/AGDG/ HOW THE FUCK DO I TEXTURE EDDITON

pic very much related

I made an elevator. Took longer than expected but it's functional, can be extended in Blueprints (base classes + default behavior you see in the video are in C++) and I even wrote a complete documentation so I don't forget how to make custom elevators/moving platforms later.
Oh, and I also refactored my ladder Blueprint so it supports a variable amount of safe unmount points. It's interacted with via contextual button presses. Usually I don't like them, but ladders are an exception. Everyone who tried to climb down a ladder in Half-Life while in a hurry, knows how easy it is to just run off the edge with those automatic systems. I have implemented it so you press F to mount/unmount/drop. I also have code in place that I'll use to notify the UI so it displays a warning, if pressing F would make you drop.

Next step: doors.

OBS + WebM for Retards.

Doesn't Unreal already have a door blueprint?

I have physics doors as seen in Amnesia.

Here's my small progress, I have decided to make my own engine by patching different libraries together since I couldn't find an engine that I could run on my toaster or an engine I liked. It uses Irrlicht for the graphics and soon OpenAL for the sound and Bullet for the physics. Irrlicht seems to be geared towards giant models as you can see with the sydney.md2 that is delivered with the engine compared to my 1,84m (~6ft) in blender minecrafty untextured model. That also created problems with the camera, it moved too fast, I had to lower the speed to a very low setting of 0.006 compared to what it was supposed to be (something like 0.5). I'm thinking about using Ogre3d instead.

Sorry for the lag on the vid, also it's not supposed to be this dark but whatever.

my inventory works on a basic level, although i still haven't made it possible to swap items
i'm still too lazy to work on the saveload system, that's an even bigger pain than the inventory itself

That sand is beautiful, everything else is really funny.

...

It's funny because even though the animations are shit, you know it's a 1MA gamedev and there's less expectations, compared to say, Mass Effect. You're doing good, user.

simplescreenrecorder

Why is architecture the hardest part of games?

JUST

Also re: last thread about saving shopper data

It's not really a huge concern after all. If I treat it like RCT where you can save or load at any time, that means I still need to offload every single shopper in the store and their current state (or approximate state) so it can be saved to disk and reconstructed.

My concern was about maintaining shopper data after they've left the map so that some shoppers were persistent and do repeat visits. But, if I expect the player to save/load at any time, then there isn't really any harm in maintaining a bunch of shopper data, anyways.

For example, between all my shirt variants, colors, and states, it needs about 48 bits of data to draw the shirt. That's 6 bytes of information. If I assume a hard maximum of 65536 shoppers with 128 bytes of information each (this could include income, shopping history, state, etc), a save file would be about 16 kb at most (at least for holding the shopper data). Very reasonable.

i wish

What's the biggest road-block for game devving?

Mine is split between getting overwhelmed by the scope a few weeks after I've started, and reaching the point where I've got to start replacing all of the place-holder graphics with actual graphics.

lack of motivation
i'm lucky that i no longer find videogames fun, or i'd probably never gamedev

Shifting gears from prototype to production.
You know those first few levels could be better but you have to stop deleting and remaking them at some point.

any good programs for conversation mapping? i mean, this functions, but it looks like cancer.

Ayyyy

useless/10

Dio's stand is a nigger?

The lips mang

There was one program I found called VioletUML, but it was more for programming things in C#, but you could adapt it, probably.

Here's another shot
The base mesh itself comes from MakeHuman, I simply made some adjustments from what I was given to match what I wanted.

pls no bully

it needs to look eviler

Oh hey, another question.

If I have the sim playable at different speeds (eg pause, 1x, 10x, etc), should the shoppers still look around and blink while the game is paused? It'd be entirely cosmetic. Alternatively, I can just have it remain perfectly still

He looks like a nigger because of the wide face making the mouth look more squashed and nigger-y. You should make the facial shape way sharper and narrower (look at the facial width compared to the neck in ).

I made this for myself the other day, it's not the greatest and works on the browser but I can share it if you want.

Honestly looks a bit hard to follow. then again, i'm autistic as shit with stuff like this.

that actually looks fairly nice. thanks, user.

I didn't pay attention so I don't know what sort of game you make, but I think that that would be funny, if pulled off properly. Something bit like vid related.

Don't use human models for stands.

What's the deal with Jojo?
All the characters look gay as shit.

Agreed, although my example was a bit convoluted.

Here's what your map would look like (excluding all the text).

There's something about Japan and manly men being homos that was told to me a while back but I don't remember the whole thing.

Some of the stands have tits though.

you just need to git gud papu
sketchfab.com/models/d3abbc3952cb4bfe804550991b1893c5

Making the grocery store sim

I took some time of to play games expecting to get sick of them pretty quickly but i still haven't
What do?

Then I think that would fit. Basically just have all the people pretend that the time stopped. Things fall over (if you have something like this), sounds continue, animated textures on billboards continue, etc.
I think that would be a nice touch.

Tits sure give me a stand.

Your point?

you can use human models for stands

It still wasn't done from a template.

well what do you expect a stard template?

I just grabbed an old dude I made a long ass time ago and turned him into a stand, you can use human models as a base, just need to put some effort into it.

Yeah you need to tone down those lips.
Also this model just seems too…pudgy?
It really feels like it ought to have sharper lines

So how hard would it be to make a character creator that focused on body type, and tie in game stats to the body shape?

I want to make a "survival game" where one can judge some one eases "build" before they get into a fight. just a basic agility vs strength spectrum. where different weapons have a required strength to use effectively and then slow down the more you get past that. along with slowing down the sprint speed.

based on your reference, try to make his face more thin and angular, to capture the reference's jawline and cheek structure.

Look at uh, Panzer Dragoon? That one Genesis(?) game with the dragons

I haven't played it, but you had like 4 sliders and it drastically altered your ship's model

sorta like pick related

Not at all, once you have the character creator you would be driving various parameters with floats anyways, so just hook your stats (modified by some constants and/or normalized to range) into that instead of arbitrary floats.

I have done this, though not with buffer mapping. I personally don't know it it's actually more efficient than just drawing all the static props dynamically, but my line of thinking is that if a scene has a thousand static props with ~1000 verts, that's 1M verts. Assuming a vertex format with 32B per vertex (I'll just ignore indices for simplicity) that's 32MB data in GPU memory that's technically redundant, since you keep copies of all models in buffers anyway for dynamic draw. however, you've cut off 1000 draw calls. However, depending on your scene, having the oportunity to perform occlusion culling may be far more valuable than batching. I recently cut my scene up into lots of small pieces and started performing frustrum culling and I got double the framerate on my shitty laptop. I don't think there's a straight answer to "is it worth it?" besides "try both and benchmark", which is a huge fucking waste of time, but that's the joy of enginedev innit?

Scenes are considered game content, so at that point it's a game optimization and doesn't technically have to do with the engine itself, but how the game uses the engine.

Best C++ image lib? stbi_image is a pain in the ass due to multiple definition errors.

you need more signs written to it like
"Support Israel Goy"
"Donate some food boxes"
"Sales are final"
"Don't become a bad goy"

But if you want the engine to be robust, so it can be used for both games that have such scenes and ones that don't, you need to implement both and add the ability to switch between them.

There's not a lot of resolution to go around, unfortunately. It's kind of hard to fit more than a few letters in.

I mean I can have a big sign for when you mouse over it or whatever, but probably not with native sprites

I've been thinking about making some clones of simple vidya games or board games to get some devving experience, and maybe practice pixel art as well.

user, I can click the same spot 20 times.

Oi vey!

Is Godot actually good? It looks miles behind unity and ue4

Oh yeah, I missworded it. I meant that the minimal number of clicks to solve it is at most 5.

But if it's solvable in a minimum of 5 clicks, how did you solve it in 3?

that still doesn't make any sense

Please forgive me, O happiest merchant who is my greatest ally

I also added solid and hollow shapes to the pool of shirt icons. No happy windmills, yet. I need to figure out how to include them in a way that won't piss people off, just for having them.

