/agdg/ + /vm/ ~ Amateur Gamedev General

You mean Tanks of Freedom or something else?

Find it yourself:
archive.8ch.net/#!search/1/advance wars/undefined

...

Calm down you turbo-sperg. Jeez.

you win this time…

(checked)

This is what autism looks like.

Not really, I'm talking about a game IN DEVELOPMENT.
Not some gay ass indie garbage. We are talking about freeware here, nigger.

I just made one of those posts

...

Thanks for the (you), my friend. :^)

Whats the difference, in today's world?

yo, MoM user here, with my last weapon analysis video. I have one more analysis video (unreleated to weapons after this), before having nothing to show until the first demo is done (which is soon!). This week I am showing off the Dark Tome and its Dark Magic!

Basically, I will be quiet for awhile after next week I imagine. Any questions are welcome!

one costs money (piracy aside), the other doesn't

well, I'm a big faggot who can't figure out the AStar class in godot
I'm using an AStar node, and I already added my points and connected them, the problem is actually using the functions of the class because I have to translate locations of the tank and base into ids, so I made code related function.
func makePathBase(position): position = Vector3(position.x, position.y, 0) var baseId var enemyId var basePosition = Vector3(get_node("base").get_global_pos().x, get_node("base").get_global_pos().y, 0) for i in range(0, columns * rows, 1): if baseId != null and enemyId != null: break if graph.get_point_pos(i) == position: enemyId = i if graph.get_point_pos(i) == basePosition: baseId = i print(baseId) return graph.get_point_path(enemyId, baseId)
It receives the location of the enemy(the one who'll be getting the path and is calling the function) and turns it into a Vector3 as AStar works with those whereas the get_pos() function returns a Vector2, then it checks position every point represents and finally uses said ids to get a path and return it to the tank that called the function, but it just hangs, and all this procedure seems pretty wasteful as well, I don't know how fast it'd be in-game when i have to switch targets and find a new path

Tanks of Freedom is free…

Don't have any questions, I just want to say that your mod looks awesome, and I can't wait to play the demo

Oh thank you, thats always nice to hear!

Just to give everyone a time scope, I just gotta finish a new more maps. Infact, thats pretty much all I have to do with the game is maps at this point!

the duck paid my rent

You gonna sell the game? I would pay for it

You guys ever find a game you didn't know existed that does what you wanted to gameplay wise pretty much perfectly? I wanted to make a fire emblem esque (s/trpg) game with commanders and morale in the game, with the commanders giving a buff to their troops while they were in their vicinity. Maybe give the commanders a inbattle active skill to buff their troops too, but when they die their soldiers get weakened. I also wanted to let you be able to choose different factions to play as.
Brigandine was pretty much what I was hoping to make, bar some minor changes. It's still amazing to me that devs back then were able to do what most wouldn't even consider trying today. This battle system would have been perfect for SMTxFE too, with the player characters from the games being the knights and the demons being monsters they controlled. The FE characters would mostly be better fighters while the SMT would be getter commanders I'm still mad about what happened

It will be free!

aw fuck, I had something similar in mind.

One is done for fun, one is done lazily for some easy cash, you dumb nigger.

I hope you finally release this shit nigger. You've been saying the demo would come out for almost a year now.

It still ain't Tanks of Freedom. It was WIP, it didn't even have terrain graphics, they where just colored tiles.

But cash is fun user!

LOOK NIGGER.
TANKS OF FREEDOM IS FREE AS IN FREEDOM AND FREE AS IN FREE BEER.
THE FUCKING SOURCE CODE IS AVAILABLE FOR DOWNLOAD.
FUCKING KILL YOURSELF.

Tanks of freedom is fucking gay. I would much rather play Advance Wars Dark Conflict on an emulator.

...

Look, if you are not going to post the clone I was talking about (very early WIP, graphics where very simple, as in placeholders) shut the fuck up before I ram my fist up your arse.

I know, I keep thinking I am close to it. The issue is, I keep refining the game. If you look at videos of the original gameplay videos I showed, small things such as added recoil, rotating gibs etc I added, including things such as summons and what not, adding secondary ammo types to weapons. I am at a point where I dont think I can add anymore.

Like, I will tell you exactly what is left to do; The maps, and the bosses. Each Major area of the game has 3 sections; The Main Area, the Village, and the Dungeon. They is 6 Major Areas in the game, and considering I am actually FOCUSING on that aspect now, the game actually being 100% done is not that far off.

I only have to make 2-3 more maps, and the demo is done. After that, I have to do roughly 12-15 more maps, and the game is done, along with the 12ish bosses.

BASICALLY, all thats left is the map and bosses. I AM hoping I can be done in 6 months, BUT who knows.

Ok lad, you might be a homo, but you have all my support. Make us lads proud.

When you are unwrapping props that are long and tall and can't fit squarely on a square texture, is it okay if you stretch it or do you just slap it on proportionally and leave off a section of the texture unused? Would that be a waste of memory? I could squeeze more detail if so but I also don't mind having to resize whatever details I add to it.

Unless you are going for a super serious and super big videogame, I don't think you need that much optimization for it.

So it's not that much of a waste of memory? I'd be passing it to an engine like UE4 and I assume the compression will take care of the big unused black square.

blegh but i guess there's nothing else to put there

I'd separately unwrap the indentations from the rest of the door. kinda a bad idea to not put a seam where there's a sharp edge, shit doesnt bake too well, especially if the island is stretched horizontally like that :\

Im working on my game everyday, feels good to follow the dream

Finished the Log System

The only thing i didn't like was the void effect on the enemies, they kept bouncing like fish, looks a bit weird

I don't like it either but now that I think about it I need to space to detail a keyhole and stuff.

Shouldn't matter if those indentation faces have a different textured color.

Rustfag here. Designing the building system - probably the first time I've had this much trouble with figuring out a game feature.

We need a high variety of possible ways to build with a smallest possible amount of tile types without locking ourselves to a grid. Construction parts will lock to each other, sure, but they can be placed out of alignment, as well - if need be. It's a tough issue. Doesn't help that it's practically untread ground - most games that include building stick to a grid and a bunch of blocks. We'd like a system to feel "free", like Rust, but without the horrifyingly large building parts.

Any thoughts, suggestions, or guides that you guys can give? I tried looking for autistic niche documents for toymakers but frankly I'm afraid this kind of shit might be a tad esoteric.

Having a very fine grid?

I'm surprised you'd try to go with building as an important mechanic when there's already hundreds of games that do it in pretty much every possible way, and all of them really disappointingly.

Honestly unless you already have the best idea and the whole system built out in your head, it will just turn out a mediocre copy of an already existing system.

I'd personally spend my time working on some other feature rather than spend a lot of time trying to replicate something that never works well, but that's just me.

As far as suggestions, dunno. something that hasn't really been done is having premade blueprints for entire parts of a house and being able to connect them without following a grid. But it sounds like a lot of work and I still don't think it'd be worth it.

Unfortunately thats how Zdoom handles sucking in things on an Z Axis

That would be in dissonance with the deformable terrain - it's off the grid so I'd prefer not to have this kind of shit. I'd also like to have walls set at various angles to each other, and the grid goes against it.

That's the thing, though - I fucking LOVE good building. Minecraft is the closest thing that came to it, and it gave me a taste of what it COULD be. Ever since I was hellbent on making a proper building system into reality.

