/agdg/ - Proper OP Edition

Thank you! I will keep tweaking and upgrading graphics but I will keep this color palette. I like fun colors. Some of the current textures I'll probably not change anymore.

It's going to be a comfy platformer without a time limit or any pressure, but the levels will become more challenging as the player progresses. I don't want any frustrating crap like pixel-perfect jumps or tedious time-consuming puzzles, so that's what I'm going to avoid while designing levels. The player has to keep moving (because it's fun) and interacting with enemies and the environment (because that's also fun) and optionally collecting some sort of collectibles (because that's challenging and fun).

That's the general direction I'm going in, but after I'm done with water I'm going to continue prototyping and experimenting with new gameplay elements and enemies. I had an idea of including power-ups that can be picked up to grant double-jump, wall-jump, etc., but the powers get reset when you die or advance to the next level. That could probably result in some interesting levels.

>loggly.com/blog/nine-tips-for-implementing-logging-in-games/

This is something you kind of don't need to worry about. When you build your game for release the compiler will optimize away function calls that don't do anything. When building for debug it doesn't do this because that way you can still step through the program as normal.

It might help if the verbose mode is controlled by a const (so the compiler knows the function will NEVER do anything).

sidenote: the inline keyword does practically nothing, the compiler will determine when to inline/not inline and usually (always?) ignores your suggestion.

meant for

That sounds pretty good to me. One more fun game in the world is another step toward making the world more fun. I shall pray for your success and I hope to play the complete version of it one day, user.

wait a minute
ENHANCE

Holy shit

Any advice for good Yoyo GameMaker tutorials?
Aiming to make an RPG eventually.

Or is following the website's recommend tutorials sufficient?

Meh, close enough.
PS2-tier modelling is the level of my abilities.
Sorry for the lowres, I'm trying to protect my precious concept art which I paid shekels for.

At least you have nice concept art. And PS2 tier models are okay as long as the texture work is on-point. See: Kingdom Hearts HD Remix

MEME MAGIC

fug

Which would look better; if the helmet has a space for the ears inside (left), or if the ears were sticking out to the outside (right)?

It's a cat pls no bully

I would say the left one. Try adding a few small holes for the ears though.

Draw my fursona next Left looks nicer, right one makes it look as if ears were pretty much on the forehead, meanwhile left one while cartoony and unpractical in reality maintains the feline silhouette without some weird perspective.

If anyone needs some help with texturing, I'm going to have to replace my tablet, but I'm really looking for a project to get involved with. I've got a few ideas of my own, but if you've got something going, I'd be glad to help out.

GameMaker is trash, use MonoGame, Godot, or Love2D. GameMaker might be easy to get into but it is a terrible engine from a technical standpoint, costs money, and has a ridiculous EULA.

Godot is easy to get into for newbies too. Try that if that's what you want.

Wanna thank user again for helping me with Dijkstra, in case the other thread is dead by the time he returns.

Only used it because I was recommend it whenever I asked. Then again, there was a Game Maker thread I saw a few weeks later that stank of shilling… Who knows.

Can you go more in-depth between the pros and cons?

Awsome work!
You going to move the skeleton sprites a little lower so they sit inside a square?

Yeah, it's badly positioned, well-noted, they are placeholders so I will not stress with this for now. I have no idea why is this happening, the code looks all ok

I mean just look at the pic I replaced the sprites with balls in their exact position, you see that they are in the perfect center. But maybe I'm doing this wrong, maybe the sprite is not supposed to be in the center.

Maybe the sprites are a bit tall?

Yeah, as I said I'll worry about that when I have an actual artist or learn to draw

random user here, you probably want the sprites to sit at the bottom of the tile so they look like they fill the tile

I was able to generate a pass as well but I wasn't able to make it remember the nodes that you visited.

Are you using Game maker? If so check where the Sprites origin is

Left, I think.

Left. Because the one on the right would seem to hurt the ears.

Guys. I've tried and tried but I cant find anything on what I need.

I think my problem is creating sky boxes. feel free to inform me otherwise.

I'm going to have a level that takes place on one of Saturn's rings. Its more of a joke, as the rings will be like hard disks instead of the field of asteroids that they are.

However I need to create the effect that they are legitimate gigantic rings. How do i bridge between the terrain and the sky box. Should I. and whats the best way to go about this? I'm stumped.

I tried for a while to make a 3d scene in blender then render it to a cube map, but that turned out to be bogus and above my understanding.

send help.

I need to pick a comfy name for this game. Any suggestions?

I explained a bit about the direction of the game here:

The playable character is a robot, which needs to be named too…

I'm looking for graphician (256 colours only). I'm working on a Top down view shooter. Huge maps, fast action and loads of FPS. The engine is coded from scratch in C++ It's in early stages, but a gfx guy would come very handy, I'm not sure how to do certain things, because that'd be up to the graphician to decide. Here's a small webm-demo of the level editor. You can add me on steam : steamcommunity.com/id/amigarules/

What's the best cross platform way to go about looking at the names of files in a directory?
I'd like to just toss files for entities in and have the game find them on its own.

boost::filesystem

something light and cute
it reminds me of chibi-robo a lot
it feels like a game that would have an obscure name
idk if you're okay with something with nip words in it, but i almost feel like it would fit
of course im autistic, so feel free to tell me so

Introduced melee weapons, now that I actually have a textured weapon.

I have to say, this looked kind of lame the first time I saw it. Now I really wish I could give it a try.

neat

Comfyorama Origins

Hey, I'm the guy who made those maps! Once I can replace my tablet (hopefully before the 10th), I'll be happy to make some actual textures for you, assuming I'm able.

Your OS most likely has a setting for how much time is required to doubleclick, you could lower that setting. In windows it's in Control Panel > Hardware and Sound > Mouse > Activities. For Loonix, I have no idea, probably depends on your DE/WM/Distro anyway.

Speebot. Your game looks like a lot of fun to speedrun.

What type of game is it going to be? What sort of story is there (if any)?

And most importantly, do you need any help?

Why do I always have to suffer from crippling anxiety whenever I have to do something that looks relatively complex but is actually simple.

I know that feel. I once procrastinated for an entire day because I was so intimidated by a thing I needed to do, but I ended up completing in literally less than 10 minutes.

Sometimes it just seems more complicated than it really is when you actually start writing it out.

You could say that secret to making game is "like just making game".

Thanks, though I guess I'm retarded.
I keep getting "fatal error: boost/filesystem.hpp: No such file or directory"
Can you spot what's wrong with my makefile?

do you have another boost folder inside boost ? Cuz you set the include path as boost_1_61_0\boost and the file included is boost/filesystem.hpp so it's gonna be boost_1_61_0\boost\boost/filesystem.hpp

Oops
Although now I'm getting
"c:/mingw/bin/../lib/gcc/mingw32/4.8.1/../../../../mingw32/bin/ld.exe: cannot find -lboost_filesystem"

Actually I think I know how to take care of that. Adding it to the windows PATH right?

same with libs, fix your make, delete the \boost

I did that already

I never actually used boost so idk, but I'll download it now and try to get it to work.

Ok i've downloaded the package and I don't see any libs in there, only source. I think you need to compile them yourself. Here's link to already compiled package sourceforge.net/projects/boost/files/boost-binaries/

I am making good progress on everything, so far.

oh nice job, that looks pretty cool

what kind of libs are you using, hows it all work?

Unfortunately I am also not an artist, so I couldnt help you.

All of my art has been ripped of royalty-free websites and zdoom forums without asking anyone if I could use their sprite sheets

I just realized, after seeing that screenshot, that I really need to put mip-maps into my game

Lol disregard my last post. But i'm getting there, hang on.

It's pure C++ with SDL2, but I don't use standard SDL blit functions. In fact, this whole program doesn't use graphic card almost at all. I have my screen in memory, I do all the blits myself (256 colors only, so each byte is one pixel) and when the whole frame is ready I convert this memory blob into an SDL texture and display it.

IDK now, but when I just started coding the game, only rendering those background tiles would set me on about 60 fps (at 512x288 res), with my method it's over 300fps.

Software rendering sounds like a lot of fun. I want to try it once I get to take linear algebra, but right now my math skills are very poor, so I couldn't do anything 3d, or anything with rotating bitmaps like your player.