Maybe I can sneak it in with a religion development arc or something, then I can just be overt about it (since it was originally a Buddhist symbol, after all)

"You can solve it in 5 clicks or less"
That's because to make sure that the initialized board is solvable, I simulate a random number of clicks (5 in this case) on the board.

But Lights Out is mathmatically solved - you should be able to determine the number of clicks as soon as the puzzle is generated.

I know. Implementing a solver is something I'd like to do as well along the lines, but I need to read the papers first.

...

Right, I need yarmulkes and burkas, noted for when I do hats

where is it

(checked)
Shalom goyim, nice chosen nite get you've got there.

OH SHIT HITLER HAS SPOKEN

I use it out of principle

MUH DICK

(checked)
shieeet

I'll accept your apology for now goyim, but do another screw up and we will cut the funding and publication for your game.

oh god why user.
i should not have seen sonic's feet.
user why.
user why.

Do you think Kirby gets embarrassed when you see his feet just wondering haha

...

reminds me of this.

if you don't tilt the swastika then it's technically not the reich associated one, and is the buddist symbol.
Also there's variants that are clearly of buddist/hinduism origin like attached pictures (last one means eternity, nodev for eternity).

Unrelated, but I've always enjoyed how the celtic cross looks

For me, lack of ability in certain areas.
I can't draw/sprite or level-design worth an arse. So to rectify this, I chose a game that requires a whole bunch of both. This is a good way to force myself to learn!

Shame I'm still rather arse at it.

honestly, who doesn't, it's a beautiful symbol really.
my personal favorite is the norse valknut

At least your "arse at it" is still better than most of the "good enough" indie pixelshit pumped out nowadays.

Personally, I think the octopus heart looks fucking rad.

Or you optimize it for the use-cases you actually have in mind and don't feature-creep the fuck out of it, Satan.

Doesn't sound like a really good reason tbh.

I removed the ears, added some more detail and tried to straighten out the chin. Could still use a lot of work though. And I'm not sure whether to just start with an entirely different base mesh, or just continue.

I-is there any way of giving my game visuals if I genuinely cannot draw?

if you pander to furries, leftists or both, you don't need to have art skills.
see: Undertale.

Rustfag here.

I swear to god please stop. My model is not as bad as the one on the right, is it? Fuck, I'm starting to get really fucking paranoid.

U anons think there's enough detail for limbs?
I think it's on the edge of being annoying to animate, but easy enough to draw for different poses; while having enough detail to see the hands/feet.

Also… this is by no means finalized… it's amongst the 10 other concepts I drew up for limb sizing testing; this is just the one I like the most thus far.

it's slightly less terrible.
honestly, when you add details and such, like hair, sunglasses, tacos, guns and mexican kids, it'll look way better.

suicide meter

[||||||||||||—-]

...

eh, here's one with a shadow too carlos.jpg

Trump says you should start off with a good, well-researched foundation instead of taking chances. Does this imply that if I need a very specific type of engine for a project and the only libre solutions out there are questionably built, I should write my own? There's also proprietary shit like Unity and Unreal, but those would be overkill and introduce too much overhead for the game to run on older PCs and toasters.

have you started cutting yourself yet? in that case, carve my id down, will ya?
oh, and post pics.
i'm joshin ya, mate. calm down.

it's a long running joke, user.

tbh user im running out of options, and now that I'm dieting I can't even binge. I really have no coping skills or activities anymore.

There are tons of engines available and your first project is never going to be perfect, do a bit of research on engines and you're bound to find something good for your project that won't be as bloated as unity/unreal.

user, jeez. calm down, man.
keep making games, it's cool. just some friendly bullying.
nice digits, too.

nah user it's not you or even the thread or the bullying, I'm not a little bitch

its just life, user, that's it. Shit's mounting

Take a break from dev, mate. It's all right to step back a bit and play some vidya you like or think is cute.

take it easy, user. nothing wrong with a short break. might bring a new perspective to gamedev too.

ye I know, im just mentally ill and im stressing over nothing.

post cool reload animations to soothe me pls

Don't kill yourself user, I want you to complete your game to be able to pirate it.

Will you give out steam keys for this thread?

of course

...

The second one and the last one killed my sides.

Is that a beard or the shadow of the neck?
Is it like some sort of pink skin alien person?
Are the hands so tiny that you would not noticed the character making a finger pinching action?

your model is a good model of a really ugly person
Personally I like it

protecting the user's freedom is always a good enough reason

automata mods never ever :^)

underside of the helmet/cloth, so shadow, albeit poorly done due to it being a concept piece

nah, just color of the clothes, during movement frames you'd see them move so it'd be clearer; as that's where the "life" is breathed into pixel art imo

that would be possible w/animation + shading, humm, that's a useful take on it; thank u.

I appreciate the constructive criticism user, thank u

that literally extends to everything that you do not just vidya dev. Theres always risk of failing in doing something, most of the risk comes from outside, our job is to minimize the risk with things that we can control.
Also just use unity or unreal and be done it with lol

I did a thing
Also a gondola

For what it's worth, I enjoy that shade of dusty purple

how do you batch render in SFML?
say i want to draw a gazillion tiles, how to do that?

haha, now i need to make the materials not suck and add furniture, i even bought an asset pack for this

smh fam i was just joking around, now if you really wanna be bullied post your drawing, and let the autist screech to no end
stick to 3d, less armchair critics

looking good, time for wall texture

Tonight, I hope to have my shoppers rendered arbitrarily with clothes and stuff

I actully thought it was some weird insect person

I really like that purple too, w/some monochromatic'ish style, and honestly setting up the proper color palettes is really hit/miss for me.

If anyone has tips of who to look at for quality vids on color theory, and such related topics; I'd appreciate it.

silhouette is looking quite good, nice

i believe in u

haha, nice, I'll work on that

vertex array or render texture there are probably more methods but those are the two I use

When you're ramping between two colors, don't just go from A to B in a straight line. Make it kind of curve a bit. Also make sure that A and B have different hues, it's easy to vary the brightness and saturation, but a lot of people just keep the hue the same.

Also also, don't be afraid to mix in grays. The left half of the color picker/mixer is usually ignored by people because they focus on saturated colors.

That's a smart way to visualize it, I appreciate the wisdom user

WHY ARE 3D ASSET PIPELINES SUCH A PAIN?

I was trying to make Blender and Unreal play nice with each other.
Fuck. That. Shit.
Importing models? That works.
Importing models with materials? That works.
Importing models with materials and collision models? That works.
Importing models with materials, collision models and multiple LODs? Fuck you! Why would you want to do that anyway?

While I was able to get Unreal's automatic LOD import to work thanks to a patch someone wrote for Blender's fbx exporter, it only works for models that consist of a single object. If you have a hierarchy of multiple objects, Unreal only imports the topmost one. Blender exports the entire thing though. I double-checked.

This is from one of the gamedev Tumblrs I follow
I thought it was amusing.

maker of thread and the shitty-er of the two models up there

i just wanted some pointers on how to texture so i posted what i had and now i don't even want to work on it anymore switched engine and probably going to just fuck around with other peoples shit from here on out. don't even have the energy to play games anymore. probably go back to my 20 hours of sleep sleep schedule here pretty quick.

Ask me computer graphics things or complex issues, I'm bored.

help the texture guy fam

help this guy

how do i procedurally generate realistic terrain? i wold like to create a island chain or peninsula that is packed with mountains and valleys that empty into the sea.

SFML last time I checked is pretty bad (I believe the drawing instructions use deprecated OpenGL and such for example). It's like a shitty middle ground between GLFW (almost barebones) and SDL (some things). Just use SDL for games.