Brushes
developer.valvesoftware.com/wiki/Brush
and

It's kinda sad to think that the most perfect refinement of everything good in a survival crafting building game could have just come out of Early Access and I'd literally never find out about it.

This is a building system that must be intuitive to a player, so it cannot be something vague. We're talking about a building game here - it must have a low skill floor to allow newbies to build a good looking hut on the first try.

Anyhow, I'll let you guys think about it if you so happen to be interested enough to give it actual thought, and in the meantime…

I posted this in the last thread, but it was nearing the bump limit, so here goes - I've made a UI mockup for the game, what do you guys think? Any criticisms?

Keep in mind that it's a mainly building/resource management game with elements of lovecraftian horror and frequent PvP.

Look into Planet Coaster's Building system. Each building has its own grid, but you can build off the grid as well.

Would it be bad to force my Players to use a gamepad

Yes. It reduces your potential playerbase.

Gamepads are also inferior to M+K

Brigandine was so fucking fun, i loved that game
Im going to do something like that as my next game

it'd be fucking terrible

Depends on the type of game. Character-action games practically require it.

I want to make game but I dunno how to make models and animations and shit. What's that about? I'm looking to make a resident evil clone but I don't even know if I have the sheer artistic autism to make such detailed pre rendered backgrounds. Maybe I should just stick to rendering simple shit. Shouldn't be too bad if I stick to low polygons.

I'm going to use Unity because I'm retarded and don't know any better. Hopefully this can springboard me to a better engine once I'm done with this garbage.

I'm here to the fag that said kamiya/taro stole his work.
Where is he?

I beat Cavestory+ on hard with KB only.
La Mulana and Binding of Isaac: Rebirth, too.

Character-action as in Bayonetta and MGR: R

If you don't know jack shit about anything that goes into gamedev, and your first idea to make a game is to create a "RE clone" you're in for a disappointment. If you don't have tangible practical skills needed to create a game, you need to have strong passion and understanding of game design to actually make moves in gamedev.

You really cannot go in and expect to "just like make gaem".

How the fuck do you know if Unity is bad or good if you don't know anything about gamedev?

Spectacle fighter*

Start small or you're gonna crash and burn, keep it simple: 2D, shit art, shit music, short length, easy to play.

does anyone have knowledge of advanced or higher intermediate level blender animation tutorials?
I know plenty of the basics but I have some random problems I really dont know how to troubleshoot.

Things like strange jerks in the animation when there's no keyframes on any of the bones or discrepancies in the F-curves.

What annoys me most about being a nodev is all of my problems end up as technical hold ups that I cant solve.

Like what?

Depends on the game. 2D platformer, why the fuck wouldn't support keyboard input? 3D action game, yeah it's probably better to require a gamepad than make people pissed at subpar keyboard controls. Either way you can worry about this when the game is further along. Focus on the fundamental stuff now, the tedious stuff can come later.

Spectaticle

Resident Evil is a very simplistic game. You move around on a plane, pick shit up, use it at designated use spots and has some pretty basic combat. I mean, I ain't trying to make some open world do everything game here. If you can't even do that then what's the draw of making fucking space invaders a thousand times? I've been making mods in Quake C and honestly I don't think my goals are too lofty.

Is that seriously where everyone has to start? Did everyone who claims to have learned on Unity just lie about jumping into 3D or something? I've made some falling rocks games and some mazes on an 8x8 led matrix and a microcontroller, can I play with the big boys yet?

For a while it was bone constraints.
Then it was blueprints in UE4 (still having trouble there)
now its animation. Which I can cheese and make it look passable but I want to learn to do it the right way.
Also things like directions on specific ad-dons like box cutter.

Maybe I suck at researching and/or learning but I can search for hours and only find the super entry level basic stuff.

If you want to start with 3D is alright, my main suggestion is just to not have high expectations.

I think you'll need to use addforce. Unity activates all of its physics when rigidbody is used I believe and they advice you to use it when moving an object.

I think using position and rotation with rigidbodies messes with the collision prediction.

docs.unity3d.com/Manual/class-Rigidbody.html

Bruh, I'm gonna make a masterpiece first try. You're all gonna know my name. I'm fucking Miyamoto and Ulililia's satan spawn.

Real talk though, would you suggest I just drop into game maker first for 2D? I heard the Unity 2D is a bit janky. As a guy coming off of some college C and Arduino courses, I'd like to dip my feet further into actual coding. I heard GM actually has a decent scripting language.

I LIKE THAT ATTITUDE KIDDO

Unity is a 3D engine first and foremost.
2D is a set of features tacked on to appease the people demanding such features.
GM has a custom language, and is from what I hear an amalgamation of python, and java.
Unity uses standard C#, and would be - for me - the preferable choice for someone coming from a C background.
You could always check out the open source engines. One I'm going to try in the future is xenko (similar deal as Unity, but open source).

Just constrain all movement to a 2d plane and use an orthogonal perspective lol

I never said it wasn't possible, and I've actually done something similar for a demo project.
Though, again, it's a 3D engine first and foremost.
Let me point it out to you, as you don't seem to get the implications: 3D users will always be first class citizens, and 2D users are second class citizens.
This is true for new features, fixing features, and all that.

I started in Unity on a 2d project with minimal prior experience and haven't run into any major issues.

I say go for it.

There's a game I really wanna make, but I know it's way too big in scope for a first game. Since I don't want it to be full of spaghetti/Yandev-tier code and shitty game design, I wanna make a small game just as a test run, but I've just got no inspiration. What's a good kind of game to make as a first?

What kind of game do you want to eventually make? The small project should help you work your way towards the larger project.

A 20 hour JRPG. RPG Maker is absolutely out of the question. I was thinking about maybe doing a dungeon crawler as a first, but that also might be a bit much.

Programming wise this is not terribly difficult as long as its in 2D.

Why don't you make a smaller RPG, then look at all the things you did wrong?

I'm assuming you want this in 2d
List of games to help you build up to what you want
Pong Clone
Tetris Clone
Simple Twin Stick Shooter
Rougelike
Small RPG

Put your skill points into charisma and pull together a competent team.

Really? I'd assume it'd be fairly difficult with shit like damage calculations and turn order and all that.

Considered doing a 4-hour thing, I've just got no inspiration plot-wise.

Well shit, how hard could this possibly be right?

Some things you should decide that will help you going forward…

On the overworld:

In battle:

You'll also need to throw together a dialogue system. It's not too hard… There are plenty of tutorials out there.

I was thinking 8 direction movement and enemies on the screen, but I want the battle system to be like FFX with more emphasis on fucking around with turn order, like what Tidus typically did.

(checked)
Read more books and short stories
Just don't get stuck on the rougelike. It's happened more often than you may think.

Definitely go with this
Smaller projects help you maintain scope, and to not overwhelm yourself (as this happened to me a lot when I first started); in addition to being good learning material so you won't be afraid to make abundant mistakes + experiment.
As it's best to mistakes early, quickly, and often; so later you'll be less prone to said mistakes.

Definitely write it down into a succinct design doc, get a clear idea without overdoing the notes, and then go for it.
As shit changes rapidly, and what you initially planned on is never what you end up with.

This is the logo I made for my game's map editor, which i'm working on. I cant begin building the rendering engine without some basic maps to render.

Just do it to learn how to make games, it doesn't have to be original or anything. You'll have plenty of time to think about a plot.

Its not a big challenge to write an RPG system, its in the "beginners projects" category of programming, basically once you start getting a handle on programming you will realize that an RPG system is most definitely within your grasp.

