I mine? There isn't any constructions for my materials to build.

Why should I play? There isn't anything to do.

See, let's look at Minecraft for a minute.

Why should I build a house? Because it's getting closer and closer to night and you need protective shelter.

Why should I build an animal pen? Because it turns out that you need to eat. Animals are one of the few ways to get a hot meal on the table, plus you could use a bonus leather or wool to make books and beds.

Why should I build a city? Because NPC's can inhabit it, and you can fill the city up with brewing bars, washers (cauldrons), farms, etc.

Why should I build at all? There are plenty of reasons to build. Including using each building to get to a goal like defeating the ender dragon or a wither. I find it fun to build places specifically for just one kind of crafting, like making bricks, even though the basic needs (a furnace) are found in almost every other building.

Why should I mine? Because you can't just pull these resources out of nowhere! You need to put some effort into getting them!

Why should I play at all? Because there is so much to do! Want to mine in an underground and challenge yourself to how much cobble you can get? Go ahead! Want to fight a LOT of mobs? Take a sword and some armor and have fun!

Okay let's look more closely at the reasons. The first three obviously require mobs. Since in most games, you aren't alone. It's fun to capture and breed mobs, it's almost like a mini-game where you take care of a pet. And no mater what your life is like, everyone likes to fight things. It's human nature. If it wasn't, shooter games wouldn't be a thing. So fighting zombies and skeletons provides the fun thing people naturally like. And NPCs? Some people like to talk or help other people. It's fun for them to build new things for NPCs like houses and trade with them.

But now, let's look at the fourth one. The fourth reason for Minecraft says that there are many options for building. See, you can build big buildings just for brewing. Imagine a potion factory with a hundred brewing stations going on at the same time! In Minetest, though, there is only one kind of non-player crafting. The furnace. See, even in Minecraft you could build a stair crafting station, because of the crafting table. But not in Minetest. In Minetest, you can craft basically EVERYTHING from your inventory. And because of that, you only need to craft one furnace because you don't need to make space for other things like cauldrons and enchanting tables. Here's an idea: In Minetest, we should add some new methods of crafting or upgrading. For example: Tinkering Tables (Like enchanting tables, but you can add new technologies), Grills or another method of cooking (Make tastier and more hunger satisfying food), Mill Grinders (Grind wheat to make flour), Alloy furnaces (Pretty obvious, right?), etc. That way you can build Research labs, Kitchens, Mills, and Factories. Plenty of things to make right? And by adding things like weather or mobs or airplanes, you can build Traps and Weather Radars and even Airports!

See, about the fifth reason, one of the things about mining currently is that there isn't a lot to mine. See, if people are looking for diamond, they get bored really fast. What I suggest is that we add a whole new line of ores. People love to discover new things. They like to read the news and magazines. We can keep people interested if they keep discovering new ores. People could explore the land more if we added trees or generated cities!

The last reason is pretty obvious. We should play if we can do all those things. However, another thing to note is that even with all these things to do, people may still prefer Minecraft. We've already seen this, since even with technic on servers, most people don't do a lot there isn't much to do. So, one of the things about Minecraft is that it has a goal. You kill mobs and breed animals to work towards your goal. Maybe it isn't always fighting the ender dragon. Maybe you are on a server and you want to become one of the most popular, powerful players on the server. Maybe you are in singleplayer and you want to build a giant metropolis in survival mode. I think if we can achieve some of the reasons I mentioned, Minetest will become more challenging and more fun. Because if we go back to puzzles and metaphors, the thing is that a more challenging puzzle will be more fun to do. Adding new rules to the puzzle like mobs to fight and hunger will make the puzzle harder to solve, which makes it fun.

So, this is a really long post, isn't it?

Wuzzy2 commented 9 years ago

I agree a lot with your 5th point. Not only isn't there a lot to mine in Minetest Game, but the entire ground is pretty monotonous. Almost everything is stone with rare exceptions being desert stone (only at surface level), cobblestone and mossy cobblestone (only in dungeons). Minetest Game does not play around with different and harder stones. It basically all boils just down how long you take until you hit diamond. This is not really difficult, it just takes some time, the only danger is lava which can only happen to you if you are very careless at mining. ;) Also, the mining levels are pretty unbalanced. IIRC you only need a steel pickaxe and you are already able to dig anything except cloud (which can't be dug anyways). Even the diamond block is broken by a steel pickaxe. The better tools just do it faster. So the digging thing has no real sense of progression either.