Now what you are asking here is more related to OpenGL (or DirectX if you plan to change) than SFML. For batching you need to use vertex buffer objects (VBO) to define "batchs", and a sort of "batch manager" to manage / create multiple vertices that will be sent to the GPU during render cycle. There is plenty of tutorials on this on the web it isn't really that complex (specially if they aren't animated)

Will there be a any decent replacement/alternatives to unwrap as the same way vertex animation got replaced by skeletal animation? I am curious about this. So far there is that ptex thing but it doesn't seem to be well suited for games since and it doesn't work so well either with hard surface models.

Also if you want linear ramps, this is an okay site
zonums.com/online/color_ramp/

How autistic do you want to be? You could go full sim mode and model tectonic plates and ocean currents, or copy what DF does.

What I usually do is make some kind of noise map with values between 0..1. Do some kind of smoothing and optionally do some kind of perlin/fractal shit to create variance. Repeat this for height, humidity, etc., and you can create biomes and oceans and shit.

Good site:
redblobgames.com/maps/terrain-from-noise/
www-cs-students.stanford.edu/~amitp/game-programming/polygon-map-generation/

www-cs-students.stanford.edu/~amitp/game-programming/polygon-map-generation/demo.html

For height based terrain generation, you first need to create a mesh, this mesh is divided depending on how much level of detail you want for the result (more vertices = more detail / slower), you use a "height map" to define the height of every vertex in this mesh.

This same idea of using a map + setting the height of a vertex is used also to generate procedurally generated terrains, the difference on this case is that you have a function that you evaluate at certain point x, z, and it gives you the height f(x,z) = y.

I recommend starting simple, first with a simple mesh and height map (height map based terrain), generate the terrain mapping the height map to vertices' height. Then start to understand texture mapping based on height, then introduce slope based textures (using the vertex normal). At this point you should really try to understand how to divide the mesh in different "batches" and utilize a level of detail algorithm. Understand that we are just using height maps, just a random static greyscale picture, Once you get all of this now try with procedurally generated content.. Usually the same algorithms used to generate these height maps can be used to generate procedural content. Libnoise is a decent library for this kind of experiment they also got a terrain generation tutorial libnoise.sourceforge.net/

And here we are talking most about 2D / height based stuff so… no caves, for more complex things you will have to read some papers and graphics gems. Even this height based approach I recommend you take it seriously and read the papers where all this stuff comes from.

Is it 3D or 4D perlin noise that's needed for caves?

You can really do whatever you want to achieve caves. Some even just first do the height based terrain and then do another pass to define "caves" using different algorithms, shit like DLA or Perlin worms can work for this kind of stuff. But usually 3D noise is used, it gives shit results from what I heard though.

just want to make a survival game where the mountains will restrict movement to the point people have to trade in their valleys and whomever controls the mouth of the valley would control trade to the rest of the server. and if you have a survival game like ark people just keep building in a handful of "optimal" positions never exploring for a good place to settle. and no context is ever given to neighbors beyond who you can loot. but if you make people play nice with each other locally and let off steam with players from different valleys… you can give conflict more definition. make mountain passes matter seal them off or march threw them. inspired by the Greek city states and their history. i am bound never to play a game with such supreme autism.

If you generate values between 0..1, then it's compatible with pretty much every existing noise method. You just need to tweak the input/output so that it gets biased correctly, and then do something like anything between 0.80~1.00 is a mountain, and anything between 0.00~0.20 is a river. Then you'd just output this into the appropriate terrain data structures, and discard the noise map.

are you rustdev?

danks mang, already seeing some legit choices

3D noise is easier, 4D is preferable for effects over some rate of change (time is more common 4th-D).
You can also use 2D for caves, like slices of noise (XY slice, XZ slice), but it effectively becomes 3D noise just interpolated in slices; although, it can be more efficient than a blanket layer of 3D noise if you have a specific range that you want caves to form within.

diamond square alg is one approach, and can easily utilize a heightmap

mips mang, those high frequency/aliasing textures in the distance burns
looks neat though

just use world machine for the heightmap, and in-engine tools to form the mesh if available.
Proc-gen is a fucking rabbit hole user, just get the map setup asap, and implement your core mechanics; then go from there… do proc-gen shit after you've finished the main mechanics.

Ptex extensions are kinda recent these things take time to get mainstream. I remember using a computing extension to write onto images using atomic operations and the extension has been there since 4.1 OpenGL I think, it was so fucking useful and had so many uses for computing things in shaders and it wasn't really something mainstream and lacked support. Another case there was a extension in 4.5 (I think) for sparse textures I got excited because I could really break the standard using this for a college project but then I realized I couldn't really use it, it was just too recent and it wasn't properly integrated with other functions. It takes time and tears. UVs still really flexible and also transferable, I think it will take something else to replace unwrapping.

that was a college assignment I didn't really care about aliasing at the time, the whole thing was about generating terrain + lod, thanks mang

This seems more like a problem for a 3d modeler to be honest. My bet is that the UVs are all shit, I heard texturing isn't that easy.

no i'm the other guy the one with the shit texture

Proc generation isn't that easy specially for someone that hasn't worked with terrain generation before (that why I recommend starting with easy/nonprocedural things). As the other user said it's a rabbit hole (a neat one though), there are some tools to generate these things without too much headache.

this looks like a some giantess porn fetish game that the tag on DA would love

Got a palette for my hair done

How the fug does ptex work without a uvmap? Is it like subdivisiding the model in texture space only and painting the new vertices? Nothing online seems to explain this shit.

I just learned about Triplaner Projection.
It's like fucking magic.

Triplanar
Looks like top quality work, nice m8, and yeah triplanar projection is pretty useful

god damn it

Progress! I can import stepmania SM files now (but only normal notes, not holds or things like that).
I have almost enough setup to begin working on real features. The last other thing I need to do is make a file dialog that isn't shit on mobile (I put the sm file in the top of my phone's storage because otherwise it was a pain to navigate to).

Disclaimer: This post was copy-pasted from my post on halfchan, but one of them had to be the first post anyway.

Oh ho ho, you're gonna get bullied, faggot

what the fuck is this.

You tell Mark that he's a jewish faggot, in order to make him feel emotionally alone and insecure

...

I giggled

Alright, so. I've got my palette saved as a bunch of hex values and it loads correctly. Now I need to assign weights and semantics to groups of values, and I have no idea how to do this sanely.

For example, I want to have a "blonde" hair group which includes the base blonde color, and each color in the blonde ramps. I also want to be able to assign a weight to each one, so that it's more likely you'll get the more blonde tones, than the brown ones for example.

First pic is how my palette is divided up, second pic is how I'd weight it in the case for the "blonde" group. (it would work similarly for the "brown" group, except it'd use different ramps).

I'm looking for a way that makes sense to put it in the data files, ideally without having to do something like 0=32, 6=16, 10=16, etc for every fucking color/weight, for every group I want to do. I mean I'll do it if I have to, but it seems awkward.

Fuck me, what to do.

Oh, I know. Since it's laid out fairly logically, I can do something like
… for each group that I want to do. I don't know if that's 100% perfect or too arbitrary, but it does give me the control I want without having to do an index:weight for every value.

Pretty sure there's a way to use nearest neighbour interpolation for GM's scaling. Maybe.

Also, why not just make your shit tiny, draw it to a texture, scale it up, like you're doing. Then, draw regular sized text at the native resolution? You can even break your UI up to use its own texture so you get something like:

Yeah this looks reasonable, I think

How viable is it to use downscaled textureless raster voxels a la Voxelstein 3D/Voxlap nowadays? Would like to use them for something that's not a shitty do-nothing sandbox.

Guy making the VS mod for Ark Survival Evolved here

I got the remaining model files for the rest of the VSs that the first collection did not have.

Friend still blowing me off. Frustrating.

I'm no expert on this, but my understanding is that GM has linear and nearest neighbor. By default, it uses linear, but you can turn it off via texture_set_interpolation(false), which sets it to nearest neighbor. If my game is pixel-based, should I assume that linear interpolation won't look good at any res?
I don't know shit about layers yet. (Or Surfaces, as they're called in GM.) I'll have to read up on them. I've tried handling this in a similar way, by making the font's default size huge and then using draw_text_transformed to scale it down, but the results were lackluster.