This makes it easier because you won't have to worry about collisions at all (either coding your own or fucking with the built-in physics engine). You're going to need to create some sort of tilemap… A two dimensional array works well… and have it hold a custom class that has data for that tile… Whether you can walk on it or interact with it or whatever you want it to do.

With movement confined to the grid, this could make things a little wonky, but it's certainly doable. Will enemies only move when the player moves, or will they move constantly? If the latter, how can the player avoid undesired encounters? How do the enemies "notice" when the player is in range? How far will they chase him? All things you'll need to decide. And of course, you'll need to look into pathfinding to actually have them follow, but again, grid based makes this easier.

Haven't actually played it, but presumably this could be handled by giving each character in combat (or perhaps action taken. Or both) an initiative value and having actions taken/turn order stored in a list, then sorting the list by that initiative value. I'm no sure exactly what you're going for, but nothing you've said sounds like it would be too difficult to pull off.

What kind of 3d modeling software do people use? Zbrush? Blender? 3ds max?

blender here. But I have to say that I mainly do animation

anim8or.com/

as well as the Sigma Map Editor (vid related)

I am wondering if individuals are really paying autodesk $180 dollars per month for Maya or 3ds Max. I have been using blender too. Fun stuff.

The individuals either pirate it or buy it lifetime. The latter are generally massive corporations.

Can your game do this? I bet it cant. X11 can do this without any issues, though

I warned you about the longjmp, dude.

I modified SDL to allow window dragging.

SDL's main problem is that its written for single threaded applications. You can't have one thread fetch input, and have another thread write to the window. This is because SDL actually resizes the window during event fetching.

My solution was to modify SDL, and basically give every window its own built-in mutex that can be locked/unlocked. Event handling was adjusted to lock this mutex before modifying the actual underlying window.

I'll probably phase out SDL entirely at a later point though since it is pretty shit for multi threading. Not to mention some awful decisions made for the sake of compatibility.

long jumps are fine, normally, its just that it was being called from some kind of windows lib

I forgot that this problem can be fixed with multi-threading when writing the post. I'm only on one thread still.

That's pretty bad. I can only think of one thing that is on that level.
>friend is trying my game spinning cube out

And I literally told you, in a previous thread, that longjmping through another library is not allowed, since it does not do stack unrolling.

It's not that bad if you use SDL in single threading only. In that case, SDL falls in the "just werks" category.

The resize is a thing that kinda has to happen, since Windows' function for changing a window's position (which SDL does for you when you drag around a window) is also the function that allows resizing to happen.
As long as the program is limited to a single thread, this behavior is pretty ideal. It means that users of SDL don't have to worry at all about windows being moved around. SDL already translates all mouse coordinates into window coordinates, so the position of the window does not matter.

sorry, I forgot

I'm using GameMaker.

I created an object called 'obj_droptrap'. It has the same sprite as a floor tile and when the player steps on it, it spawns a bunch of enemies. I can get them to spawn.

For the object I have these variables in the create event.

enemy = 0;enemy_num = 0;spawn_y = 0;spawn_x = 0;

Then in the creation code of the instance I change the enemy to whatever enemy I want, in this case 'obj_blob', change the number to spawn, then set the other two variables to how far I want them to spawn.

Then for the collision event I have this.

while enemy_num >= 1 { instance_create(irandom_range(x - spawn_x, x + spawn_x), irandom_range(y - spawn_y, y + spawn_y),obj_blob) enemy_num -= 1 }

It works fine except for one thing. My game is tile based. So when you move it jumps 32 pixels in that direction. It's kind of like the original Ultima game. So what I need to do is have it so the enemies spawn randomly but only randomly every 32 pixels from the origin of the trap.

Anyone know how I can do this?

anyone has any advise or guides on low poly modeling, im trying to make an rts and i want to try working with an aesthetic

from what I've been told you usually go

Unless you're a genius of 3D modeling or an amazing artist that's kind of your only option

P—please cease your heresy at once.

Choose between this
if (op == "faggot") { ban_user(op);}
And this
if (op == "faggot"){ ban_user(op);}

Generate a number between 0 and N, then multiply by your grid size (32). You can use decimal numbers with the RNG but round down after scaling it

Or better yet ignore the grid system and logically set the XY position yourself to be in grid units and on the draw event do the 32 scale

thanks

seems so simple now i dont know how i didnt get it on my own

After reading a bit through Unreal Engine just for fun, i only now found out that you can actually convert blueprints into C++. Anyone has experience with this? Does this mean that some cucks can use copypaste visual spaghetti to compile C++ code that's actually faster than C#? Or is there some drawback?

Never made a game before
I'm planning on making the underground chinchirorin from kaiji s2.

Lying is not nice user.

I'M NOT LYING, HOW DARE YOU

Very nice.

When I messed around with unity I personally used vectors and forces, yeah.

I'm currently halfassedly learning how to into unreal, and the eventual goal is a climb n kill.
If you remember pic related being in the catalog, that was that thread. It also explains the game.

traps aren't gay

Big milestone for me today. I've finished making 200 levels for Speebot.

Faggot

faggot

...

faggot aids-boy being butthurt as usual

Please don't post homophobia on our christian message board.

Yeah, pretty good. Love the Diablo look you went with, even if it's a bit dated, maybe make it transparent with some random old art pic in the background.

Great job so far on the physics, it's a nice touch but I fear it might get in the way when in combat but if it's rare enough it wont be a problem.

...

The best way to handle that is to make smaller projects out of the individual mechanics you want to put in your bigger project. Work your way up.

Quake fans just wanted a prettier Quake Live. What's so hard to understand?

indented braces is god's way of coding

Unfortunately Quake Champs's audience is mostly people outside the Quake community. That's an actual quote by the way. Someone was spamming ESports Reality with those some time after QC was announced.

Blender does everything.
If you lean it right you can make it proffesional.

Start here
youtube.com/user/cgboorman/videos

Actually, he's doing it correctly. The proper ways to indent are either your first example (preferred) or his. Your second example is wrong.
See Steve McConnell CC ch31 for examples and rationale.

Good Job
But I still think you should keep improving the game, adding more levels and mechanics and shit before going to the next one

Fuck C++. If I ever finish my shit and make some shekels, I'm definitely investing in a Ryzen processor to shorten these ridiculous compile times.

When C++ Modules TS is supported compile times will be reduced dramatically in C++, if you take advantage of them in your code user. Should be ready in your local compiler Any Day Now™ :^)
isocpp.org/std/status

C++22
Socal media integration
Github integration
Package manager

only if you can befriend the mook village before blowing them to pieces

That's C++22.5 to you buddy :^)

"Sure, let's see."
"Actually, I'm kind of busy so maybe later."

He's a born 'ideas guy'.

...

Err, there isn't a handler 25.

Due to a microcode bug, compiles can randomly fail(at least with gcc) on ryzen.

So I got a pretty good idea for the building system going - but I need a little tool to test it out quickly without having to fuck with implementation at this point in time.

Do you guys know if there's any tools out there that I can use to construct shapes out of lines, where I can define precise length and angles?

I don't wanna have to resort to fucking blender.