Is the world stored in memory using a grid based system, or are you using something else?

It doesn't look like you are using windows, but with winAPI, you can just render a bitmap in your window without going through any libraries or the graphics card.

I'm on Linux, but the code should compile just fine on Windows and Mac. The great thing about SDL2 is the portability. You just as for a window and it will figure out under the hood what to use. I'm not really good with graphic math, I'm learning as I go. Rotation is easy, especially this kind I have (no antialiasing etc.) scalling is different beast. BTW, you can add me on steam if you want to chat about whatever. I sold my old steam account and I don't have friends :(

heh yeah, it says on the site that you should only need the headers for most things.

I was doing the same earlier but caved and started using SDL's OpenGL wrapper since I don't feel like optimizing a software renderer for my first game.
However I feel like an idiot for not thinking about limiting things to 256 colors earlier. Too late now.

And just the fact you don't have to deal with the clusterfucks of each OS API yourself.
I started out with a windows platform layer and dropped that shit hard.

Having true 256 color mode is kinda pain in the ass though. You can't just take random placeholder assets from net, unless all the colors match or you're gonna fix it (that's what I'm doing + my poor painting) but it's very fast.

Modern systems will not allow you to have 8 bit textures, but you can get 16 bit, which will be faster than 32 bit. That's what I use.

Let's say my "8-bit" screen is char screen[320*200], each byte is one pixel. Then I convert it to the 16 bit SDL texture.

So, you have a palette of 256 colors, this array is 2 bytes (cuz the pixels are 16 bit) so it's short palette[256]; and then you read one byte from "screen", this byte is an offset to palette and you store that into the texture. But there's even faster way (about 50 FPS boost in my case)

You make the palette 256*256 colors (64k), then you read 2 pixels at the same time and story 4 bytes (int) into the texture

void blitScreen16_2xScale(unsigned int* dst)
{
WORD* screen_data2 = (WORD*)theSurface->pixels;
for (int y = 0; y< 288; y++)
{
for (int x = 0; x< 256; x++)
{
*dst++ = pallete16_2xScale[*screen_data2++];
}
}
}

It's hard to explain and my english isn't best, but you sure get it !

Yeah I know. I actually never learned WinAPI. When I was still on Windows and wanted some GUI in my program (almost everything I wrote ever was for BSD anyway) I just used some GUI generator. It's funny how a simple window with few buttons can generate hundreds of lines of code.

Pretty cool, I'll keep this in mind

It's not only a fun challenge to write your own bliter functions (SDL will clip/rotate/scale your images… I have to do it on my own) but I'm also stuck with old laptop with integrated Intel GPU… so the software rendering is way fucking faster.

I have this shit as well

That was it.

I don't get this.

They are fine but putting the sprite in the middle of the square doesn't work well, I made minor adjustments and looks better

Ok about the boost, I think I've got it.
Go to your boost folder in cmd and type

bootstrap mingw
b2 toolset=gcc filesystem

I didn't test it though.

I tried that and ran into an error

post it.

...

that second gif is legit creepy

Try just bootstrap and then b2 toolset=gcc filesystem

It completed although I'm not sure what it did.

It should have build the filesystem lib, try compiling your program now.

Still getting "cannot find -lboost_filesystem"

I don't get Data-Oriented Design at all. I understand the concept, split everything into separate arrays so they can be processed without making the CPU jump around and avoid cache misses, but then they mix shit in like components which use pointers that MAKE THE CPU JUMP AROUND ANYWAYS! Apparently all the articles about a structure of arrays I.E:
std::vector x;std::vector y;std::vector z;instead ofstd::vector xyz;

Completely miss understand the point of DOD, but that's like 90% of the articles on it. Then they add a handlemanager (which isn't fucking anywhere else in programming that I've seen so it's obscure as fuck), which further complicates shit.

Btw is MinGW/bin in your PATH ?

...

I haven't done research into DOD yet, but from the way you're describing it I think I'm seeing one of the problems I have with OOP.
Try not to treat any of these things like a dogma that runs your whole design and just makes things shittier to work with.

You'll likely find some situations where it makes sense and others where it isn't worth it to try. Remain curious about the ways you can handle something and follow your gut as well as your understanding of the computer.

Yup

I'm currently trying to just like make game with SDL2.
Whats a good IDE I can use for C/C++ development on Linux? Vim isn't quite cutting it for me.

Code Blocks is a C++ IDE that's open source and cross platform. It's what I'm using and I've yet to see a single flaw with it. Give it a go

You can leave IDE alltogether, just use notepadqq (clone of notepad++) and g++

It seems like components invalidate the purpose is really what made me confused. I was already using components on an earlier project with something called aggregated design, but it seemed like an OOD concept. I guess I'll just stick with the way I was doing it and assume it's DOD, but the use of pointers seems to defeat the purpose.

Hey /agdg/eeks, is there an ideal 3D engine to prototype a survival horror in the style of Silent Hill? I have few experience with 2D and it doesn't seem too much difficult considering the simplicity of the mechanics. Good lighting systems are encouraged.

added mipmaps to the world textures, it really fixes all the issues that I was having. I'm still using nearest neighbor filtering, instead of bilinear.

I don't know what are mipmaps but it sure looks better

Its like this magical setting in openGL that makes your textures look nicer

en.wikipedia.org/wiki/Mipmap

Fucking boost man. I'll save it for later and just deal with loading shit manually
I want to make enough progress to talk about my game

Urho3D
LumixEngine
Crown
Panda3D with RenderFramework if you want ridiculous lighting and high system requirements

Never heard of those, I'm gonna check them out.

Tempted to use unity with all that premade shit.

heres something from the last thread for anyone interested in joining my project

this the most recent one i have saved it shows off mining and a lot of the basic functionality

i've made a good bit of progress since then but most of it isn't really visible

shit didn't mean to reply

I've been watching for progress for a while now and there's just one thing that bothers me about what you've shown so far: will there be free-look camera controls? I don't know about other people but whenever I platform, whether it's one of the 3d marios or even dark souls, when I go to pull off a tricky jump, or any jump where it matters where I land, I rotate the camera so it's behind the player character, and I'm jumping directly into the screen, holding the left joystick straight up, if you know what I mean. I don't know if I could play your game if it meant jumping in arbitrary directions with a fixed 3rd person camera.

*your progress

If you're looking to pure prototype, Unreal + blueprints will save you a lot of work right away.

A video of me fighting the new enemy that I added into my game- still needs some minor work.

Why your game still looks and plays like shit? I understand its your engine and shit, but isn't you taking too fucking long to implement simplest fucking things? Your enemy doesn't even fucking shoot back, moves and teleports like a retard, everything moves incredibly slow just like in your earliest beta you uploaded few months ago, and you are lazy son of a bitch.

You should stop working on this garbage, or get yourself together. Because its not worth it to make your own shitty version of doom from a blank page. Fucking hell this is autistic.

Nice music. What's up with the enemy randomly teleporting around?

It looks and plays like shit. Shame all that work was wasted.

I already tried it, my toaster barely runs it ;_;

the guy shoots back, he was just set up earlier to not be very accurate- When I was first playing with it, he was very accurate, so I wasn't sure how good his aim should be. You can see that he damages me once.