Why would you use a gamemaker anyway? It's a huge piece of shit. Love2d is far superior.

also a way to set it to default to nearest neighbour in the settings. and yes, linear will always look like shit.

for that, i'd suggest either picking a different font, or figuring out how to draw letters from a sprite sheet or a texture.

(checked)
Hmm yeah so far only Nvidia showed a minor interest for it and made a small presentation somewhere, for AMD one I couldn't find one at all.
Sparse textures huh? From the first few quick glanze I made over it, it looks like it is a function how a texture is rendered and is thus more in the programing section so nothing at all for modelling process.
did you split uranium-325 for that one? :^)

That's a relief, in some ways. If it's assumed to be shit, then I don't have to worry about trying to make it not shit. Now if I can get fonts working, I may not have to rework the rest.

Because it's what I know how to use. If I switched languages/engines every time I saw somebody call one shit, I would never get anything done. I do intent to try other engines for my next project, though.

Gamemaker isn't just shit, it's hindrance because of its dogshit language with which you can't do anything useful. You'd know what I mean if you tried using actual fucking languages and not this special snowflake abomination they have in there.

This. Using gamemaker when you can use Godot is insane.

Not sure about Godot, but, having used Unity, Java, and a tad of Python, I appreciate that GM lets me get simple shit done quickly, without having to on ridiculously long, convoluted bits of code for it.
I'm almost curious what GM prevents you from doing by being so "limited", aside from being able to have Massives.

godot is like unity/GM but with python style scripting

plus it's completely libre and gratis. pay nothing, use forever and for anything. it's free software. and it's getting very popular.

That's more than a little hyperbolic. I can name several decent games that have been made with it just off the top of my head.
I've already said that I intend to try other engines, and therefore languages, when I'm done with this project, so you don't even have to convince me, but I'm also not about to throw away many months of work without very good reasons.

Nah atomic operations is a multi-threading term. In a shader you are basically working with a shitload of threads, that extension let me write in a certain pixel atomically, without it you would get different results on every frame at that pixel. Also shit like a moving average operation of neighboring pixels, which you can't just do normally.

In computer graphics there is always a sad history of some random extension that does everything you want but you can't really use yet.

Ah alright.

Is it because not enough Accelerated Graphic Cards or Programs are supporting those extension yet?

Put all tile textures in a texture atlas (no worries if it gets huge, even a 4096x4096 atlas is just 64M), for each tile construct a pair of triangles at the corresponding world position, with texcoords set to the offset in the texture atlas, put this in a buffer and draw. You can either fill the buffer with the entire map and just pan the camera or add/remove tiles as the camera moves. You can do hardware-accelerated layering by simply placing multiple tiles at the same xy, but with different z, then do alpha blending like usual.

Another option is to abuse glTexSubImage2D to draw the map directly to a texture. You'd have to perform all layering and blending manually, but it's possibly faster (though even laptop graphics should easily be able to draw the few thousand tiles you're working with). Additionally, camera panning is more tricky. You can't scroll the texture data, so you'd have to essentially allocate macrotiles of some suitable size, scroll those and when the player gets close to the edge, blit the upcoming tiles. I have a feeling that's gonna be slower than the vertex solution, since for instance if a tile is 16x16 with alpha, that's 1kB, whilst a tile described with vertices could be as little as 32B. The blit solution may however be good if movement speed is guaranteed to be small and the tiles are also very small. You can also cache macrotiles generously, possibly even to disk.

Fuck me.
Well, I need to redo the powerup entirely, so I guess this is just more reason to.

friendly reminder

NO GAME, NO LIFE

Quick question, what kind of techniques do you guys know of for compressing textures and file sizes so that a game can be run on toasters? I really wanna know how it's done.

MD2 is a Quake 2 format, the Id scale has always been 32 to 1. This has nothing to do with the Irrlicht renderer.

Good to see people appreciating LP.

LP3 never ever

I always save my things as color indexed png

Fam
You mean U235. Even then, thats the non fissionable waste product after the reaction. People want U237 and occasionally U238 to power their reactors

There was a Lost Planet 3. It was western developed, and sucked.

infogalactic.com/info/Lost_Planet_3

Never Ever on LP4 though.

That Nicolas Cage simulator ruined everything!

Nick Cage is good at ruining everything.

I've got another question for you AGDG eggheads: why is it so hard to develop enemies that act and react believably? I don't mean having realistic AI, just enemies that will try and take basic cover or run away once they're in danger. Why do all enemies in most games just make a beeline for you and try to hit you next-to-point-blank?

like the picture and models appear, but only ambient lighting reeee

I would imagine most AI is programmed to prioritize shooting the player over preserving itself, since enemies are infinite and death means nothing to them.

if i recall correctly, the gears of war AI was quite good at taking cover.

but gears of war also had a proper cover system, I think that's the difference because it lets them program the AI to take cover much easier.

not mentioning that waiting for the enemies to pop up their heads is both annoying and boring

I assume this is a game design decision over a skill hurdle. Turns out lots of things are annoying and boring when you translate them 1:1 into a game. Driving in a city isn't very fun, but implement it like GTA or Crazy Taxi and it becomes much more fun at the expense of realism.

AI that is designed to run away at any sign of incoming trouble is absolutely annoying to play against, compared to AI that actively tries to attack you

Mostly because of lack of compatibility with other extensions, poor flexibility and/or inability to use certain tricks/instructions/objects from the standard with whatever a new extension provides… aaand then there is the hardware support problem. At the academic level people do crazy things with new extensions and possibilities within newer GPUs but these things don't become standard until even decades.

It'd only be good in a controlled environment like a roguelike where resources are extremely limited and you need to make every item count.

If I make any significant progress in creation of my shitty RPG Maker VX Ace game, should I share it here or not? I just started doing it yesterday, because I simply tired of my meaningless existence and need to do something.

Sorry for my poor English.

look at Halo:CE/H2/H3/Reach's AI.
It's very good in this regard, as each "type" of alien has different behaviors that it fulfills; per the perceived personality/race/marine type.
Which actually, this isn't that hard to achieve, and the algorithms used (behavior trees, but with "unlockable"/"locked" branches to said tree) are very easy to use; in addition to having a "perception system" so any behavior is prompted by "in-game perceptions" (not cheaty, as that was what they were explicitly avoiding).
Even then, they dumbed down their AI for the player's benefit in Halo:CE.

Some races are cowardly, if low rank, and some are courageous, if high ranking (in the covenant), or some are even suicidal if given such an order by their "holy overlords" (as really the covenant was waging a jihad on the human race), and so on.

Because you aren't doing the lighting stage?

Taking cover as an example, the more complex the surroundings the harder it is to create an AI that finds cover intelligently without killing your CPU. Complex meshes make it really hard to determine what is decent cover, and what is leaving the AI's ass exposed.

Generally though it's because normalfags don't even know what an AI is and by the time they have played the game enough to start noticing, media told them to get excited for another game and they completely forget about the current one.

LP3 does not exist

I agree needing to kill enemies that run away is annoying, but that doesn't mean that enemies with self-preservation are necessarily bad, just that mechanics need to be in place that reward the player for exploiting it in their enemies.

FEAR also had pretty innovative AI for it's time.
I understand that enemies would take note of the last place they saw the player, if you were to run away or hide. Then they'd aid their weapons and grenades accordingly.

In PoE, humanoid and monkey enemies will automatically flee if they've been ignited (usually from a critical hit from fire damage). Fleeing isn't TOO annoying here, since they basically take a few steps back, pause, and then return, but it fucks up their clustering and makes AoE skills harder to aim.