But this is a general problem with Minetest Game. What you are talking here about is “depth” (game development term). Minetest Game has a low depth because there is simply not that much to explore and discover. Eventually, you will have discovered everything which can be discovered (all blocks and crafts), from which point the game will repeat. The fact that the map is automatically generated is great, but Minetest's Game generated maps still feel way too repetitive after a while; while nothing is actually the same, everything looks the same, especially when you have seen the 10000th desert, the 500000 jungle tree etc.

But the real question is whether it is actually desired by the developers to turn Minetest Game into a game or leave it as a sandbox as it is now. As a sandbox, Minetest Game does kinda okay, but there's always room for more stuff.

Oh, and please stop refering to games as “puzzles”. While this is not false, since all games are also puzzles (because they have a goal), this does the games unjustice. There are 2 fundamental differences between games and puzzles, and that is a) games have non-trivial challenges and b) games have meaningful decisions.

C1ffisme commented 9 years ago

Puzzle is more of just a word. You could think of other ways to describe it. Problem, situation, challenge, etc.

I see the difference. Because a puzzle like sudoku has one solution, a game has several methods of winning. For example, you could win a strategy game with a huge force, or make a small strike force disctract the army and take over the city / capture the flag / whatever.

Also, does minetest have any meaningful decisions?

About repetitivity, I agree. Maybe minetest should randomly generate new kinds of biomes, with a mix of generate-able blocks (different types of dirt and sand, stone, many plants and vegetables, etc.)

See, one of the things about Minecraft and Terraria is their high depth. There are so many things to do with your life in those games...

C1ffisme commented 9 years ago

By the way Wuzzy, about mob mods, we should make our own API from scratch. A clean, smooth start.

minetest.register_mob(name,def)

minetest.register_aem(def) (Might help for growing mobs into other mobs or on-chance entity events.)

C1ffisme commented 9 years ago

Just a thought to ponder, when I play minecraft, I tend to find it fun to play in survival when I'm alone. But when I start an LAN server, I find it fun to play with friends. Why?

Why is it fun to play survival alone?

Why is it fun to play creative with others?

Also, I play minecraft a lot, but sometimes I feel like playing a strategy game like Age of Empires. So I play that and then after only a couple days I want to go back to Minecraft. Why? Age of Empires has strategy and such, but Minecraft has creativity. So one possible development direction is to make this a voxel strategy game! You could lead armies, build cool buildings, and all that!

Lymkwi commented 9 years ago

Just a thought to ponder, when I play minecraft,

It seems to me like all your arguments are based on your minecraft experience... So, just remember that Minetest != MineCraft. Even the community and the players are different..

kilbith commented 9 years ago

@paramat

We could add savanna with Acacia trees and the red wood items (see watershed/paragenv7/riverdev)

Seeing this video, I'm all for this biome in MTG.

paramat commented 9 years ago

MTG team let me know about savanna, now is the time to add it while i'm still working on the biome system. I'm going to sink some rainforest into water to create occasional swamps too.

C1ffisme commented 9 years ago

@LeMagnesium Minetest != Minecraft, but we unfortunately are ~= (partially equal) to Minecraft. We are based off of it. We are a clone of it.

Just not pulling features from minecraft is trying to cheaply deny the fact that we are a clone. We have just about the same basic type of game. So why don't we at least try to copy it without directly copying it. We don't have to make ender dragons and redstone, but if there is one thing we need to copy from minecraft, it's fun.

Because the truth is, I haven't played minetest for a while. It's hard to get into a world without quitting. It's just not fun. If we don't know how to make a fun game without any copying, then copy.

Also, we technically aren't copying minecraft by adding a goal. We are making the game into an actual, well, game.

And minecraft has been most of my life for the past few months, so I don't really have anything else to compare against. If I had been playing Call of Duty for the past few months, my suggestions would include guns and violence.

nanepiwo commented 9 years ago

@C1ffisme Minetest is not based off of Minecraft. However, they are relatively similar. The question is whether the dev team want to make Minetest more for the mass market of kids (who usually can't pay for Minecraft) or as a fun game that the developers of the Minetest community like to play and code :P

And as a note, there is an achievements mod which I think would be great for a goal added in Minetest.

On Sun, Jun 7, 2015 at 9:16 PM, C1ffisme notifications@github.com wrote:

@LeMagnesium https://github.com/LeMagnesium Minetest != Minecraft, but we unfortunately are ~= (partially equal) to Minecraft. We are based off of it. We are a clone of it.

Just not pulling features from minecraft is trying to cheaply deny the fact that we are a clone. We have just about the same basic level of game. So why don't we at least try to copy it without directly copying it. We don't have to make ender dragons and redstone, but if there is one thing we need to copy from minecraft, it's fun.

Because the truth is, I haven't played minetest for a while. It's hard to get into a world without quitting. It's just not fun. If we don't know how to make a fun game without any copying, then copy.