I can post my logs of everything that I changed since the past two weeks, if you want to see what I have spent my time doing:

7/16/16:

Version 0.6

Added support for audio playback using openAL
Added support for .opus audio files
"data/sound/test.opus" is now played when the application is started
Blood effect dissapears after 0.3 seconds of contact
Blood effect no longer appears when shooting dead bodies
Blood effect no longer appears when not firing the gun
Added new entity ENT_ARMORED_INFANTRY

7/19/16

Version 0.6.1

Added sound effect when firing the gun
Added sound effect when ENT_WORKER switches to S_PAIN
Added sound effect when ENT_ARMORED_INFANTRY switches to S_PAIN
Created rountines for a menu
Added "Sigma Engine" splash screen at program start
Added level select screen in the menu
test.opus is now the menu music
User can now choose either the test map tower1.s3bsp or warehouse.s3bsp
Mouse cursor is no longer hidden or frozen during the menus or when the console is active
Fixed bugs with enemy gravity
Fixed bugs with enemy path-finding
Added jumping
Extended the "tower1" level

7/20/16

Version 0.6.2

Added "facility" level
Added entity ENT_KILLER_ROBOT
Added entity ENT_ELITE_GAURD

7/24/16

Version 0.6.3

Added ambient music to play when the maps are loaded
Added firing animations to enemies
Added firing states to enemies
Fixed skybox clipping over enemies
Fixed aigraph on facility map
Enemies can now damage the player

7/25/16

Version 0.6.4

Added the new map "outside" to the availible maps.
Added new class of "projectile" entities.
Added entity ENT_AR_GRENADE
Added entity ENT_ENERGY_BALL
ENT_KILLER_ROBOT now shoots projectiles at the player.
Fixed framerate bug, game now will run at the correct framerate specified by vid_fps
Added shotgun weapon - currently incomplete implementation

depending on how simple it is you might get away using blender game engine

Wait, Blender is a game engine now?
Alright, added to the list.

The music is: freemusicarchive.org/music/Taiki_Ozawa/nx2016-07/nx2016-07A_02

In my game, the quality is lowered to mono at 32kbps.

The gimmick of the enemy that I showed, is that it will shoot a projectile and then teleport to the projectile after either 5 seconds, or a collision with a wall. It is meant to throw the player off guard.

What should be changed about it to make it better?

You can download the newest version of my game here, if you want to play it yourself:
mega.nz/#!EIZB1J4D!805qPsH30CYQQnLTU5Ub2faeC4uKC3OT7jaUtTHc5BQ

Blender has had a game engine for a while now. But it's in Python so it mostly goes under the radar

Wasn't that supposed to attract people since it's easy?

Why does nobody here use UE4?

It drives them away because it's single-threaded, interpreted ass. In the real world, no one cares if you have to put a semicolon at the end of a line or not, performance and libraries is what matters. There's no shortage of high-performance, free 3d engines in this day and age.

well it has a node editor, and can be used for simple projects with tank controls and alright graphics

PROPRIETARY SOFTWARE

I probably would be, if I had a computer that could run it.

Same

Most (all?) devs here aren't making anything that would require the capabilities of an engine like UE4, and trying to use it just adds unnecessary complexity and bloat. For example, compare number of posts with pixelshit and simple meshes vs the number of people using physically based rendering.

While your labor and software tools might be free, you really can't expect to develop a game with a 0 dollar budget.

I'm on Game Maker at least partially because of my limitations. While I do want to make the game I'm making now, I'd be doing much more crazy CUH-RAY-ZEE projects if I had the tools necessary.
I know. I'm saving up, but I've got so much other shit (medical bills, car repairs) that need paying for, it's put a delay on things.

I know that feeling. I'd be doing a lot more ambitious gamedev if it wasn't for the time and skill involved with producing 3D assets of even passing quality. I really wish we could get ahead of the curve in this regard, like if open-source, easily modified assets were a thing, or procedural generation tools for more assets besides shit like trees (speedtree) were available.

I'm not really trying to, I just don't have a 3D card at the moment and I'm waiting for the current gen to shake itself out a bit more before I decide what to get.

Butthurt millennial fag detected.

They're being needlessly harsh about it, but they're not wrong that it doesn't look like it controls well. It feels extremely stiff.

fugg

Space Pirate in Space has a pretty firm foundation, I think, now.

God damn there's a seriously bad case of "everything is a fucking placeholder" and god damn everything is rough and unpolished and needs a ton of work and god damn is there zero fucking content in anything, but at the very least there's only one way to go from here.

Didn't the dev basically stopped working on it?

Eh? I'm still working on it.

Are you thinking of another space pirate?

Maybe

The gimmick doesn't seem very threatening, especially not after the first time. The player would have already figured out the gimmick and would jest keep an eye on the floating orb. Now if you made it move faster and set it to move behind the player, that would be more threatening. Or maybe keep the motif going with a couple of additional skills and use him for a boss of some sort. Also, it's difficult to tell that this guy is trying to fire at the player.

What this guy said.

The game is looking pretty good, placeholder aside, but those characters in the opening look atrocious.

Just watched this gameplay and oh wow am I laughing. Opening animation is top. Serious advice though: make the destructible box items actual cubes and not sprites. As sprites they float around and pop between angles a lot worse than a round or organic object might. A cube with a minecraft-tier low res texture on it would fit into the world better imo. Also the 6 gorillion gems bursting out of them seems a bit much, placeholders or not. It takes you forever to pick them all up. I didn't watch much further than that because like this user:

While I can get nostalgic over low-poly, low-res 3D games from around the turn of the century (like pic related), I really don't understand the appeal of the Doom-like FPS as a game genre; it seems genuinely autistic to me. That and arena shooters are games wholly constrained and defined by the technology of the day, and recreating them today is like someone going for the ZX spectrum 2-color-per-sprite, neon everywhere look over say, the NES or SNES look for pixelshit. To me personally, at least. Feel free to sell me on it though.

Shit I wasn't the only person who played that game?

I laughed my ass off animating it. It's the most low-effort way to simulate talking, I love it.

I tried giving the crate a whole ton of rotations at 32 angles, but if it ain't working then it ain't working.
I'll look into voxels and figuring out how they work, then. 3D modelling a literal square in Blender would probably be easier, but I think voxels would blend better with the pixels of the environ.

This is fair. Especially since it directly contradicts the rest of gameplay, with all the fast movement and jets and etc.
Think a Metroid Prime-style beam charge up attract item thing might be good?

Drakan was the shit. My fondest memories include backtracking all the way to the beginning and hitting the little pink and yellow goblin dudes with the biggest sword and having them go to gibs in one blow. Can't believe it came out in 1999.

Plus that fucking "blergh…" noise they make when they die. I gotta play it again.

You don't need "voxels" for that, just a textured cube that turns to wooden sprite gibs when struck. I can't make an honest critique about how many gems should come out of crates without knowing about your game's economy and what all the ￥ is for, just thought it broke the pacing.
Maybe, if you're not worried about ripping off Metroid. A lot of games solve this just by having items within a certain radius attract to you like a black hole.

If you're gonna have that many gems, have them drift towards you like you were drawing them in with a magnet. It's anti-fun for them to be annoying to pick up.

I do, and I keep seeing people posting UE4 stuff or asking UE4 related questions.

I do. Almost done implementing an HTN planner into the engine to use for my games AI. Got most of the gunplay mechanics in game and working with multiplayer. Once I've got the AI I'm gonna move on to expanding my magic system, adding in hacking and a hacking minigame and then add leveling up and shops.

Also might as well post some progress I figured. Bullet penetration, ricochets, and shattering (and simply embedding into the target, which is just the bullet doing damage and being deleted) are mostly done, I can fine-tune the values for materials and such but the functionality is there. Now I just need to network it (due to random values being used I'll likely destroy the cosmetic bullet actor on clients after the first impact and then it replicates from the server). Ideally I'll add some more shit (like bullets shattering in the material and shooting fragments out on the other side) but that will be much later.

Were you the guy doing that UE4 shooter with destruction and sniper rifles, something like that? Can't really describe what he did.

Speak for yourself, buddy.

Good stuff. I've got a ballistics system going too, very similar except mine's server side only so it works with multiplayer just fine. Is the way your penetration changes the trajectory of the bullet done using actual physics formulae or just randomized?

Ah, whatever. Can anyone explain to me how to get a properly magnifying scope working in UE4? I do know that I need a camera with a lower FOV at the front of the sight, then I need to display what it sees on a texture, and adding the reticle as well. How do I go about on doing that?

Camera with a lower FOV in front of the scope. Then draw a texture to the screen for your scope. Reticle can be a part of this, or separate. Can use a post process material to 'blur' the edges and make it look magnified, you'd have to ask someone else about that since I'm not an artist.

Ain't that guy, but if I recall it was similar to this.

Well, by the end of this, this one will likely be as well. The only thing that will happen is the client spawns one as well which will handle the first impact effect. My hopes is that it will diminish the delay between firing and hit effect seen when shooting something up close (might end up doing the bullet purely serverside replicated though, will see).