Actually that's another thing with PoE; enemies are fairly passive. Some bosses and uniques are quite aggressive, but regular white enemies tend to do some kind of animation, walk up, stand beside you for a second, and then do a slow swing, giving you plenty of time to evade. What's great is that you can physically move away and this is a perfectly valid form of defense. The downside is that getting hit hurts since armor mitigation is pretty ass in that game.

yeah, GOAP is pretty neat, and especially so if one harnesses a type of perception/blackboard (basically memories) system.

Do stuff, share it, ask for feedback.

Thanks, I'll do it. Sorry for IP hopping, I've only dynamic VPN

MoM user here, and Hello Again Everyone, its that Time of Week!

This time we will be looking at the Reptilian Warrior Class, his gear, and his perks. I hope you all enjoy!

On an Unrelated Note, Happy Mothers Day, Everyone!

Get it?… Mothers Day? HAH

Will your artist draw more lewds?

I am, if I wasn't it wouldn't be rendering anything at all. I have no idea why it's rendering the ambience but not anything else.

I had to make sure it wasn't mother's day where I live, u freaked me the fuck out, u mofo hah, get it, mofo, I'll show myself out…

Seriously though, glad to see you're still alive, and lewds… we need them.

He posts pretty often in his own place!

It depends on who does fan art for me. They was a cool guy last year who did some lewdish fanart, but I haven't seen him around here lately. He was working on his own game here too, so I hope he is okay.

It's more than a month away where I live but still made me check

It actually is Mothers day where I live right now, in Bongland

Is it good?

Fair enough.

That's sad, it's like menu designers, they get no appreciation.

Sure. You'll get a couple people jabbing you because RPG Makers are pretty casual-tier engines, but a game's a game.
I'd be happy to play it, at least.

I'll play it as long as there's no furry shit, and you don't use the default assets.
That's like, ResistJam tier of lazy.

Yes share.

Great paper on it. Really cool AI.

here's a good general resource for game AI

No furry is guaranteed, though.

Thank you.
Just to be clear, I've no experience whatsoever in game development. I want to try few experiments with this engine. My idea is too advanced for vanilla RPG Maker, and I need to know can I do it or not.
It will be a really shitty game, of course, but I need motivation to do something with my life, and best way to do it is just trying.
And English is not my first language, so it will be additional challenge for me.

If it become slightly playable, I'll share.

If you need people to proofread stuff you can always make a thread on >>>/agdg/

webm related

Literally cannot figure out why it wouldn't work.

So I just took the tutorial source code and refactored it to be exactly like mine, swapping out all of it's classes for my own and now I've got it working.

I fucking hate OpenGL lol, I bet somewhere I missed a number and I would never find it in a million years.

Does anyone know of a good (non-hentai) RPG Maker game?

Thanks.

...

Any anons using monogame?
I'm debating using monogame extended, but it looks like I would have to refactor a lot of code if I wanted to use it.

I just use XNA because I can't fucking set up environments to save my life

Also I feel like I might be approaching this the wrong way.

use enums

Enums are functionally identical to bytes (or whatever backing type they use, default of Int32) after compiling. I'll still have a bunch of indexes at the end of the day.

I did write up on paper a bit format for serialization, which accounts for 140 bits, but that doesn't help here, since it's wasteful to parse out the bytes from condensed bits every time. This would be the live object used by the game, so I can be a bit more verbose with the memory it uses, but it still feels a little … something off

I've been playing with it lately, it's pretty nice and I'm considering using it. My second choice for 2d is Love2d but Monogame has some nice benefits. The obvious downside is that it forces you to use Windows and Visual Studio, but you can at least export to Linux. VS does have its benefits and C# is easy. Try it out for yourself. This is the tutorial I used and it walks you through a little sample game, it should only take a day or so to copy it for yourself to see how it works especially if you have some C# experience already: rawgit.com/MrGrak/Monogame-Getting-Started/master/index.html

this is pretty much a correct way to approach this, but you might want it to be a struct instead

I think you read that wrong, I said "monogame extended", which is a community extension of monogame.
My project is already well under way.

Java > Mono/C#

space combat progress

Needs 3D holographic sights.

My mistake. I hadn't heard of it before. Better to learn about it now than have to refactor later and end up in a similar situation. Depends how much time refactoring will save you later on.

Provide proofs

OpenGL is really neat, you might be doing something wrong don't blame the API, it isn't that much different to DirectX when it comes to setting something as simple as deferred rendering.

Is something like this typical? I'm defining each image on my spritesheet with a name and coordinates, rather than just shitting out a ton of coordinates.

hastebin.com/nizepuxahi.txt

The camera feels bad, otherwise, it's patrician quality.

The API is unironically pretty awful and most people acknowledge that. But still I lay the blame on me being unable to spot the problem, this isn't the first time though I've had to rebuild something from the ground up to get it working with OpenGL.

you don't set deferred rendering, you set it up.

not controlled by microsoft, faster, more portable.

Java is a shit language and you should feel bad

C# is literally a bad clone of java so microsoft can force their "ecosystem" onto people and lock them into windows world.

...

No signals work for this button. It seems to happen to any scene I instance under a container in another instanced scene. Help I have no idea whats happening. The properties are unchanged please never ask about them to me or my wife's son ever again.

im talking about godot here

do everything in script stop using the gui

Not sure if you simply haven't added sound yet, or you fell for the no sound in space meme, but a fair amount of those things should be making sound:

According to the filename it's VR user, that's just what it looks like when you move your head.

how would the gui be the problem here

conceals information and doesn't really let you know what's happening.

if you do it all in the script you know everything that's happening.

Also, the levels were constructed around it. Yahtzee makes fun of chest-high walls, but building your levels in tandem with the AI that's supposed to act in them is consistently more effective than trying to jam a general purpose AI into whatever area the level designers happen to make.

I'm taking a break from deving and i'm going to idea guy for a little while. I've been trying to figure out a boss encounter for the time slow mechanic I had planned. Tell me if this sounds cool as it seems in my head

It wasn't working because it was in a container

what did it mean by this?

Idea is this: Boss rush game wherein you play a sniper in a world-spanning Kowloon-like city. The game, unlike normal autismo sniper games, revolves around going 2fast4u because your enemies are all mechs. Each level is a boss arena wherein you are chased by an anime-like mecha who you have to take out with your outdated Anti-Tank rifle. Anyone got advice/mind if I dump some ideas for critique?

Why you do this? Set up demo server so we can shoot each other in space together.

Agree with this, you don't need to go full "Elite Dangerous canon fake sound system inside the ship", just those things user mentioned is ok.

That sounds really cool!

...

i like it

Placeholder model. I just took the kar 98k from DoD: Source and made everything a bit bigger. I'm more of an audio/level guy than anything else and I like to get that stuff out of the way first.

Sounds like it'd take a lot of work, but if you can get it going it'd be rad as hell.

apart from that it sounds pretty good tbh f a m

I mean I can't make art apparently either but it'd be nice if I could.

It's arbitrary as fuck, but its interactive now

I feel like I can make a very clever joke about this, but nothing comes to mind right now.

Nice dubs.
The idea for the mecha is that each one will be a different beast entirely. This could affect weapons, speed, design, weak points, etc. Your gun is designed to be used to take out specific bits and pieces. You would also have mines, traps, tripwires, and grenades to purchase or find which will give stun or knock down your enemy. This would enable you some time to shoot off an arm or a leg.

You can make games in watercolor, it just takes a metric fuckton of effort.

Or if he holds his hands up and was black

Why is input handling so annoying?

Does anyone know of a good hierarchy/architecture for input?

Here's my problem:
engine has 3 callback functions that receive input for mouse, kb ,etc

I need to handle this input for the GUI as well as to control the players entity. I also need the input to tell the camera where to go/look. And I need the camera to tell the renderer how/what to render.

Should I just say fuckit to the camera being it's own thing and instead put it into the user entity's class? Should I add a user controller class that controls the user entity and the camera instead?

How have you done it?

XNA has 3 different input types (keyboard, mouse, gamepad) and 4 different streams (keyboard, mouse, gamepad, chatpad). A chatpad is for the x360 controller, but it's really just a keyboard with a player index attached to it. A gamepad also has a player index. Mouse and keyboard don't.