She sent out FB invites to like 20 people, 8 accepted at most, nothing was ever posted, and the one time I talked to her about it was basically just meme anime ideas

Even then years ago I thought about making a Morrowind clone. Now obviously thats a lot of work and way too much scope for one's first game, but she was still there injecting awful anime tropes into my process

If you want to make a game, you need to lock down creative control and micromanage every idea that makes through. Allow leeway if you don't see harm in the ideas of your colleagues, but never back down if you believe that their addition is detrimental.

As the game director, it is your job to direct the process and maintain control.

She was the project leader and got exactly nothing accomplished. She specifically had me critique their ideas for feedback and acted on none of it. Her project lasted about three weeks before it stopped being talked about

If the project leader is retarded, the product will come out malformed and shitty. Take Rust for example.

Everyone should remember this the next time they try to create a team effort project, whether offline or online. I have first hand experience with a project that turned into a trainwreck because nobody was there to say "this is what we'll go with" and tell anyone who doesn't like it to fuck off. Everyone wanted different things and nobody knew what to do.

Doesn't that mostly affect overclocked processors?

As far as I saw, it happened on stock clocks, and the only way to get past it is to cripple your CPU(smt off, u-op cache off, lower clocks, enable llc)
community.amd.com/thread/215773?start=30&tstart=0

Ryzen suffers from the hyperthreading problem like Intel?

They have no idea why it happens. Disabling random features is a way to mitigate it I guess.

The fuck is that? I don't know enough about compilers to call bullshit, but it seriously sounds like something Hollywood would make up.

On windows how do people handle windowing and input? Win32, SDL, or GLFW? I am really just interested in the easiest to use solution. win32 is a pain.

SFML

what the fuck
HUHUHUHUHUHUHUHUHUHUHU

CPUs have what's called a microcode, which is a very simple set of instructions that allows for the reprogramming of CPU logic, to make it easier when things like this happen.
A microcode bug is when this code contains some sort of bug that impedes the intended operation of the CPU.

HUUUUUUUUUUUUUU

Turn the h upside down to reveal the hidden message.

oshit

WEW

what the fuck is going on

Do you not see it?

I do see it, I just don't get what the big deal is.

Some bug with filenames >>>/sudo/40749

took me 20 minutes to realize that was a "do" and not "don't"

What is the simplest programming language there is? C seems pretty "simple" but surely there's something even simpler.

Brainfuck, there's only 8 instructions.

what the fuck

I want to try making a game in this

Please don't do this to yourself.

I made an extremely short CYOA (like 12 states) with it back in high school, but that code died along with the harddrive. I did write a script to generate a lot of the code for me though, so it wasn't as elegant or impressive as it could have been.

this. Cinder is also quite easy to use.
libcinder.org

It would have to be a text-only game, as brainfuck only does string IO, there is no C interop, and you can't define or call functions.

If you're going for low-level, assembler language is about the simplest you'll get without just writing raw opcodes.

wtf I love brainf*ck now

Any kind of assembly language is literally the simplest you can get, except esoteric state machines.

just make grass

couldnt you use secondary programs in a different language to interpret that code and apply visual aspects sorta like how you can use texture packs in dwarf fortress or how you can enable AO in older games using nvidia's options

im just an artfag so i might be talking out my ass here

now make blades of grass interact with one another, remove the delay before they start rising again but make the rise slower

Goddamn why is he so mean to that poor anorexic girl?

you are a special kind of sadist, aren't you?

do it

oh and give them all dynamic AO

How hard would to be to make a low-poly shooter like embed related?

I've only done little 2D stuff with game maker, but low poly stuff like this really appeals to me. I imagine a lot of the learning curve would be in modeling and texturing/animating?

You could, but it's a little different. You'd have to either have the brainfuck spit out some sort of IPC language with escapes (or just fully framed messages) that are to be interpreted by another program, or have another program inspect and act upon the running Brainfuck program's running state (by reading its memory, for instance).
Dwarf Fortress is a bit different, because it already operates by loading in image packs. They are mostly small hacks in the way that the game already loads in its graphics.
The NVidia and AMD settings that allow you to override some game settings will usually simply force-toggle an OpenGL setting (such as is done with Antialiasing). With something like Ambient Occlusion, the graphics driver basically injects a shader stage that analyzes your world and then runs an AO pass either before or after the game's own shader passes.

In all these cases, you either run the program wrapped by another program, you are taking advantage of something that the program already does, or you are injecting behavior in an already existing program.

So yes, you could theoretically, but you could not do it with pure brainfuck. You'd have to either wrap it or find some extended brainfuck implementation with some degree of C interop.

Consider the following:
cs.utah.edu/~jsnider/SeniorProj/BSP/default.htm

If you're using an engine its a lot easier though.

Dusk is a terrible example of a low poly shooter. Pick up a Quake/Quake 2/id Tech 3 fork and a BSP editor if you want to make an authentic one and don't fall for the retarded flat shading = low poly meme.

Also, i'll add to this: modern engines just buy a licence to umbra3d. But that tutorial is what old-school FPS games did.

i don't think Dusk looks too bad. The textures aren't just flat shading.

Would using a Quake fork be easier than trying to make this kind of game in Unity, as the devs of Dusk did?

Unless you're an autistic nodev/enginedev stuck in the year 2002, it'll be infinitely easier to make something like that in unity than to mod quake.

Unity is more newfag friendly, but your game would have ridiculous system requirements for a low-poly FPS.

what if I channel my autism in order to optimize the game in Unity? is that even possible?

It's less that every operation is slower (It is slightly slower, but still incredibly more optimized than homegrown trash like the sigma engine) and more that there's a large base price you pay in every category just to get the engine running.

In my defense, I didn't even program in any kind of culling in that engine aside from back-face. That's the only reason i'm slower.

fair enough

I think i'll beat unity with this new engine on the account of implementing culling. I don't even have to beat umbra (unity's culling engine), I just have to be competitive on top of the gains from not doing all the heavyweight things unity has to do.

I still can't visualize how BSP/quadtrees can be used to make spatial checks between points.

Like if you have an entire plane, you divide it in the midpoints or whatever. If you have a point juuuust a little bit to the upper left of the midpoint and a point juuuust to the lower right of the midpoint, they're basically as close as they can get, but you still have to go up to the root node and then back down, which seems inefficient and then do it for every node to guarantee that

there is some explanation of this in part 1:
cs.utah.edu/~jsnider/SeniorProj/BSP1/default.html

The collision detection works like this: With the BSP, you can tell what is in front of and behind each point. If you have point X; you know what polygons are in front of and behind it. Now, consider that to be in solid space, you cannot be in front of any polygons. This means that, if we can tell that X is behind a polygon and there are no polygons that X is in front of that are also behind this polygon, then we know for a fact that X is in "solid space". Since the BSP can be used to determine which polygons are in front of and behind any other polygon, this is how it achieves this. Now, because in the tree, every single time we go deeper, we eliminate half of the polygons. So, eliminating the polygons we need to consider is O(log(n)) complexity.

I'm probably explaining it in an odd manner. But, the crux of it is that:

computers are fucking fast, man

I don't understand your point. Even Unreal 4 uses BSP.

are you a woman? because I mean literally what I said

no
I know, but its totally non-sequitur. I thought you were trying to express something else.

...

thank, umbra sir

>(((cloud based))) lod system
This really only seems helpful to developers of AAA games with millions of polys per scene.

LOL wouldn't you rather have INFINITE DETAIL?

You hit the nail on the head:
umbra3d.com/clients/

This is something that most games use, but isn't talked about, since its just a component of an engine. People might know what Havok physics is since valve really hyped hl2 on that, though.