Currently it's randomized in angle, where the maximum angle is based on the penetration (the smaller the ratio of penetration/materialthickness the larger the max possible deviation). The penetration calculations themselves are based on Newton's impact depth formula with a few additions, to take into account bullet vs material hardness mostly (since most bullet materials would rarely hold together inside the target and instead break up) and speeds which can vary a lot.

I could do a whole lot more with it, I've got a professor or two who do research in this field who could hook me up with whatever research papers if I really wanted to go overboard, I just don't know if I should really go that far, since this isn't really a simulator and this system comes up with at least semi-realistic results given the right values.

Robot Vault
Robohop
Upsurge
Joltspring

I meant the specifics, how can I draw a camera's vision to a texture in Blueprints?

Sounds good. Very similar to mine. Mine's based on the same formula from Newton but simplified as I'm not making a sim. Ricochets are clamped to angles based on material, but for now it makes it impossible to ricochet and hit yourself by firing at a flat surface in front of you. Doesn't prevent multiple ricochets from hitting you, however. I just wish I had an artist to make me models and anims. Don't exactly need one right this second as I'm still implementing a lot of core mechanics to the game, but it'd be nice.

this might help

Similar to mine too then. The chance of ricochet depends on how hard the material is vs how hard the bullet it, so wood is less likely to ricochet, while stuff like steel will ricochet a lot if you don't have enough penetration (which you likely won't have without a .50cal, steel's a tough motherfucker). I do have the possibility of a bullet ricocheting right back in your face but it's rare at worst and nigh impossible if you aren't shooting a steel plate head on (and the bullets lose so much velocity from it they are unlikely to be lethal).

Assets are the bane of every programmer.

Pretty much. Your system for ricochet sounds pretty cool. Mine will do it on any surface right now, but only if it has the required energy and if it isn't going to penetrate, naturally. Loses a bit of energy but not all of it. Had a projectile ricochet up to 11 times before. Wouldn't do much damage at that point however. For now, it's just a random chance to ricochet, about 70% if it's able to ricochet. Considering clamping it down to up to 3 times, but I don't really care all too much to do so right now as I'm balls deep in AI shit.

Speak for yourself. The only thing that's fucking me over is time.

(More Data-Oriented Design shit)
Now I just need to figure out how to do polymorphism without basically invalidating the entire point of the system.

pls no

Make small, single purpose re-usable components instead.

The Dev who is doing a Sonic Physic Platformer.

Anyone got a good formular for bouncing off diagonal obstables?

I'm trying to figure out how you're suppose to separate character physics from prop physics.

These are two systems that need to interact, however the character physics model is subject to change where as the prop physics isn't. Also the character physics model is upright 100% of the time.

The only solution I see at the moment is making them a special case with their own array that they go into.

Isn't that just swapping x and y?

Thought of a simple solution. Constraint system.

Are you writing your own physics engine? Otherwise any decent one should already have constraints.

Yeah. I understand the mathematics behind it I'm just trying to understand how it'll all fit together with DOD.

It's been one for ages, it's just so shit even the creators themselves don't count it as usable.

I use it, but I signed an NDA.

For any angle, you take the normal of the obstacle, then reflect your motion onto that using this formula:
d = current motion vectorn = normal of obstacle, normalized of course⋅ is a dot productd−2(d⋅n)n
If the obstacles are always 45 degree angles is right though.

Never mind I was right the first time. All unique cases (Like NPC physics) need to be stored in a seperate array

The renderer makes a clear distinction between non-animated meshes and animated meshes. And so do the underlying systems, hence the mapping is always a 1:1 mapping in that case:-) The renderer renders all non-animated meshes, each mesh having its own transform.-) The animation system only touches N animated meshes, updating their transforms.-) The renderer takes those N animated meshes and their transforms, and renders them.

I've used it for some previous prototyping. Tried blueprint and after playing around in a variety of engines, I'll probably use it for my next project once I'm done with this C++ book.

I did what I wanted.

I am doing in 3D with x and z. Yes is swapping, almost. Two angles need to be positive multiply and the other two negate multiply. Thanks anyway.

Thanks for the formular. I could use it for other games.

Ogre is getting shit on all over the place.

Usually it's ogres taking fat shits.

Note that I actually want the speed, so I can proc-gen'd the data for terrain/structures far away (obviously lower res, but larger size… so same amount of data); so I can have a relatively large view distance.
Though, the old system has its uses so it's not all for naught.

Also, I have an odd collision manager system as the way I'm proc gen'ing all terrain data/meshes on the GPU adds some unique constraints, and I need to do collisions on the CPU (without passing said data from GPU -> CPU… that's too slow).
I had to fuggin' re-invent the wheel to make it fast, but I think it's pretty cool as it scales super well plus makes the collider generation extremely low overhead compared to more generalized systems; while allowing for easy pathfinding.
Now all I need to do is redesign my main algorithm… then tie everything together. The how to do this + algorithms are planned, but not in yet.

[/blog]

It can't do what I need it do… unless I do some crazy time burning workarounds that end up being less efficient than if I use an engine that natively supports + updates + bugfixes what I require.

Engine is cool and all, but while it has its pros it also has cons; same thing with all engines pretty much. In this case, the cons outweigh the pros.

Yo man that's pretty cool. Would like to see it in action. Nice dubs btw

Takbir

Its not supposed to be a particularly strong enemy , because of the hitscan, but it should be harder to keep track of when in a room with a few other enemies. Its hard to tell if the enemy is firing or not because the accuracy was set to be lower than it should have been.

the gems seem annoying to pick up, most enemies are hard to see, the boss fight last a minute more than needed and your framerate goes to shit with the whiskey
other than that your game looks like fun

And another boring task of UV-mapping…

why do all the textures look better than the collective model? usually I have the opposite problem

How would you make the texture animated?

I see those green bilps, I assume they're different frames, would you just run a shader on the model to adjust the UV mapping?

I am working on muh microtonk mod for skullfuck actually.

The only thing that is animated are the track parts, which I have to separate it and making 4 frames then referencing them on the modeldef file 4x times, hmm actually I could cut the reference to 1 since I can load the model/texture 4x times per reference.

zdoom.org/wiki/MODELDEF#Examples

Don't ask me why the tracks are invisible ingame, it worked before I decided to atlas texture it.

GZDoom/Skullfuck OpenGL render works in a mysterious way.

Have also some ancient webums I made during the fucking Windoze 7 era, I am on Loonix+Stallman now

Unrelated note: It still pisses me off that there is no way to prevent sprite clashing that and no sprite sheet support because I would have added software support by now, since ZDoom Software render can till to this day not render any models. Which means also I need to cut down some of the sprites I have and making it a bit more standardized.

Yet funnily enough there is voxel support.

Thanks everybody for the name suggestions!

I don't mind it being obscure as long as it's catchy and pronouncable.

I like Speebot, it's #1 on my list of potential names. It's cute, short, memorable and fitting.

It's a platformer, still thinking about story. I don't need any help, sorry.

Fixed camera that follows the player, the angle you see in the pics/videos is what the player always sees. The jumping frustration is a level design fault which I'll try to avoid.

Will consider these as well!

Really? Sounds plain silly to me.

Not that I can think of anything better right now.
Looking at him I get the idea of some kind of highway patrol robot. So in all my creativity I thought of Highjump Patrol.

speebot is really good, if you can't find anything else go with that
i was thinking of some kinda english written nip in it, something akin to "doki doki panic" for example.

Your water shader looks pretty dope.
Will there be a boat mode?

Boat mode seems like it'd be pretty comfy and not too hard to implement
in that game.

dammit

Reading my mind! I think it will be fun so I'll have to find a way to weave it into some levels, or maybe into some bonus levels.

Time to start looking at either how to make things drift to the player or make some kind of vacuum tool to suck them up, then. Thanks for the input.

No clue what happened with the framerate on the booze, though. FPS was perfectly fine while recording and in the recorded video itself, either YouTube cocked it somehow or Sony Vegas did while rendering.
Either way, I'll have to fuck with the settings.