So what I did was make a wrapper that holds a current and previous state, and exposes a WasJustPressed, WasJustReleased, WasJustHeld (all which poll both states) and IsPressed and IsReleased (which poll the current state only). Then I made a manager object which can create an instance of each input stream (4 chatpads, 4 gamepads, a keyboard, and a mouse) and it updates the state every frame as needed.

XNA internally holds the state as a uint, and the state manager I made still lets you check individual current state so you can poll manually against non binary things (such as analog stick state)

so did you player controller character just check the wrapper to see what it should do?

and how did you handle camera, it's probably a bit differnet in XNA, but do you have it owned by the player character class?

When I check input I do something like pic related. I mean, something has to check it, but it always made more sense to chuck it on a global object, rather than a player instance.

XNA uses matrices and shit. I just call something like Matrix.CreateTranslation(x,y) and give it to the start of my sprite batch call, and it offsets the coordinates for everything appropriately.

To be honest I haven't had much experience with that, because I usually give up on projects before they get that far

Also have more spastic goodness

I should make a game about horrifically bad watercolor art.

I'm using just matrices too, I just have a camera class atm to hold them, but I think you might have the right idea. I don't really need a dedicated class for it.

fuck

just like install pulseaudio mang XD

How does this look, any advice?

Turn down the fucking brightness, holy shit. Then we can actually see it.

i will never grow old of the source aesthetic

I have all my camera related stuff in the rendering c file, and have a global function to set it's position/rotation. You can probably find a safer way to do it, but all the camera shit has no reason to be anywhere other than the renderer. One thing you can add is a debug mode where it'll print an error if it's called more than once a frame or something like that. The renderer can't render more than one camera at once unless you're doing something like reflections and even then theres probably a more optimal way.

If you want to update the camera to the players head every frame then you get the players head position, however you want to do it, and update the global function.

One thing I've learned in the years I've spent engine fagging is mindlessly throwing in OOP just fucks everything up. You'll never need to modularize half the shit in the core engine since you will only use it once, and wanting to change it will result in a complete rewrite. A lot of console devs like Insomniac and Ubisoft's Assassin's Creed team uses C for everything and breaks the rules like crazy. Even fucking source and unreal have globals everywhere, like for their console commands.

If anyone screetches at me, I learned OOP is a meme on here when some fag posted Mike Actons "Data Oriented Design" talk.

By brightness do you mean the fog?

I mean this stuff.

Reminds me of SotC tbh.

Post the vid.

It can afaik with viewports, but won't bother with this so your point still stands.

Literally me right now. I began (sort of, it's a project built of a uni project) last year and went full head over heels C++ OOP. I burned out in like 2 weeks after passing pointers like crazy. I've finally come back to it and I'm making it much more "C with Objects" this time round. I'll take your advice and the other guys and I think I'll go with global camera.

That's fog. There is nothing actually over there yet. I am basing this off of the Kowloon Walled City that used to be off the side of Hong Kong. It was a no mans land that neither the British government or China wanted to touch for fear of starting a war. As such, due to the intensely small area available combined with large influxes of Triads mingling in with the population of political dissidents, poor, and refugees, they simply built upwards until they risked hitting planes. According to former residents the city was a maze and you could not see the sky unless you went up. There is a Japanese theme park based off of this since the whole of it was smashed down in the nineties.

What dynamic range are you using, like, half an f-stop?

Don't listen to the other guy, the look and feel is amazing just from this webm. Flesh out the city streets and alleys, add some chinese character signage, I'd play it.

It took me forever to wrap my head around how to pull the shit he's talking about off. There's a thread about it here to
>>>/agdg/27468

He even trashed the ogre engine for this shit bounceapp.com/116294

My first engine I had a ton of shit in all my constructors because I tried encapsulating everything, only to learn everyone uses globals eventually.

Really? I thought it was in Chinese territory. Granted, said territory was surrounded by British claims, wasn't it?

Bonus vid of Ubisoft talking about how they use a global array for entities with a singleton to initialize it.

Thanks.

I just read the bounceapp thing, and damn, I need to get good, I can't even understand most of what he said or why certains thing in it seems to be bad coding.

I didn't get much out of it either other than the main point. The point really is they forced OOP so hard that they overcomplicated the FUCK out of their engine and had a ton of redundant code that was both slow and inefficient.

What do you guys think of using stacks like this to show shelf space, and using colors to show fullness?

Inefficient as in, hard to work with. Hard to read, hard to debug.

This is why neither of them wanted to touch it. The Hong Kongese were the only ones to ever do anything for the city; having done occasional police sweeps and busts to try and curb crime. This city was Bladerunner in the flesh.

I'm planning to. The main theme will be hiding in alleyways to find the right spot to avoid fire from mecha you are sniping.
Here is the ambiance without the Chinese music.

Elaborate?

I know sometimes OOP is annoying but I just cannot understand what the guy was trying to say about certains parts and that's what bothers me. I want to understand, I want to git gud.

gonna go a little devils advocate here and say he's a bit of an ass to trash on software like OGRE when he doesn't write any rendering engines that are open source. We can't exactly see his right way to do it.

This is an easier to understand one. It's not really relevant to the original question, but it shows why OOP and software rules gimp you in terms of performance. This stuff mostly has to do with performance, but after I read really deep into it I kinda learned that mindlessly using OOP will fuck you over. I use C for most of my core engine. That means, instead of a collection of classes it's a collection of functions and static important variables. If you do it this way you kinda learn that these are what classes are under the hood, but copyable.

Yeah, the code he gives is usually really small samples and stuff. The key thing really is that if you follow the rules mindlessly you'll fuck yourself over with over complicating everything. It also fucks you over in terms of performance because of how CPU's work. When you branch I.E. calling a function, etc.

As in, it's important that they are static. In C static variable are basically global variables that are only global to the file they're in. I didn't know that before I started switching over.

Do you write your engines from the ground up or do you patch rendering, audio and physics libraries together ?

The things that are bright are really bright and the things that are dark are pitch black. There isn't a lot of in-between meaning any detail that was there before is lost.
Saying you were using half an f-stop was me being sarcastic. Dynamic range is measured in f-stops and a disposable camera with it's terrible dynamic range still has 10 to 11 f-stops the human eye has around 20 or so.
If I'm correct in assuming you're using Source Engine, this should help
developer.valvesoftware.com/wiki/HDR_Lighting_Basics

fix the perspective

It depends really.

Haven't gotten to audio yet, but I'd just use a audio library because I use linux and don't want to deal with the cross platform shit. The most I've gotten done in physics was collisions and ray tracing before I started rewriting engine. I might do it from the ground up again just because I don't really mind math. I do rendering from the ground up, although buffer management is always a pain in the ass. I did my window system in X11, but I'll probably switch over to SDL because managing cross-platform crap sounds like it'll be a lot of work for nothing.

Basically if it's something cross platform, in my case because I'm a linuxfag, I'm just gunna use a library.

Seriously though, managing buffers is the next big hurtle after figuring out how to structure your engine, because making calls to your GPU is slow as shit.

So you did the whole rendering with OpenGl and not with a third party lib ?

It's just a mockup in PS
What looks off?

the shelves and the NPC are in different angles

To also add to the discussion, most proprietary engines for AAA titles still use a lot of third party libraries for skeletal animation/rigging, audio, physics, etc.

I can't think of any other ways to represent the shelves, though.

you could change the people sprites

Ehh… that's like the one thing I actually have done and made good progress on.

make them vertical stacks and when you hover over them either the adjacent shelves become semitransparent or a bubble of text apears naming the shelve's stats.

In the long run it's better to tie player input to the player instance since at some point (I assume) you'll want to disable player input for something like a cutscene.
My game in monogame wraps my player.Update function with a single boolean, be that good or bad, it will make things easier in the long run.