I completely forgot that even existed, thank you user.

I remember that shit being hyped in an game informer magazine back in 09. Fucking ugly like GTA V's graphics

Thinking about doing these directional ports next. Because of the lighting on them, I can't really be lazy and just flip/rotate them if I want it to look consistent, so they all get individual frames.

Do you think it's worth adding an animation to it, maybe a blinking light or arrow?

Is dat sum Supaplex I see?

SubTerra, which itself is a clone/reimagining of Supaplex

...

For what purpose does -> exist in C?

Why can't you just use dot notation like normally? Is it supposed to have a different meaning somehow? I mean they do the same thing, I don't get it.

-> is a derefrence. . isnt

I have very little C knowledge.

From what I can tell, -> is a pointer of some kind, and . refers to values and properties contained by an object.

-> might be a smart pointer, though

What's the difference?

In both cases you're modifying the property.

Skull of Death!
Old and new

I mean "myobject.thing = 2" and "myobject->thing = 2" do the same thing don't they?

no. think of it like this:

when you have a pointer to an int, you can show the int with *my_ptr and the address with my_ptr.

When you have a pointer to a struct, you show the value of the field with -> and the address of the field with . . You can still use . though:
(*myobject).thing is equal to myobject->thing. Its only that -> is nicer to type.

Oh okay I think I get it now. It's very confusing with all these different ways to write the same thing, like how you can access arrays with *arr or arr[0], and getting the address can be either &arr[0] or just arr or something like that.

It's a great improvement user.

it's a consistency thing is suppose i guess they didn't want the same syntax for accessing members of pointers and value types.
I mean they effectively do the same thing but behind the scenes something else does happen.

...

rate the first draft of the building system pls

Because (*x). is annoying to type.

Hold onto your butts, boys because I have progress for you.

I've got two new subweapons: The Trine Volgue, which is a spread gun style weapon that eventually starts shooting big hard hitting projectiles, and the shatterbomb, which is frag grenade that sends out sharp crystal chunks everywhere. With the bomb also comes a new swordtech: The rocket sword which is self explanatory.

Also featured is a WIP of the Pavise Guard, a fat armored bastard that shoots cannon balls.

The text sound in the beginning is a bit annoying maybe try lower the sound or change to something more subtle but otherwise it looks like fun and the gunshot sounds have a nice kick to them.

I'll be keeping an eye out as this has potential

The only thing I really see possibly causing a problem is the 45, 30, and 15 degree triangles.
The 3.0m 15 degree triangle is about .80385 meters tall. That's far from a round number and may trigger some people's autism where everything has to fit together perfectly.
The hypotenuse of the 45 degree triangle is the length of one of the other sides times the square root of 2, once again, not a round number.
You get the idea.

Please never stop working on this. I've been wanting to play a Konan wad for doom for a long time.
Nice.

That's actually a mancubus replacement. Have a few other random webms.

Hey thanks user. Are the sprites original, or did you take them from random games? Anyway, keep up the good work.

Most are rips, some are edits and frakensprites, but a handful were special made for GMOTA

Would anyone recommend taking classes to learn how to code vidya? I feel like this self-taught thing isn't working as well as I'd like. Money isn't an issue, I have a lot of money saved because I was going to go to an art college, but the only one I actually liked(the only one with professors working in their fields and wasn't overrun with SJWs) was too expensive

Always makes me laugh

and by too expensive, I mean I didn't think it was worth what they were asking

No. They'll take months to teach you less than a week of tutorials.

Alright, will just work harder then

Just use brushes like a normal person. Play with a quake level editor.

here's your (you)

I'm finally making the time to dig through the full Blender codebase. It's something I've been wanting to do for years tbh. Glad I learned about clang-format beforehand. From what I'm seeing so far Ton seemed to have given the many different developers lots of leeway when it comes to style and code formatting. It triggers muh ocd tbh.

Careful of "art universities", as most of these slide by as non-accredited schools.
Which means the diploma is generally recognized as just a piece of paper, credits do not transfer to accredited schools (real universities), and they're only meant to build your portfolio + network (which you can do on your own, and learn how to do with the Internet on your own).

When it comes to coding tutorials look on the paid sites for courses taught by professional teachers (general rule of thumb is, if they're formal, they take it seriously, and at least provide solid information; though teach style varies so that's hit/miss).
I used Lynda.com for "paid" coding tutorials a few years back when I started, but I had a free sub because of my university.

Also a quick FYI, never use paid sites for game engine related tutorials… they all suck as it attracts non-professional teachers (as it's a less mature/formal topic of study compared to pure programming).

meh some minor progress now.

It's a first-person building game, user.

you'll get it eventually.

This is correct. The next thing to understand is pointers to pointers.

It'll close the gap, at least. I suppose we'll cross that bridge once we get to it.

That's a pretty neat SU-100 model, user. And this is a non-dev user looking in with no context. However, it looks a tad bit too high; the frontal part of the tank should be pressed a little tighter. The gun manlet isn't as big either.

this looks like a cgi cover from the PS1 era

Another dumb C question

Why is the incorrect block wrong here? Does the content array get deleted when the function ends or something, unless I allocate the space manually?

Disregard the structure itself, it's not meant to represent anything real.

Isn't it just a local variable to the function, which gets removed as soon as the scope leaves the function call?

Thanks.
The height should be identical almost 1:1 since when I do such kind of models I always use a blueprint for it. Now that you mentioned it the belly part of the tank is a little bit too short and I forget to slope the front belly.
I had trouble making the gun mantle more close to the original one so I chose to simplify that part.
I hope the included screenshot with the blueprint should clear up a few things.

I am not sure if I understand you right but when I scale the "turret" part of it then the back chassis will become visible when viewed from the front.

I wish I had their texturing skill since I am not good at it at all.

Your placement of the pointers is driving me nuts. One side or the other, my man.

That was my guess, but I thought it would work since I'm essentially putting it into mynpc.inventory which then gets sent somewhere else.

That's what caused the array to be filled with garbage even though it sort of works, I think I posted about it a few days ago before I ragequit.

I'll worry about autism triggers after I learn to write something that actually works.

content is not dynamically allocated, so a reference to content[0] after you leave createNPC will be gone.

I know that one guy whom I sent my deusexclone's builds for testing (pic related, people remember this project, I'm sure).

He always had pleb opionions on the game, on how it should be more like call of duty and how my ideas won't sell (I kept telling him I'm making games for fun and not really for profit, which would be a nice addition doe) and how RPG elements should be like skyrim and the inventory is stupid etc. It wasn't anything near the deep discussions we had here. It was petty shit that often indicated he didn't know how games are even made.
I gave up on sending him anything after a while, as he clearly didn't understand the very premise of what I was trying to achieve he only lasted 15 minutes playing Deus Ex, never bothered to play Rogue Spear. Pleb 110%

But he got fixated on one thing: that guy he knew on the internet had a dream of making games in his spare time, had a stable job, posted on /agdg/ etc. So OF COURSE he dropped out of school, moved back home to his toxic parents and did "the same thing" full-time.
With no experience.
And pleb ideas.
But he was gonna be an indie star!

A recipe for disaster, right?

But he did get some dude to program for him. The guy was super pumped and motivated and willing to learn. And my friend actually started working on his modelling skills.
And I even agreed to help with the wriging of what was supposed to be a sort of small-scale Skyrim-clone with "beautiful graphics, original setting", "something noone ever saw before"

Under one condition. He had to prepare a small vertical slice for the project. Something simple, like an intro containing the dialogue system he'd like. Something I could work with.