I don't know shit about ZDooM, but maybe this tag would work for non-missiles too?
zdoom.org/wiki/A_SeekerMissile

What sort of mechanics, themes, or storie would you want in an old Nordic, somewhat mythical game? I've been working on the basics of my game using C++ and SFML. Curious what others might want to see. Image related, although doesn't need to be Varangians.

Did you check Banner's Saga? It has a cool premise.

I'd like to playthrough Ragnarok (especially the Battle that ends time), but that might be a tad too mythical considering your use of the word "somewhat."

adding some new enemies into the game, right now their attacks are not fully implemented.

Your art is pretty good; Have you done other stuff before?

Why not an adventure game about varangians under the employ of the bizantines instead of another typical viking game?

Whoever made this fucking Image needs to be executed for being a dumb bitch.

As the others said, but in one post:
Make the game fast. The jetpack is especially good for a fast paced game. To achieve this, less is sometimes more -→ less coins (maybe they could be magnetic, as said. Other than that, the basics looks pretty damn amazing.

[16:32] anonymous_coward: it's a bug in luajit that was introduced at some point and i can do nothing about it
[16:33] since noone else went as far as I did with its ffi/bytecode exploitation, my bug is very hard to reproduce anywhere else
[16:33] and whenever i change something even a little bit, it stops happening :|
[16:34] a workaround is to build luajit 2.0.3 and point the engine at it… but i'm not going to deal with this anymore, since the work in progress next generation of the engine doesn't have a luajit dependency
[16:35] it's pretty much a reboot, and it will provide direct native API
[16:36] then if games want so, they'd be able to integrate their own scripting, but it's not necessary
[16:37] there will be a lot of new features, portability to more platforms, and proper multicore support
[16:37] and loadable plugin system with on-the-fly reloading
[16:38] i already have most of the design planned, and implementation is ongoing
[16:39] also the project will become a lot more split up, with several components becoming standalone and useful outside of the engine
[16:40] (the C++ utility lib is standalone now, its custom build system is a standalone project, its cubescript interpreter has become standalone and librarified and mostly made threadsafe etc)

I'll stick with Tesseract for prototyping until OctaForge's reboot is further in development.

how do I stick to it

basically my cycle is

I don't really have a validation that my work is great. maybe mediocre at best

in all honesty, I just want to make a game for myself mostly.

Work on it every day whether you feel like it or not.

my longest streak was 3 months
but I couldn't escape the 'sad thoughts' or just can't shake the fact that it's shit no matter what I do with it

I just want to get out of all day grinding hell and make something without worrying and shit

Thanks user, and I'll upload some webms once I finish the new implementation

Why dont you just buy all your animations, art and whatever else of the EU4 Marketplace and make your own game Holla Forums ?

How much of the art did Ching Chong actually make himself?

I really dont know how the marketplace for unreal engine works, I am assuming he bought everything and just copy pasted it together?

I just checked his >Twitter and apparently he at least did the main character's design and art himself.

Those are some strange symbols on that chart. Wouldn't the end-game be boring? Or is the goal eternal peace?

I was thinking of making it more like a villiage or settling game where you play the head of a Norse family trying to make their way in a new northern land while beset by all those old Norse myths and more mundane things.

I've played Banner Saga. The battle mechanics fall short often, but the rest of it feels solid. I don't want to make a huge RPG though. I want it to be more 2D adventuring with some villiage settling in the process. Maybe expand to more villages, but I need to start simple.

Data-Oriented Design is such a mindfuck. I think it's just because I don't understand how cpu cache works fully. However, I see them doing shit like using an array of indicies, which seems like it'd fuck everything up anyways because it's bouncing back and forth between arrays

Wait, you mean he didn't do the combat programming? That system just exists out in the wild somewhere?

Can I get lawyers on my ass because of mechanics? My tactics game is almost identical to Fire Emblem.

No, game mechanics are not copyrightable.

No. There are countless carbon copies of other games out on the market.The only thing that matters is that you aren't using their assets or their exact code. As long as you're doing that they can't really touch you.

Thanks for the answer, anons.

Then stop worrying and just make it. The only thing stopping you from making something good that you want to make is yourself. Once you get past that barrier you'll be able to make it.

How did you get those movement indication lines to work?
I've always wondered how to do that the best way and those look pretty smooth.

He draws a radius of tiles (which is wrong, the right side should be "shorter" than the left side, because of the wall blocking his movement), generates the fastest path from A to B to create a list of coordinates, and then draws a line from midpoint of each tile in the coordinate list.

Not him, but they appear to just be drawn from the center of the current tile to the center of the next tile in the path, and then to the next, and so on til the last tile.

It looks wrong to me too but for another reason: it doesn't let you travel as many tiles diagonally as in straight lines. The movement space should be more of a circle than a diamond, unless moving diagonally costs twice the movement points.

But they are patentable.

If you're still around, what are you using for the gpgpu side of things, language/library-wise? I've been wanting to get into better exploiting the GPU for algorithms myself.

But it is.

Oh I guess you're right. Character is too close to the edge of the wall to make it noticeable at a glance.

I use compute shaders for the GPGPU part, i.e. constructing vertices/normals, and other parts of the algorithm I'm using.
Language is HLSL, and it's all DirectX based, but it also "sorta" has OpenGL support.
The engine I use, Unity, has native support for it; so the lower-level stuff isn't really a topic I've delved into as it's a non-issue.

Another alternative is OpenCL, which is more of a computing centric API (has access to lower-level details, easier to implement certain algorithms that require local memory, etc); as compared to compute shaders which is a mix of a computing and graphics API.
Although, if u can do what you want2 do with compute shaders it'll more than likely be faster (I've compared them in rudimentary tests, but only for a C# port of OpenCL vs native compute shaders… so it's a very small sample size).

I'd recommend to only do GPGPU stuff if you can generate/keep the data on the GPU.
Although, if you have a requirement of transferring from GPU -> CPU then you either need to not do it too frequently, or need massive gains in the execution time of your algorithm(s) to make it worth it to do it frequently… as that's transferring over the bus of the mobo so it's quite slow.
As, it's less of concern of "how much data", but a constant delay due to going all the way from GPU to CPU; so transferring one value compared to 512 values has little difference in overall transfer time.
Although, if u do some massively costly algorithm that you execute every few seconds then it's more than fine as transfer times for small to large sets of data range from 2-4ms in my experience.
Though, massively parallelizing algorithms is quite fun, so I'd highly recommend it, and it makes those tantalizing options that seemed impossible before within reach.

Also, if you want some good resources for compute shaders I've got a bunch bookmarked; lmk.

I keep seeing the timestamps on these and crying inside
I haven't felt the urge to dev in almost half a year

I remember your screenshots. Whats your game about?

It doesn't, but it gave me a wonderful idea that works beautifully.
Thanks very much, mate.

How about something like this?

Fuck, I think at the time I wanted to make a Viking-themed roguelike called Vinlandsaga.

It was easiest to work on the lighting system first I reasoned, and I hit a snag with fogs and other materials that introduced a "cost" into the light source floodfill, since neither XNA nor C# provided a simple sorted heap structure, so I had to implement that and a bunch of other shit, then I lost sight of my goals, and also couldn't reproduce the fog of war I had happening.

My issue with the FOW was that I had a few different textures, and some combination of blend settings allowed me to bias the scene between ambient/global light, and specific light sources, eg to emulate day/night or interior/exteriors.

I noticed that all the screenshots were just of the lighting system

The lights look good enough to me, maybe you should focus on making the game part and then polish it later?

How make time inbetween procrastinating, watching animu, playing vidya and working?

Had a good laugh at that. Good job. I'd turn off the magnetism after a specific distance however.

Well the thing is I know I have a history of abandoning projects, so I've found it more helpful to focus on a modular concept with a "game theme" in mind and build towards that; when I inevitably get bored, I'll at least have a piece of an engine I can easily salvage.

Although the problem with that lighting system was that it was primarily tile-based, if I had to redo it, I'd probably make it more open. I did have constant/linear/quadratic falloff as well as user-customized flickering enabled, too

quit 2 of those things.

Work out at least a rough schedule that you hold yourself to, and give yourself ample breaks when you reach a moment you feel overwhelmed or are pushing yourself too far.
After a time, it'll essentially form into a perpetual machine of motivation due to having dedication; which the schedule helps to build.