Oh shit, that's a good idea.
I can also have multiple "views" the player can cycle though, too.

Thanks!

opengl. Keep in mind most graphics cards are optimized with OpenGL and DirectX in mind. I doubt you can even do graphics on the GPU without using either of those libraries.

and vulkan. Forgot it existed.

i did a thing

physics library?

Update. Using Tf2 kong king models as placeholder.

shaping up nice

Oh baby that's starting to feel good

>itch.io/jam/agdggj01

HOW PROFESSIONAL

should I attempt something?
I've got nothing better to do.

In 24 hours? If you're some kind of maniac.

This starts in 4 days though, and has 160 participants. Shitty meme contest, though
itch.io/jam/1-bit-clicker-jam

I think that "clickers" aren't necessarily bad, just how you go about them. Stuff like clicker heroes is a direct example of just using time as progression, which is cancerous.
It would be interesting to try and throw in elements of strategy and the balancing of resources to that kind of format without using cheap shitty mechanics to make it last longer or drag out.
now thats some bullshit. I'll be saving myself for ludum dare at this point.

I have a rough idea of what I could do to make it interesting. No keyboard is kind of dumb, but then, I could always just make it entirely mouse driven and not have to really worry about things.

Anyways, I was thinking of having it progress in phases. The first is resource gathering. Set up mines, logging camps, whatever, to get your set of resources. The next phase would be crafting, and the third phase would be assembling your crafted parts into specific items. These items would equip some kind of robot or fantasy party for the fourth phase, which explore dungeons.

Or fuck I don't know.

That sounds pretty close to what I had in mind. Some sort of survival/resource management ordeal.
If you want to try that jam out together, shoot me an email

I'll politely decline, for now. I'm flaky enough with my own projects as it is.

so how hard would it be to have multiple hit boxes for different organs and have an armor layer that one would have to penetrate?

like an action game where someone swings a sword or stabs someone and then you (compare the two players relative speeds * and inherent bonus) to check if you penetrate lets say a breastplate then see if you hit anything important + stacking bleed effect based on weapon. i know world of tanks has something similar but how do they do it? do they have different colliers for each pice of armor?

I fucking wish man, but then my Dell Latitude would crash faster than a penny stock.
embed unrelated, just somebody's loonix music

Last for tonight.

What do you plan on doing with that ?

...

H-have you got any of the mech designs worked out…?

I'm planning on drawing up some concept art soon. As it stands I'm a really big fan of "junk yard" looking design wherein things are made of scrapped together hunks of steel. The idea for the first level mecha is that it was once a smooth, shiny, military looking mech with a very sleek look until it was left rotting in a pile of trash for a hundred years. After this the new owners made a suit of armour for it out of random bits of sheet metal, neon light signs. etc for a mix of protection and "street cred". It will chase the character around the streets and smash into buildings/shoot into corridors/grab the player off balconies. Its weapons are a flamethrower and a slow firing minigun.
This is, of course, up to me 1: actually modelling the damn thing and 2: scripting it with Hammer entities in Source because fuck programming new entities.

Sorry I'm a fag, I should have read the thread.

Make it so you can slide between their legs and put a sticky grenade on their crotches.

What kind of newfag made this? You?

4chan agdg has had all kinds of jams demodays notwithstanding.

If it's for Holla Forums, then I'd expect someone to actually talk about it before making it.

It's no issue. Most of an AGDG thread people tend to quickly skip through. To be honest I really like that idea.
As it stands there will be hub areas for every mission wherein you can interact, do side quests, accept missions, buy shit, and stock up for the next go. A big focus is to trip up your enemy and throw them to the ground to buy you breathing time and to have a real honest expense for every mission. If you need your items you need them.This, once again, is all based on whether or not I learn to actually fucking program.

I'd code for you but my computer is shit so I'm not sure I can run source.

I was making fun of 4agdg you goddamn autist

…by making a game jam? Because I was talking about the game jam.

I can run Source on a laptop but in-dev maps made on modern BSP engines have awful performance until you properly optimize them. Thanks for the offer.

* most Source games.

Main problem I see between what you're making and what you plan is that Mechs tend to not move well in Kowloon-like environments. Mechs are big. Kowloon did not have big spaces for the mech to stand.

This is actually a terrible way to handle input. It's all fine until you've used an event-based system once, then you'll never want to go back.
I actually wrote one for monogame (the version of XNA that's still supported). I meant to open source it, but never got around to doing it because I'd have to implement mouse handling first and I have more important things to do.
It was basically Unreal's input system ported to monogame.

That sounds… Really radical m8

bullet

That's sort of the idea. They get a lot of verticality so it would be a lot of running into cover to avoid being turned into swiss cheese. As well as this they'd be able to "video game logic" their way into buildings and smash down set-piece walls.

What's that technique/setting for lowering the resolution of textures the further from view they are? Mipmapping?

Yes.

this is really interesting thanks for posting this

Does it really matter if I have a long-ass list of if statements, or do I have to change them all switch? I don't understand the practical difference between the two. I've seem people say that switch statements are just easier to write (I disagree), but I've also seen others say that it matters on some kind of computational level. If I recall correctly, this was one of the grievances that people had with Yandev.

Frustum or Perspective, give me the red pill.

switch is more of an ifelse chain than an if chain

chaining ifs lets you handle multiple buttons in one loop, so I think you've done it the right way

(checked)
switch statements can use a jump/lookup table

mostly due to retarded repetitive nested if statements, using strings for comparisons, etc.

if it impacts performance/readability/usability, yes, otherwise if you prefer that other structure; then no.

Just looked at the code in question, this user is right about the series of if statements being proper here
u want to be able to handle multiple input directions in a single update tick

I see. After some brief googling to understand what that is, I suspect that this is the kind of thing that won't matter for the button input code that I just posted, but may matter very soon, in a segment where I'll need to write very similar code to loop through approximately 300 possibilities.

I mean, I suppose technically they could also be else ifs, and therefore could also be a switch, because it's only reading one button input from the buffer at a time anyway, meaning that only one of those statements will become true per loop, despite all of them (if there's more than one) ultimately happening within the same update once the loop finishes.

No, I posted their jam here for us to laugh at. Thats what I've done with every jam so far, including resist jam and the 1 bit clicker. I dont intend to host or manage one.

On languages that use a virtual machine (eg pretty much any modern language), a switch and ifelse chain should compile to the same thing.

You should avoid repeating yourself.
In your case, you have four almost identical lines of code copied four times. Something isn't right here.
You're also hardcoding constants.

ODE or Bullet Physics?

Give me a quick rundown.

the only VM based language that I really care about is C#, and generally this isn't the case for release code compiled in VS.

benchmark it, and decide from there; though the general rule of thumb is if there's more than 5 sequential conditionals use a switch statement.
also, if it doesn't have considerable overhead for what you plan to implement for your current method call it good enough and move onto the next thing.

probably better to find a blog of some physics engine guru than to ask here

well, one way of handling this that i ecnountered was something like this:
...#define VK_KEYA 0x41#define VK_KEYB 0x42#define VK_KEYC 0x43...bool m_bKeys[256];void GameOGL::HandleKeyboard() { if(m_bKeys[VK_UP]) { m_Player.Move(MOVE_FORWARD, (float)m_dTimeDelta); } if(m_bKeys[VK_KEYW]) { m_Player.Move(MOVE_FORWARD, (float)m_dTimeDelta); }......
the guy used if statements because you can be pressing two button at the same time

I'm making an FPS RPG to muck around with the unreal engine. I've set it up to use character attributes that fit the intended gameplay (Strength, Constitution, Dexterity, Intuition, and Perception). My intention was that they directly affect one thing in particular, and indirectly affect (or are pre-requisites for) many other feats; and that making any of them a dump stat would have greater repercussions than those seen in most other games.