Also, we technically aren't copying minecraft by adding a goal. We are making the game into an actual, well, game.

And minecraft has been most of my life for the past few months, so I don't really have anything else to compare against. If I had been playing Call of Duty for the past few months, my suggestions would include guns and violence.

— Reply to this email directly or view it on GitHub https://github.com/minetest/minetest_game/issues/515#issuecomment-109820200 .

C1ffisme commented 9 years ago

@nanepiwo :

Minetest.net:

Minetest is an infinite-world block sandbox game and a game engine, inspired by InfiniMiner, Minecraft and the like.

Yes! An achievements mod! That would be great. (Though we still need the features to have them... :P)

paramat commented 9 years ago

I'm adding savanna to the biome system, thanks for the encouragement, with more research on biomes i discovered it is essential for the biome system and cannot be left out. It is also beautiful and the alternative surface node adds refreshing variety. The only terrestrial biome missing now is chaparral, but that is a minor desert-edge biome that could be created with existing nodes.

nanepiwo commented 9 years ago

Personally, I love acacia, and new trees are always better. Adding pine trees and snow biomes was a very good idea, so I don't see why a savanna would be a problem.

@C1ffisme https://github.com/rubenwardy/awards

RHRhino commented 9 years ago

BTW we have currently 3 types of wood which are very light - a darker one would be usefull too.

asl97 commented 9 years ago

@nanepiwo that awards mod is limited to giving item iirc

it doesn't have an api which let us define a callback to run when someone achieve a goal.

we couldn't do something fancy like using spawn_falling_node to make node like steelblock, copperblock, goldblock, mese, diamondblock fall from the sky one after another making a fancy pillar for the player to dig up

or award people with currency from a mod like money2

(ps: i got that fancy pillar idea from lucky blocks if you know what i mean)

C1ffisme commented 9 years ago

@paramat I hope they get added!

@asl97 Yeah, but code can always be edited.

rubenwardy commented 9 years ago

@nanepiwo that awards mod is limited to giving item iirc

it doesn't have an api which let us define a callback to run when someone achieve a goal.

Do you mean something like awards.register_on_unlock() and on_unlock() in the def table? That would be easy to add. The awards mod is now LGPL 2.1 and CC-BY-SA, so could be included if people wanted it.

(I created the awards mod)

kilbith commented 9 years ago

@paramat

Hope this helps for the Acacia textures... acacia_prev

Raw file : acacia

rubenwardy commented 9 years ago

@nanepiwo: I've added register_on_unlock and on_unlock to awards, 0.6 should be released in about a week maybe. (It's exam period so I'm slower than usual)

However, I don't think awards would fit Minetest Game, as it it a base for modding rather than a game as such.

rubenwardy commented 9 years ago

@est31

We'll should chose (or write) one API with good design, and then all mobs mods have to use that.

xkcd 927

C1ffisme commented 9 years ago

@rubenwardy LOL.

nanepiwo commented 9 years ago

@kilbith I personally like the acacia tree textures in more trees and paragenv7, as I like having a red wood to build with.

paramat commented 9 years ago

Thanks kilbith but i will be using the moretrees textures and am working on the other textures needed.

C1ffisme commented 9 years ago

Maybe it's not quite the same subject, but what are we planning to do for 0.4.13? I suppose our development direction includes the next version.

Should we add new blocks? A new crop? A new biome (savanna)?

I think that it would be nice if we could have something cool to use in 0.4.13. After all, we haven't done a new version for at least 3 months.

est31 commented 9 years ago

@C1ffisme could you open an extra issue for that?

C1ffisme commented 9 years ago

Sure... Eventually. I'm learning HTML right now.

rubenwardy commented 9 years ago

Versions don't matter. We shouldn't be driven to add new blocks / content every release - it's silly and ends up with a bloated and bad game.

kilbith commented 9 years ago

@paramat I don't like this ground (too clear). I prefer that for the savanna :

5hwst

paramat commented 9 years ago

'Too clear' as in not enough long grasses? or the ground texture is lacking contrast? I agree more long grasses are needed. I might also make the colour of long grasses closer to that of ground.

kilbith commented 9 years ago

I was more reffering to the ground grass, the tall grass looks good. But the ground may not be faithful enough and not "wild" feeling.

paramat commented 9 years ago

I agree, i have increased the contrast and made it a little darker.

paramat commented 9 years ago

Savanna and rainforest swamp are added to biomesdev. I will now be adding another biome similar to the Antartic, thick icesheet/glacier on stone under a dust of snowblocks, with thick iceshelf on the ocean. Tundra will therefore lose its ocean iceshelf. We will soon be up to 10 biomes.