Out of curiosity, for those making shooters, how do you guys handle recoil?

I've noticed a lot of newer games, particularly console games do it in a way where it just adds screen shake rather than moving the characters aim, and then over time it returns to the same position. That way you don't need to ever move your aim if you wanna hit a target, you just have to wait between shots for the aim to return to original position.

I think that system is gay as fuck. What I have is the same as STALKER, possibly other games but none I can think of off the top of my head. Moves the players camera vertically by x amount each shot. Also can move it horizontally but I don't like the feel of it too much.

Each game should be looked at individually. A lot of AAA games focus on high speed toybox rampage style gameplay these days, so springy aim and regenerating health are proper design choices for them. It really depends how you want the game to play.

My game will initially play as more of a tacticool. When you get further in, depending on choices and how good you are, it'll become extremely fast. Sort of how Deus Ex progresses, but even further.

I think the way I have it with recoil is fine. I can't think of any other recoil methods that are suitable.

Man, I know fuck all about shaders.

I'm actually working on something nonvidya related at the moment, a predictive modeling project in C, that has benefited greatly from parallelization, and I've been eyeing my GPU for more power. I've been looking into OpenCL but haven't actually started using it yet. It involves an awful lot of building very large matrices, multiplying them together, then breaking them down again.
Eventually, though, I plan on getting back to work on vidya and hope to use some of what I've learned to build some graphically undemanding sim projects and use the GPU for simulation.

Yes. Yes it is.

If you've got anything that you feel is appropriate for someone who is generally comfortable working with C and C-likes but doesn't know shit about shaders or graphics programming in general, I'll take it.

Found this while lurking epic forums, it's a cat simulator made by two people (and two cats)

hk-devblog.com/

Maybe I can turn down the realism a bit, and make that Gnome game this user ideaguy keeps reposting for some time.

Right now though I want to finish my current 6-month project

...

I haven't touched the UE4 market in months, most of the stuff there are high end only.

I had to buckle down and learn Blender. Got good at unwrapping pretty quickly.

For models with animations and physics I still need to hire outside help, but for almost all the props I can just paste the textures from the concept art.

I can model basic stuff in blender but I've never done any unwrapping or texturing. I want it to be pretty high quality in the end too, so that makes it tough. I'm really not wired as an artist anyway, my creativity exists in other ways, not for aesthetics.

unwrapping / texturing is less of a creative endeavor, and more of a tedium though. That's why I can do it, when most of the beautiful designs have been designed already.

If you're willing to shell out a little, make a list of the props you want, then hire an artist to draw the concepts… especially the textures. Then you can model the shapes and plaster those textures. You'll benefit both from an inspired custom design as well as the satisfaction of having put the end result together yourself.

That sounds bad ass as fug
I have something similar in that realm planned, implement procedural modeling that's an amalgamation of a few different approaches in architecture/biology for procedural modeling.

OpenCL would be perfect then, and in passing I've seen a lot of papers on parallelizing matrix maths; so there's resources specific to what you're doing out there.

You'll fit right in. HLSL is based on C.
Also, the OpenCL language is essentially C(99).
As are just about all mainstream shading languages (CG is the only other one I use, same deal here).

Note, that the idea behind CUDA, OpenCL, and compute shaders for memory, threads, and such are all essentially the same.
The only differences are their syntax, capabilities, and hence what they expose in their APIs.
So - for at least the concepts + way parallelization works - any resources centered on those APIs works.

Intro:
channel9.msdn.com/tags/DirectCompute-Lecture-Series/

Entry to intermediate topics, and source code to learn from:
blogs.msdn.microsoft.com/chuckw/2010/07/14/directcompute/

Advanced learning resources:
developer.nvidia.com/educators/existing-courses

More source code:
developer.nvidia.com/directcompute

OpenCL references:
khronos.org/registry/cl/sdk/1.1/docs/man/xhtml/

There's plenty of OpenCL resources + tuts out there because it's in heavy use by multiple industries, but I didn't save any because I decided to go with compute shaders.

Would you mind elaborating a bit on how you get a texture from some concept art? You just recreate the texture yourself? Or do you get the concept artist to actually deliver usable textures himself? Also, how do you find your concept artist?

Hmm. don't we have code tags?
x = 1if (x): print("Hello world")else: print("No")
testing

In the past I used to ask the artist to give a texture set as well. But recently I just ask the artist to give me orthographic views, so I can just cut out the texture from the concept itself. When unwrapping, I try to make the seams so that the cut up parts resemble the general outline of the object I'm texturing. I then patch up areas using the clone tool.

I find the artists from looking at Patreons and Kickstarters and looking up the portfolios of the artists they hired. If you want to save, get someone from Vietnam, Thailand or Indonesia.

Would be possible but I don't think I'm good enough at modelling all the high poly shit I want. It just adds another huge sum of time onto development that I'm not sure I have.

Still cool though

Thanks, I'll take a look at all those.

Procedural modeling sounds cool too though. I had to look it up, I really am totally ignorant when it comes to graphical programming.
I really am a predictive modeling/machine learning/AI/statistical physics junkie.

Have you looked into CSG-based level editing/modeling? Shit like gtkradiant and similar. I used to go nuts with valve's hammer editor when I was a kid, it's minecraft-tier in terms of ease of use. I've been meaning to get back into it as a way to rough out levels and objects, exporting the resulting mesh to blender then to my engine du jour just to give me something to walk around on and experiment with run speeds, jump height, swimming in water volumes, that sort of shit. You may have some luck with that: making a bunch of small textures, blocking out your model in CSG, exporting as mesh, importing into blender or something for an ambient occlusion bake or whatever, then importing into your engine. You can scale, rotate, etc your textures on the brushes, no UV unwrapping involved.

I really am lacking sleep

UE4 Starter Architecture does some of that though, for basic building blocks.

I've considered it but haven't done it. I've just done simple blockouts with brush tools so far for bits of testing. Not really at the stage where I want to make levels yet as I can't really play them and I need to design scenarios around the levels too.

Pleasure m8, would be nice to have someone around here doing GPGPU stuff too hah

I feel ya user. I love AI too, and tbh while I enjoy doing graphical stuff I'm excited to get to a point where I can get back into AI again, and flesh out + program what I've got planned.

I really am in the same boat.

I imagine it does, I used to do it for UT2K4 levels back when that game was fresh. I really liked the ability to bake lightmaps in-engine, but it was fast and crude. You can get much better results with a proper global illumination setup in blender now. I'm actually playing with installing gtkradiant right now but it's 5am and I don't have time to get into it. I'll try to crank out some textures from cgtextures or some such and see if I can't make anything presentable from my "edited CSG -> blender for lightmap bake -> back to engine as mesh for render" idea.

It never hurts to experiment with a test level. Not everything you make will end up in the final game. I feel that same sort of hesitation a lot, like a fear of wasting effort and time, so I spend all my time planning and not doing, when I should just be cranking out placeholders asap. Hell, I wouldn't even sweat the quality of final product too much either. People still play whatever the latest counter strike game is and it looks blocky ass.

While it's kinda blog-y, the look I'm going for is similar to the metroid prime games: relatively simple textures on relatively complex geometry + a good light bake. I remember being absolutely blown away by how good it looked, especially in morph ball mode in a tunnel where you can see how every little nook and cranny is modeled in, there's wires and conduits and shit underneath, unlike games today that do the same thing with complex meshes and normal maps. /blog I really am gotta sleep too.

sleep sounds like a good idea
also CSG is awesome, props

Said patets are pretty easy to get around, by changing some tiny things. Nintendo has a copyright on insanity meters causing visual effects, yet Amnesia has that kind of insanity too. That's because Nintendo's includes in their definition that your insanity goes up when enemies see you, while in Amnesia it's when you look at enemies.

(checked)
This kinda reeks of rendered non-RT "gameplay" but I don't care what it'll look like as long as I finally get a catformer.

Well, we'll see how long it takes me to get on that. I'm slogging through data validation and formatting routines right now, boring but necessary.
Either way I'll be here I've hung around /agdg/ forever, even though I haven't vidyadev'd much in a long time. I just like it better here than /dpt/ or /lam/ or the like because they tend to be more interested in writing a smaller fizzbuzz where aggydaggys always seem to be interested in making something new, even if it means banging their head into a wall for a while. Plus I occasionally get to be helpful.