He spent 2 months bitching about the programmer guy's work. It wasn't what he wanted, it was ugly, it didn't fulfill his vision of a single, pointless detail he was obsessing over at the time.
I saw the guy's work. It was breddy good, actually. He made a nice, little snowy village with nice snow particles and simple AI roaming it.

What did the project lead to in that time?
A rock model.

But it was perfect. It was exactly the way he wanted, fulfilling the impeccable artistic vision. It just wasn't textured.

So after endless nitpicking, the programmer guy left to do his own shit and my friend is now alone being a NEET, still dreaming of making his game.

What you want to do is change the calloc to:
calloc(1, invlength * sizeof(int)); by the way. I guess its just personal preference though.

So, this is the difference between the stack and the heap. This is how scope works:
You have a thing on your CPU called the stack. When you type something like "int x" it gets pushed onto the stack and you can use it. An array like "int content[length]" is also on the stack. When you exit a level of scope, all things pushed onto the stack on that level of scope are popped from the stack and you cant use them anymore. Since "int content[length]" is on the stack, once you exit that scope level, "int content[length]" is deleted from the stack. When you call "calloc", you are not allocating memory on the stack, but you are allocating memory on the heap. Nothing is automatically removed from the heap until you free() it. The only way to refrence things on the heap is through pointers (and if you want such an explanation as to why, it will be even longer) Therefore, the scope you are in does not matter. When you exit the function, mynpc, the return value is placed in the EAX register on your CPU (at least in the stdcall convention), and so it is not on the stack anymore, which is why it is not deleted, even though it is originally allocated onto the stack. Since mynpc.inventory is a pointer to something on the stack, and not the heap, the memory it points to has been deleted when the function is exited. Also, you cannot allocate an array on the stack with a non-static variable line invlength. #define a fixed number to allocate it.

The solution to this is:
#define INV_LENGTH 64typedef struct{ char name[128]; int x, y, inventory[INV_LENGTH];}Npc;...int content[INV_LENGTH];...memcpy(mynpc.inventory,content,INV_LENGTH);

sorry, there was a mistake in my code. You want to use:

memcpy(mynpc.inventory,content,INV_LENGTH * sizeof(int);

Although I would use uint32_t instead of int, here it doesn't matter.

Welp I guess the user here is right that my only option right now is doing a highly detailed/high polygon model, I just found a few screenshots where the Panzer Front devs had a much higher polygon model of those tanks. Well is there any user that cares enough that has a simple high poly to low poly baking example in blender so that I can learn the process myself?

That actually clarifies things A LOT. What about the npc that's being created? I'm not allocating space for it, just the inventory. Will it get deleted too when the function that called createNPC exits? Or "global variables", i.e if I do "int test = 15" outside any function, will it be in the heap or stack and what's the right way to create those?

Not quite. Rather, the space it occupies isn't reserved beyond the end of the function call, so it's likely going to get reused for something else next time a function is called.

No, because you're return the value of it.
Personally, what I'd do is return a malloc'd pointer or pass an empty object reference

void createNPC(Npc* npc, ...) { npc->x = ... npc->y = ...}Npc npc;createNPC(&npc, ...);orNpc* createNPC(...) { Npc mynpc = malloc(sizeof(Npc)); return mynpc;}Npc* npc = createNPC(...);

The struct is being allocated on the stack.

Even though you allocated the npc on the stack, when you type "return mynpc" in your code, it is moved to the eax register. (you know about registers on the CPU, right?) Thus, when you return from your function, you can still access mynpc through that register, even though it isn't in your scope anymore. Now, since you don't actually interact with the registers, the compiler will do this for you. But, that's what is happening if you say: Npc new_npc = createNPC(…) . If you allocate the space for the inventory on the heap, you need to manually free this space before you stop using the Npc struct. Since your Npc struct is on the stack, it will automatically be purged at some point, you dont need to delete the struct, but you do need to call free(npc.inventory). If you use the solution I wrote in which the inventory is an array on the stack, this is not necessary.

Global variables are not on the heap, I know that there are several ways of handling global variables depending on what OS, compiler, executable format you use, so really the implementation is not a concern, you just need to know that they are always in scope and get deleted when the program is killed. If you did "int *test" and calloc'd it, then the pointer to the heap would be a global, but the actual information obtained from derefrencing test would be on the heap and would need to be free'd.

I have to wonder. Why limit yourself to C?

Why not C++ and use shared_ptr or unique_ptr?

Okay I get it now, thanks a ton. I should have just asked from the beginning.

I guess one of the downsides of learning yourself from tutorials and books is that you can't ask questions from the book if you're confused, so information just keeps piling up on top of already incomplete understanding.

Many reasons, mostly I just feel that I'd learn to hate C++ in the long run. I get the picture that it's bloated with features that don't really need to exist, and the people who advocate for C++ and thus teach it often teach you "the C++ way" of making software even if it's backwards in terms of how computers actually work. That's just my feeling as a bystander.

I'll look into other languages if I ever end up feeling limited by C.

C shill pls… Leave this ancient language and start using Rust. C is unsafe garbage, how can this retardness be tolerated in this thread ? Off yourself my man. Reported. See you around, bucko

Am I in Holla Forums?

It looks cute. It looks like a PS1 port for Flames of War.

Honestly, being able to create and just pass an object without worrying about pointers, or malloc()/free()ing all over the place. It may seem trivial, until you accidentally free() something somewhere you shouldn't have. This becomes exponentially harder as your project grows in both code size and complexity.

I rarely use 'unsafe' features considering the 80/20 rule. Premature optimization isn't needed. Though I'm not advocating iterating through hashmaps or sorting linked lists…

tl;dr:
* C++ is truly a hack of a language, some argue dangerous practices like the quacks they are
* There's little to none performance gain when using C over C++, optimization should be done where-needed and C++ offers the same kinds of optimization features.
* Being able to use SAFER features saves you debugging headaches and allows easier building of complexity
* Write your OOP code as generic as you need, but no more generic. (Over-generalization leads to scaffolding in your code that you're NEVER GONNA USE. Build it with eventual generalization in mind, sure, but that's not the primary objective.)

Just don't go overboard and singleton everything and you oughta be just fine. Feel free to merely use C++ to allow you to pass objects from functions and forgo with the memory management. There ain't no wrong answer in that case…

Pretty sure each and every feature of C++ was important to someone on the Committee at some point. And like C, C++ is very aware of 'how computers actually work'. C++ simply defaults to the premise that in the typical case using well-made libraries are the best approach rather than relying on developers to roll their own each time. If you want to open the hood and do it yourself, you always can in C++, even to writing inline ASM patches directly in the code.

I don't get why you guys are writing paragraphs about how he should use C++. The guy just asked for some help understanding pointers, its pretty bad to create such an environment where someone cant ask for help learning something like that without being confronted about not using the C++ STL.

Yeah, sorry about that. I just don't see the use of writing a videogame in pure C if there's so much shit in C++ to make your life easier (like being able to just pass a god damn POD from a function.)

You might be right. Just explain the man pointers and carry on. Don't switch languages until it becomes a problem.

I really don't have the time or energy to explain pointers to the guy, I really am sorry but I'll leave this to someone else. (You, perhaps?)