CON/DEX/INT were the easiest to implement:
CON:
DEX:
INT:

Pretty simple stuff at the moment, but I'm not entirely sure what Strength should directly affect. It's going to be a requirement for what guns you can adequately use, amount of crap you can carry, and a bunch of others. Should I make it directly affect recoil received? High values reduce recoil, low values increase recoil, really low values knock you on your ass?

And then for Perception my intent is that it affects your senses, though admittedly I'm not entirely sure how I'm going to implement that. I was thinking it would sharpen your hearing/vision at higher values, but at low values you would be harder of hearing and short-sighted. Can what a player hears from a sound source be overridden on an individual level? Any ideas that would make that sort of idea better?

This code would make you walk forward twice as fast when holding up and W.

That's why you put the two values in a vector which you then normalize.

thers a lock on movement speed later in player class

in this case you couldn't use a switch statement since you need to it to be if and not if else
unless there's something im not seeing that is cut off.

I'm pretty sure that modern C++ compilers still put code into a VM to make it platform independent, if possible.

ree

tryin to get some practice in(still not texturing) so modeled a pig last night and animated it today what you guys think?

online jepeg to gif site cant seem to handle more than 30 fps so the timing is a little off

They don't, not by default anyway.

Nice start. Currently it looks like the head has no weight, since it's completely stationary during the animation. It just moves up and down along with the rest of the body.
Subtly delay the movements of the head. Make the head follow the body.

like dis?

honestly didn't touch it if i was going to use it in a game i would probably have a "look" animation or something attached to it/ i don't like having extra motion even thew it lacks weight many animals already compensate their head movements for their stride. and while no movement is "wrong" as you still bounce up and down, the opposite, weighted motion can also be wrong.

Looks better.
You're right when you say that animals compensate, but it's not perfect. The neck still moves a bit. Your old animation looked like the head was attached to the body with a steel rod.

yea, i didn't touch it added a new ik bone in the back of the skull and fixed it up a bit. but the whole compensating thing is the head moving in a strait line vs up and down with the body meaning the had looks more like pick 1 and setting the head before the shoulders gives the view of the head leading the body in its movement.

this looks really nice now

would you buy that for a dollar?

if it was under a freedom respecting license and i was making a game about hunting pigs

I wouldn't and I don't get who would
Unless you're planning on buying all of the models for your game
If not than it's just a waste of money and an opportunity to get better

What should I waste my time on, working on my game engine, or making a CPU architecture?

i'd say CPU architecture, we need some new open/libre hardware.

just like this guy?

wouldn't mind trading something for something i'm not as good at. swapping models and animations for textures/"materials?" or help their of. to getting good enough that i can just swap models for the rest of my game. would not mind making a ton of shit for people.

(Heil'd)
They're directional buttons, so they're going to be identical aside from what direction they move you. Why is this not acceptable? I can't think of any way to make it a for() or something unless I were to somehow make the button input evaluate to the distance that should be moved, which would be likely fuck other things down the line.
If you're referring to the distances being moved, there's no reason they should ever change, so I don't understand why I would need to make them variables.

Have a Move(float speed) function that can be changed.

Kowloon Sniper: Extra Chinky edition. It looks better in game but the webm washes out the colours.

But why? This is the only time and the only circumstance in which that code is run, and there's no reason it should ever be anything different than what's there. This even isn't a game where you have a speed, but a set distance that's traveled per button press.

i really like the webm look

What's going to be the solution to getting up to those high places in game? A really high jump? Jet-pack, grappling hook, becoming Spiderman?

...

Jumping on whatever is in sight as you can leap quite high as well as stairways, elevators, escalators, ladders, etc.

I by no means a pro at animation, but I just found this nice series of videos for animation tips; in the embedded video.
It may give you some direction on what you need to touch up to make it more lifelike (probably just arcs, slow in & slow out (easing in/out which can be done automatically in the animator), and anticipation).

Also, honorable mention for an awesome game related animation video:
youtube.com/watch?v=y2vCko6V4pQ

Neat.

crit?

He looks like he would crit, yes. Jokes aside, he looks like he has gigantic eyeballs underneath his helmet, for some reason.

crit?

i got pick related in paperback and its not really applicable to "game animation" because "gamer's" insist upon realistic or at least defined animations in their games. slow combat and god forbid you stretch shit. not to mention compensating for net code. while at the same time accepting shitty shit like feet sliding around when they turn. so the mythical good animations are just a fucking pipe dream.

and i forgot pick related

Fuck Hair

Yeah most of the 12 steps applies to things like pixel art games, or all apply to Animations for cartoons/etc.
Some principles can be used for 3D animation, as mentioned in my earlier post, but a lot doesn't translate to 3D animation.

I'm making a really simplistic TPS action RPG, what I found is that I can just tie increase / decrease movement speed not to the main pawn but instead set them on the enemies, tied to difficulty level. Same with gun fire rate. I don't have the luxury to speed up XP gain since I barely have enough trash mobs to slog through before the game ends though.

In summary this is all I have:

hp
boost
atk
def
critrate (higher critrate means after X shots you get critical strikes)
accuracy (higher accuracy means you go towards the higher values in the damage calculator)

So basically I calculate the damage of each bullet strike dependent on the atk / def of the player and the enemy as if it were a melee strike. But then again I have very few weapons so player has no need to discard since it feels as if they level up with him.

Welp, last one for the night will do any animal anyone wants me to except for humans or human-likes will not have texture thew.
first non sarcastic post wins

Do a gecko

it's pic

kk ill start tomaro

NEW BREAD

nice dub dubs
You could add a Precision stat that varies the damage done by a random amount after the Accuracy calculation has determines what damage would ideally have been had. So say your calculated damage from Accuracy was supposed to be 20, but you have a terribly low Precision (50%), so the actual damage dealt would be randomly chosen within the range 10-30. This means that with low precision there's an inconsistent chance at dealing higher damage but at the risk of likely dealing less; and at higher precision you are practically guaranteed to deal the median damage.

I'm purposefully trying to stay away from affecting damage through character stats, instead having that be completely derived from the weapon's ammo (and any modifications to that ammo) entirely. This way there's less of a focus on damage builds and a greater incentive to get creative in how you approach situations at higher levels.
Or perhaps I just overcomplicated everything, and I'm sorry.

Tangentially related, I plan on making enemy AI somewhat "fair" in that they too have character attributes that affect their skills and behaviors. The gamemodes aren't going to be one-man-army style, but more Player vs Enemy Squad, Sniper v Sniper, or other such things where its one versus a few (or intermittent waves of a few). With less enemy units used at a time, I can afford to make them more complex and personal. And it gives me an excuse to fuck around with various AI systems and approaches. And with less enemies, XP gained is vital, among other things.

having INT tied to xp gain is shit, itll railroad everybody into maxing int first then making their specialized build later

should be something like
CON = increased max hp, more resistance/immunity to certain effects such as pain, knockback, stuns etc
STR = carry more shit and use bigger guns, also affects recoil control (ie how fast the guns crosshair settles out again after firing in full auto)
DEX = move faster and have faster reload times, effect is much more prevalent on weapons that require a lot of hand movements to reload such as revolvers
INT = modify your weapons/armor with better attachments, better at manipulating people in conversations
PER = find more stuff while scavenging, higher damage from exploiting enemy weaknesses (ie if an enemy is vulnerable to fire and you have high PER you might do 3x damage as opposed to 2x damage for a low PER character)

Is it possible in blender to give each animation action or layers it's own frame length? I am asking this because I would later try rendering a Tiberian Dawn 1 resolution infantry and it has several animation types with different lengths.

new bread when?

because of say animations, effects or sounds
If you want a sound associated with moving it's nice to call it through a function so you can add a single call to make a sound to the function once
Also the reason you shouldn't hardcode for variables like that is partly because of moddability and also you say now there is no reason it should ever change, but what if you decide the character moves just a little to fast or to slow you are gonna change it and it will require to recompile every time you change not the case if you read it from a file of some sort. Which makes tinkering with variables a lot faster.
Ultimately it's not that big a deal but it's good to get in the habit of doing this kind of thing