Most are, some aren't. Namco is pretty notorious for annoyingly enforceable patents.
Also most mechanics can't really be patented because they proliferate too quickly for someone to claim originality/ownership, which is why companies that are really serious about it (again, Namco) get their patents in line before even releasing a product.

But someone already released a good catformer.

kill me

anyways, after minor tweaking, i got the rpg system working with the boxing game
now it's time to abandon everything because i don't know how to actually make a game out of all of this

Call it boxing simulator and add a few more stupid memes and some levels and watch as tens of thousands fly into your account if not hundreds of thousands.

but i don't want to make meme games
ignoring the fact that i wanted to put van darkholme and fesh pince maymays in here in the first place

anyone have any idea how i can get A* pathfinding working in 16 or 32 directions?

But you'll be rich if you do it. It's okay, I don't want to either. What little dignity I have is worth more than money from a shitty meme game

I think there's already a meme game based on boxing. I thought I remembered seeing it when I was scouring Steam Early Access for any games worth waiting for the 1.0 release.

not quite sure if i even understand A*
it has a grid, and each square has a cost for how much it would cost to reach said cube, right?
in that case you'd just have to get a different formula for assigning the cost
no clue how though

there's this one that's on the greenlight, and i'm hoping it comes out soon, since the only other one that's out right now is Real Boxing and it's garbage
there's also this VR one, but i wouldn't buy into gimmicks

More or less but with standard A* every step is in of 4 directions which just makes the movement look stupid.
I think you need at least 16 directions to make it look decent.
The real issue is making the algorithm not favor certain directions since not all steps will be the same length

can't you just divide the grid a bit?
then the shortest path your character needs to take should be more and more accurate
unless i'm understanding you wrong and you want the character to move not entirely according to the designated path

no clue if this actually helps
redblobgames.com/pathfinding/a-star/introduction.html

Technically A* doesn't have to be a grid, you can put any arbitrary mesh of nodes and use A* to calculate the best path along those to get to a certain node.
Cost is what makes A* different from Dijkstra. One basic component of cost is distance, for a grid that would be 1 for adjacent and ~1.414 for diagonal tiles. The rest depends on the game. A TBS game has different terrain types that have different costs. A boxing game could have different costs depending on the boxer, for instance one that likes to step back a lot has a lower cost for the nodes behind him. Blood and sweat on the floor could add more cost to nodes, too.

8 directions plus a small (

>8 directions plus a small (

In your opinion, what are the advantages and disadvantages of selling an one dollar game?

I was suggesting A* methods for that guy's boxing game, my bad. For an RTS you want a navmesh, not classic tiled A*.
en.wikipedia.org/wiki/Navigation_mesh

why?
there's usually no crazy geometry in rts games, tiled worked great for blizz rts games

pros : You will most probably get more sales, so more players, and more feedback (not necessarily more money).

cons : Most players won't "respect" the game if it's that cheap, and might overlook how good it is if it's really good (i'll elaborate on that if you want).

Because it's faster (if you do it right), which is good if you want a lot of units. Because it's more accurate for better landscape. Because it allows overlapping areas like bridges. Because it works better with boid systems so your units avoid clipping into each other. And because more reasons probably, but I don't feel like thinking up more of them.

yeah i would love to use a navmesh but that wikipedia page doesn't really tell me alot
and im not sure i can figure it out myself
Since its an rts the mesh needs be generated dynamically
Im already doing something kinda of like this
Where I ad nodes around every obstacle and A* traverse the graph
The problem is creating the graph is O(n^3) at worst
with a spatial quadtree that could be reduced to O(n(log(n))^2) which may be good enough so maybe thats what i should do
though from what i gather generating the spatial quadtree would be O(n^2) which is still a big improvement but maybe not enough

so please tell me more about navmeshes if you can

Do it.

Nope. Boids. Both units and buildings repel other units. Works fine as long as you can't use buildings to build complete walls. And if you can, just have units that can't progress for more than X amount of time shoot what's in their way until they can move on.
There are ways of building navmeshes dynamically, but I'm not an expert, and I have to go pick something up in a few minutes so I can't exactly write up a lot of what I do know either right now. Try thinking if you really need dynamic generation or not, and googling around some more for now.
gamedev.net/page/resources/_/technical/artificial-intelligence/generating-2d-navmeshes-r3393

If the game is good, but it takes some effort to really get into it (for example, you have to learn to play it and get good before it becomes really enjoyable, or you have to go past the first level / chapter to start noticing how good the game actually is), i suspect a lot of players won't expect so much quality and content from a one dollar game, and might just play for 5 minutes, and then stop, and miss everything that makes the game good, and never play it again, forget about it, not talk to anyone about it, there are so many other games to play…

If the game costed them 10 dollars, then they'll be more motivated to search where their money went, and they'll be more inclined to dig deeper, progress a bit, get better and look for what's good in this game. It's OK (and i'm afraid it's even expected) to get a mediocre game for one dollar, but if it costed more, then it's not OK for the game to suck : the player will want to know if the game really sucks or not.

The example that comes to my mind is Frogatto : a wonderful little pixel art platformer, great mechanics, gorgeous art, good and fun writing, quite a lot of content. It's a good game. It was launched at one dollar, and pretty much flopped, no one heard about it, and it was quickly forgotten. The devs kept working on it (they're actually still improving and expanding it), and the game now costs 10 dollars, and is doing a bit better than at launch. I suspect it would have sold much more if it had been launched at 10 dollars from the start.

If your game is good, i think it's a mistake to sell it cheap.

You should always avoid underselling yourself or the goods you make. Although people aren't consciously aware of it, it does give off a bad impression.

If you're selling a product for less than it's actually worth, then that says a few things about it/you:

None of these are impression you want to make, and most people generally only get these subconsciously. But they are a thing, and you should avoid under pricing your game because of it.
If you decide to roll with it anyway, then you'll see a lot of people take a glance at your game, see a price of one buck and decide to skip it because if it was actually worth playing it wouldn't cost just one dollar.
Others might buy it, but then don't take the game seriously since they're going in expecting a buggy mess to waste their time on. Even good movies are shit if you try to watch them as if they're a shitty B movie.

There are only a few cases in which it is worth selling your game for cheap:

I definitely want pathfinding to work with buildings and not just bounce of eachother

I want the pathfinding to work with both building and units actually
I tried all kinds of collision but nothing really worked the way i wanted it to
basically I want my pathfinding/collision to work similiar to how it works in starcraft 2 melee units naturally wrap around enemies and units don't get stuck on eachother and you can use buildings as walls.

Starcraft 2 Uses navmesh A*, plus a whole heap of other tricks and shit.
gdcvault.com/play/1014514/AI-Navigation-It-s-Not

Dumped everything I've discovered about DOD so far in our dead board: 8ch.net/agdg/res/27468.html

if you are talking about the movable grid I used Dijkstra's algorithm, user in the past thread told how to do it on general and simple terms, but I forgot it, so I'm going to explain my logic:
- check position
if it is "empty",
- make it "movable"
- move points–
- check adjacents (recursively)
if it is "blocked"
- let it be
- end function.
if move points

actually it looks like boids is what im looking for.

this is probably more complicated than i can manage

what's a good formula for doing damage based partly on the collision and partly on some variables?
in this case i have a punch that has a certain amount of damage, and then i want it to be modified by the magnitude of the collision
however in this case, the result ranges from 3 to 40, which is a bit too much when the characters have 100 hp per limb
int rag = (int)(curpunch.Damage * ( Mathf.Sqrt(target.GetComponent().velocity.magnitude)));
for reference, the current damage is:
left jab - 3
left hook - 7
right jab - 5
right hook -10

(int)(curpunch.Damage + Mathf.Ceil((Mathf.Sqrt(target.GetComponent().velocity.magnitude)) / 10));
Would give you 4-7 for left jab, 8-13 for left hook, etc.

it kind of works but the results don't vary at all

Ah wait, I guess I misinterpreted your ranges. Do you mean you got results between 3 (slow left jab) and 40 (fast right hook)? I thought the result from "Mathf.Sqrt(target.GetComponent().velocity.magnitude)" varied that much. In that case, just the following would give the numbers I mentioned:
(int)(curpunch.Damage + Mathf.Sqrt(target.GetComponent().velocity.magnitude));
So basically just replacing the multiplication from your code with addition.

yeah, i'm gonna go without the square root
now i'm getting right hooks between 10 and 15, which i think works well enough considering health on each limb is currently 100 max
when you start accounting for dodging/weaving it should be fine

1. Create the skybox. This entails either using a precanned one or creating you're own. You'll need six images of the same size as the textures for the skybox surfaces. To start with, you can simply use six different colors. Use Krita3 or PS to work on these assets, then apply them to the cube geometry of the box.
2. Inside the box volume, place you're rings, camera, and lights. Animate the rings to rotate?
3. Render you're shots. Try different combinations of camera angles/positions and different lighting combinations.