Because it appears that the things that are giving him the most trouble in C would not be giving him trouble with C++, and because his stated apprehension of C++ are unfounded.

C++ is not a perfect language, but if you are having trouble managing your memory properly in C, you should probably use a language with stronger memory guarantees. Granted, that still won't save you from having to learn properly how the stack and heap work and how to work with both as well as stack frames. If anything, C++ forces you to learn these, as using shared_ptr and the like necessitate a strong understanding of pointers and the heap and stack (as the reference counted pointers live on the stack but point to a heap object, unless you're doing something really hacky).

You're also misrepresenting the conversation, The guy asked about pointers, somebody asked why he wasn't using C++, he responded, and the conversation went from there. Nobody just dumped C++ on him immediately after his initial question, they dropped in the C++ explanation as a response to his reasoning for not using it. I don't necessarily think C++ is better for him, but there's nothing wrong with answering the question at hand as well as proposing another approach, as long as it's presented as a suggestion.

I think pointers themselves are very simple to understand. If I'm ever been confused by anything then it's the syntax for using them. Like I'm not sure why (*thing).stuff is correct but *thing.stuff isn't.

I also didn't (until now) understand the difference between stack and heap.

Just because I'm struggling to understand in the beginning isn't a valid reason to completely switch languages. Chances are I'd have even more questions since there's more to C++ and it's more abstracted.

Whats a POD?

I posted a few paragraphs about it earlier in the thread, so its probably OK now.

I honestly cant even read my friends C++17 code sometimes, I think that C++ has some nice convenience features, but I also think it has a lot of stuff that makes it more complicated to read.

The questions are being asked because he doesn't know C. I don't agree with this kind of response because it is discouraging people from learning about memory. I think its more valid when someone understands memory management already, and wants an easier solution.

I can see that, reading it again, so, I agree that maybe I shouldn't have made such a response in the first place.

Any game I've seen done in C ends up using a million lines of code to do simple shit. I can't imagine living like that.

Plain-Old-Data
It's a specific type of type that offers specific semantics and guarantees: en.cppreference.com/w/cpp/concept/PODType
It's not really relevant, though, as all C structs are essentially POD types and allow most of the same guarantees, and can also simply be returned from a function with many of the same optimizations (GCC, for instance, will usually allocate the struct on the stack before pushing the function stack frame, and then just access that directly for a no-copy return). You don't usually have to worry about what is and isn't a POD type unless you're digging into semantics of object construction, deconstruction, copying, and move semantics. Following old C++ Rule of Three and new C++ Rule of Five will essentially make it so you never have to worry about what is and isn't a POD type.

And C++'s biggest weakness is its backwards compatibility. Modern C++ is really nice, but it is essentially a completely new language in a lot of ways, and the old C++ cruft hanging around does nothing but hurt it. That's the biggest complication with the language now.

You're right about learning about memory. There isn't a reason to not learn how program memory work, and much of what you'd need to know about C you'd also need to know about C++. C++ makes it harder to fuck it up, though, which is the point I meant to make. If you use your RAII and everything else right, you won't leak memory or access garbage, which is why I wouldn't necessarily recommend switching to C++ for him. C++ won't fix "I don't know something about C". C++ will fix "I know C and don't like the weak type guarantees or unsafe manual memory management". For somebody at his level, the extreme strong typing of C++ will only end up being a headache, not a help. I would recommend that he stay with C until he's a lot stronger, and then probably look into C++ as an experiment later.

Order of operations. *thing.stuff is equivalent to *(thing.stuff).
You can make your life easier with the dereference operation. thing->stuff is the same thing as (*thing).stuff.

C++ isn't much more abstracted, really, it just has a whole lot more nuts and bolts. It's still nearly as low-level as C (other than the vtable and implicit this), but it has shit-tons of rules and semantics. It's more complexity and strictness than abstraction.

You should stay the course. Don't switch off of C because you don't understand something about it. Only switch off if you understand it fully and don't like what you understand, and another language does it in a way you like better.

I'm at 6919 lines of code in my current game, last game was ~7k for the engine and ~7k for the map editor. This one could easily exceed 14k for the engine alone by the time its done. Its not for the faint of heart. 2300 lines are because of the Vulkan API, though. And then 600 lines of OS-specific code as well, instead of using SDL which might be less lines. At least it can do this:

Thats good stuff to read, thanks for sharing it.

I need to stop posting and start working on my new map editor more

Think of it like this:
& is a 'function' that takes an object and produces a pointer
* is a 'function' that takes a pointer and produces the object again.

& is thus the inverse of * and vice-versa.

pic related

Have you guys seen

Yeah, that would be cool if I were a big nintendo fan. Did you know Vulkan has a nintendo extension?: khronos.org/registry/vulkan/specs/1.0-extensions/html/vkspec.html#VK_NN_vi_surface

Using SDL, my game is at about 800, including vector implementation and obj loader. It's not a big game thought, only started it a few days ago. I don't think C games have to inherently be massive. I dunno thought, this is the first game I've made, so I could be very easily wrong.

...

It just depends on how far you want to go I guess. My command line - related code is the size of your game.

Yeah, I'm in the really early stages, I haven't done any cross platform testing or anything. It's basically just a test run, so I have no idea. The thing I dread most is trying to build my game on Windows. I know it's going to be a nightmare.

At least you're using SDL, but I know what you mean. For me, its the opposite, if I have to use linux, im incapable of being productive.

Hello, This is my first time posting here

In high school i was stricken like a dorf, with this idea.

The photos i made in high-school in 2008?, and the following video was my last attempt to make core features.
It currently has no visual or sound polish, so i included some of my doodles for reference.

Effectively i wanted a disoriented free running game through cosmic storms.
The player avatar would be a rayman like set of floating knives bound by force.
The faster it ran, it would look like it was taking slower strides like skating.
It would be able to fling itself after charging, or converting some of its momentum, and use that to crash through energy barriers of certain prerequisites, along with being pulled along bands of energy

I have never played a sonic game and mario galaxy was hot shit to me.

1/?

fuck me wrong video and i'm a fucking newb for not knowing embed eats images

...

Cool shit, reminds me of Zineth.

Parallax is magic.
Procedural Textures are magic.

Everything here is a procedural texture I wrote myself

The slime on the ground is even the same shader as the moon

Impressive, very nice.

welp

i'm not needed

but seriously "hard to look at for long periods without migraines", is not art

look at their other works like the sonic dream collection. surreal as fuck.

For some reason I want to fuck that thing.

...

Alright I think I am getting the hang on the process from High to Low polygon baking, but how do I make sure that the lights is spread equally? Like right now the bottom of the crate is pretty dark and the top from the crate doesn't has any fake bumps detail at all. Or do I need to create several sun light sources instead? Since I am using environment lightning right now.

Are you making Uncharted 5?

I didn't know Final Fantasy XIII devs posted here too.

What's this? Bioshock 4?

I'm going to make a game

i like your enthusiasm!

Survival horror with dice, crafting and mining to be added later

Yeah I guess I am going to do like this from now on, I might make the SU-100 again that way either today or tomorrow depends on my laziness.

How many objects do you want on screen at once where you have to make them that cancerously low-poly?

Doubt it's a poly count thing, at least it shouldn't be. The performance cost of polygons is basically negligible nowadays unless we're talking extremes. The materials applied to those polygons are what really makes the difference.