Animation isn't all that different from planning & setting up practical photography shots user. In the same fashion; models, props, lighting, camera. And (somewhat) similar to irl photography, you either have to obtain or fabricate these items before you can shoot the images.

Codelite is pretty good user if you insist on a full blown IDE. Personally I'm pretty /comfy/ using juci.
github.com/cppit/jucipp

sound effects are back in
i can't shake off the feeling that they're delayed or aren't playing as often as they should be

for what purpose

trust me, you get used to it in minutes. it can be switched off if you don't want it. it's provided by the gtksourceview library, and the juci devs make it available to the users.

not having a dark theme with the faint grid feels /blandstark/ to me now. the exact opposite of /comfy/ tbh.

Shadows are now distorted on water surface.

I lose every time I hear that fucking sumo sound

sorry i didn't give you feedback on your actual point user.

try slipping the frame by one user. that is, make video occur one (possibly 2) frame late on the audio.

the net result experientially is that the audio plays one frame early – just like it does for film. we're all used to this and it feels disconnected for most of us if that slightly early audio isn't there.

as far as playing often enough, have you tried logging both subsystems to ensure they are being consistent? ie, when ever an impact triggers an audio event log it in the impact detector. then also log it in the audio playback mechanism and compare. they should exactly match.

very nice.

It seems like you've got something new every 12 hours. What's your secret?

Are you sure it's not the way sounds are being stored or decompressed from memory? What are the base sounds encoded in?

no clue, i just downloaded them from some site for free sound effects
they're mp3s, in unity here's what it says

i should note that it could just be my imagination
i haven't bothered setting at which frames of an animation it should be counted as if it's actually attacking, the sound plays the moment the limb collides with another limb while playing an animation

I usually come up with better solutions/ideas/bugfixes when I'm not in front of a computer screen, e.g. I think about it when I'm in bed or walking the dog or at work during breaks, and do it when I come back home. I think that saves me some time.

Also, I think I can work faster because I'm using my own engine, and when I want to implement a feature I can just do it instead of wasting hours practicing shamanism, looking around docs and forums trying to work around some other game engine's limitations to get the feature that I want.

I'm also making my own engine but it takes me forever to do anything. I think trying to do things the "proper" way trips me up a lot because I spend hours researching "How should I do this thing", "How will it not fuck me over in the future implementing this new system", etc.

Hopefully now that I've discovered DOD it will be the solution to 90% of this

That could be the issue. Try .wav or .ogg.

Be careful, having a solid skeleton for your engine is important (such as a basic entity system), but don't fret too much about doing everything the "proper" way. Most common design patterns suck because they prioritize code readability over code efficiency, which is a shitty idea unless maybe you work in a 50+ man team on the same codebase.

Case in point - here's a fizzbuzz program done the "proper" way: github.com/EnterpriseQualityCoding/FizzBuzzEnterpriseEdition

If you think your code will be more efficient but more obscure, go for it - you'll still be able to navigate it quickly since you were the one who wrote it. I've had to do major refactoring of my engine once and it's not as scary or time-consuming as it seems.

Just my 2 cents.

Been looking for you. Need an artist, right? Here's some stuff I've made, I did some work for the Tiberian Sun mod Twisted Insurrection (I made both of the guard tower types in the larger image).

Also, what engine is your game running on? Also would you ever consider giving your an isometric tile grid system instead of a top-down tile grid? imo isometric looks much, much better.

Yeah I understand this, which is why I said hopefully DOD solves it because DOD basically is: "Fuck object oriented programming and all those people, code readability is a small issue performance is #1"

kek

Seriously, it does the job well, nice.

Yeah, but performance does not equal ease of writing code either. DOD doesn't strike me as something making programming easier.

This is looking pretty good user.

You don't understand it then. Now that I've got the concepts down I noticed it makes coding WAY easier in the long run. Because all the relevant data is going to be together anyways. A lot of DOD articles will point out examples of how DOD makes things simpler than OOD in most cases and the Ogre3d engine basically became a meme with DOD developers since it tied itself in knots with how hard it tried to force OOP.

Dedicated today to fixing small bugs and ended up rewriting half of combat and magic code as a result. At least it works better now.

advantages would be more 12 year olds will buy your game, more people might buy it on a whim

disadvantages would be pretty much everyone who aren't 12 year olds or buying it cause its a 1 dollar game aren't going to buy it.

The engine is sfml
I did actually switch to isometric after a short period of consideration
Though not as far as physics are concerned though pathfinding/collision are still a work in progress i will probably end up keeping the bounding boxes straight rectangles
as for the selection boxes though i think i will probably change them to projected circles to match the angle of the isometric view.
And i do need an artist actually and i like the way your stuff looks.

Not sure what you're talking about with 16 or 32 directions. Grid-based pathfinding algos are perfect for RTSs though, because most of them use grids eg. a building might take up a 4x4 space, a turret is 2x2, etc. If you don't want janky movement, allow diagonal movement and smooth the resulting path, make tanks and shit rotate when moving, etc. Anyways, pathfinding and AI in RTSs has been studied since forever and there's shittons of material out there. See stuff like aigamedev.com/open/tutorials/potential-fields/ and skatgame.net/mburo/orts/

Readability is a preference thing and ease of writing code is a meme.

well im going to bed now and this thread probably wont be here when i wake up so
email in the email field if your still interested

i mean with grid based movement u only get 4 directions and it doesn't look good
even with diagonals its only 8 which still doesn't look very good
I don't really know what you mean by smooth the path but thanks for the sources

I have made a new video of my game, this time going through a level that I made.

Hopefully it looks a bit more impressive than the last video.

Aight. Just shoot me a message and we can talk about it

It's good that you're adding more content, but you've gotta make it move faster.

You need some of this, urgent.

legit, and I agree. It's pretty refreshing imo

>you tel me ease of writing code is a meme
Barking up the wrong tree there user.

so i looked at my codebase a few days ago after constantly getting thrown errors from love and i got fed up, so i completely re-wrote all my progress into a much cleaner system, started breaking down more code into functions and rethinking how i was doing things, and am now making more progress than before

now to re-write a collision system that will either check collision pixel for pixel within a certain range or just draw two invisible boxes on everything

posted in the wrong thread and now im embarrased

How long did it take to make your own engine? I want to try it but I'm afraid it'll take too long.

Alright, I will work on this angle-

I'll speed up the gameplay, and post it later to see what you guys think.

I made this picture for you nigger.
The grey boxes are your units, red arrow is the direction they're facing.
The teal highlighted squares are its movement path.
The rocks are rocks.
The yellow line is the path if you drew it from center to center of the path squares.
The purple line, which I should have put on top and not under it, is your "smoothed" path that the unit actually follows as it moves. You don't want to deviate too far from the janky (yellow) path though or you'll appear to clip through buildings and obstacles.
The two units on the purple curve show how you keep the units facing forward in a curve, not pop between 0 degrees, 45 degrees, 90 degrees, etc.
I hope this clarifies what I mean. This is how you give the illusion of free movement in a grid-based system. If your grid is larger than the units, you can have them stop and move at random points within the squares, with their coordinates and angles and shit in floats and the grid integers, if you follow me.

Also look up bezier curves.

What page do we normally make a new thread at?

13

anyways, here's the new bread