That's exactly why I ask, because it looks like cancer so I really hope he's got a better excuse than "style"

It depends, I have yet to make several maps to fill those objects with and then doing some basic performance test. How many polygons would you rather prefer for the barrels?
Also I am modding for GZDoom which has only diffuse texturing support.

Ehh at least I bothered texturing those models a bit, which is not the case for hipster trash ""low polygon".

Great explanation of C++ strengths and weaknesses user. Modern C++ is already good and is still getting better.

I think your models would really benefit from a few more polygons.

youtube.com/watch?v=H1oNuKChsdU

I seriously hope you use EASTL or write your own containers user, because I've heard the standard STL implementation isn't very good for gamedev.

quake came out in '96, your models are literally worse than a 2 decade old game

The problem isn't the STL implementations user, as far as I've they are fine. std::vector for example is fundamentally a rather simple idea, and is superior to raw arrays in almost every case.

The problem is developers not realizing that a data oriented design is a necessity for any low-latency engineering design that deals with large sets of data–such as modern gamedev. Keeping the cache lines full is what makes a good data engine. Most developers use an 'array of structs' approach to data (whether as a raw C array or a std::vector) because this is the most natural way to think about objects.

For gamedev however, it's important to use the 'struct of arrays' approach to keep the machine's data busses loaded up, both on and off die.

gameprogrammingpatterns.com/data-locality.html

Regardless of one's philosophy about the C++ Standard Library or it's STL, the simple facts that a) basically everything in C is also valid C++, and b) C++ much stronger typing and memory model guarantees make it a superior choice to using C. There are many other reasons as well.

So how are you supposed to do timers/intervals, like for instance a game tick in C?

What's the approach for making it run as fast as your computer can process it, but not faster than your monitor refresh rate?

take a look at how SDL handles timer events.

GLFW also provides access to high-resolution timer.

Mostly on PC it's an enormous mess and a non-standardized clusterfuck. Access to a real-time hardware interrupt with a customizable period would be pretty nice. Unfortunately you cannot rely on vsync from the video card because gfx card manufacturers are chink retards.

I have tried reading C++, but its convinced me not to use it for a couple reasons;

Just makes it very uncertain to read. There is an "implicit-ness" that I think is harmful.

Hook into your Operating Systems high resolution time API's. On windows this is:
msdn.microsoft.com/en-us/library/windows/desktop/dn553408(v=vs.85).aspx
And on Linux this is:
linux.die.net/man/3/clock_gettime

The timer code that I use is based off the ideas in embed related. But, its really just counting the time that has passed until it exceeds the desired time-step, which is usually 1/60th of a second. The operating system also has features that allow you to query your refresh rate.

Well is it better now? It has 176 polygons now or 408 triangles.

Pouting and tantrums implemented

The dynamite in the middle doesn't need to be a mesh, a texture would be enough. Why are you going for a ps1 artstyle anyways? Nostalgia?

That is much better already.

Adorably tragic.

Im now making the event system, each event is an entry in an external text file for easy modding
You can make an event that causes all lost faeries to return to village in the next day you can even choose if they return alive or are found dead
The event system will be the pillar of easy mods in the Faerie Village, you can even make a state machine with it since events can enable or disable other events

i still don't see why you feel the need to restrict the poly-count at all.

So how many AGDG devs have successfully gotten a game finished and up for sale?

These are my games, but they arent good enough to sell:
ika.itch.io

speedbot dev/Keyreal released a game (first pic related).
store.steampowered.com/app/495230/Hypnorain/
There's also his greenlight page
steamcommunity.com/sharedfiles/filedetails/?id=924909034

In the other agdg related thread there are mentions of anons who released things. Risk of Rain is an example, definitely. So is Gunpoint.

Because sigmadev plugged his thing, I will shill for my itch page:
shodanon.itch.io/
and beelzebox.net/

There is far worse for sale on Steam

Anyone have any ideas on how to make my bowl look less shit? a floral pattern maybe? or some cracks?

Hey cunts, what's the best way to edit MD5's?

I know Blender has a few add-ons for MD5, but you could use IQM which is the replacement for MD5.

Cracks maybe and make it look less shiny.

make it look like clay or terracota.

Thats a pretty good idea, I'll do that. It'll give me a chance to mess around with normal mapping too.

Getting there.

b

u

m

p

Fuck, that's giving me flashbacks to the gambling hall in the yakuza series. Fuck cee-lo, worst rng gambling I've ever played.

I like the aesthetics of it the way it is

This feels like cheating

Post your progress so that you might lead others to make actual progress.

I used to want to make games, at some point, I now find enginedev much more fun.

zawa… zawa…

Thanks, I'll probably revert it back if the terracotta thing doesn't work out.

Thinking about getting into pixel art. Anyone got any recommended programs or advice? Still not sure if I want to go with 2d or 3d with my game, but want to get into something anyway.

If you want to do pixel art then most art programs are fine for it.

Asesprite is specially designed for pixelart though, so you may want to look into that. It also has probably the best animation tools for sprites.

Krita also has tiled mode drawing and animation tools, and isn't limited strictly to pixelart. It also has normalmap painting and stuff if you want to do textures for 3D models or something.

...

Nice TECHNOLOGY

Try sucking cock instead. It takes the same personality but much less skill, and is far more profitable.

Is that a FAMAS?

user, you're the one sucking dicks. No bullying AGDG posters.

can't spell success without SUCC
I know it's normalfag tier but couldn't resist

Good normal maps and specular maps make all the difference for objects that are close to the camera.

Also, a suggestion: make the bowl wobble back and forth a bit when the dice are thrown in (and possibly rotate a bit too). Make the spot light shake around too, as if its hanging on a string. The more stuff you have moving on the screen, the more pleasing it looks. Subtlety is key, try it.

Where can I find decent gun sounds? Not necessarily gun fire but rather everything else, loading a cartridge, chambering it etc.

Preferably ones used in old westerns. I like that crunchy quality.

You could try ripping them from the movies themselves?
They all got blue ray releases so the quality should be good

I could, but it's hard to find exactly what I need. I've looked for reloading sounds but they're either barely audible or there's lots of other noises going on.

Do you have decent recording equipment?
Movies usually use other things to recreate the sounds they want to use

I have no clue what the tricks were for spaghetti westerns, but you could probably figure it out

Look at youtube videos for stuff like firearm tactial training courses (lots of speed reloading, shell casings hitting the ground, etc), speed shooting videos (lots of reloading and revolver sounds), or the plethora of other gun related videos (reviews, generally has reloading sounds, loading sounds, etc).

I don't.

Guess I could do that. Thanks for the suggestion.

Youtube compresses video and audio quality, probably not the best idea

That's some good advice, thanks, I'll try those out.

If you want raw footage then go to a shooting range.
Here in the US you can go2 a shooting range to rent out their firearms, and test them out; or try them for free if you're looking to purchase.

Also the 1080p quality video/audio compression is a negligible loss of quality; as you'll be compressing it in the end anyways.
Just avoid 240p and the other low quality shit.

I'm a yuropoor, unfortunately. Even if I had recording equipment I wouldn't be able to record anything.

Ah gotcha.
Also, forgot to mention that tutorials are a good option too.
Though it's hard to find a point where they just shuddap and do the reloading so u can get the proper sound effects; though it's generally recorded close to the action so it's good for the sound related to the tutorial.

Thanks anons. These look like they could be helpful.

So do you only play games that have 3d models or are you a professional dick inhaler?

...

NEW THREAD: