Amateur Gamedev General ~ /agdg/ + /vm/

So he could take your work and not hire you? living the dream.

my work is currently shit anyways
but yeah, i don't feel like this is worth the effort
should i tell the guy i don't want the job or just remove all contact with him like some autist?

I don't know because there's too many unknowns about the job.

t. bossman
and now he gives me links to unity assets for kikebook and microtransaction integration
i just don't like it, i feel like a slut

What a retard, it's completely possible to do both, you just need to have some imagination.

Jesus Christ, user.

To whoever shilled the book "Getting Things Done" in one of the previous threads: Thank you.

Sounds like he's either nice, or desperate.
So I guess you're going with the software package since you said last thread you couldn't implement push notifications?

Also what's the difficulty with that? In general? Or with Unity?

I desperately want to go back to aping quake 3 movement code in my game but I have to draw furry porn to pay the bills.
lol

...

Is the name too similar? Will I get into trouble?
I get the impression that - is used similarly to : to denote a subtitle, hence it's franchise name would just be Imperator

...

But they've made a game

Make it a Caesar 2 clone but with dinosaurs. Call it Sic Semper Tyrannosaurus.

imperator is not a trademarked name if that's what you mean
It's from ancient history so yeah you can use it

Which engine do you think I should tinker with first, /agdg/?
I started off reading about tiled forward and deferred shading, then started consuming more papers and presentations on increasingly unrelated shit until few engines could satisfy my tech autism. I don't even need many normalfag-friendly visual features aside from GI, almost certainly using Radiance Hints algorithm, just something that runs reasonably well on GL3/4+ hardware.
The main options I'm considering are Crown and Panda3D: both of which are easily extendable and have their advantages. Crown satisfies much of my tech autism but uses bgfx instead of OpenGL, potentially interfering with certain optimizations, while Panda3D has a lot of shit already implemented but needs a GLSL shader generator to replace the Cg one. Going full enginedev isn't an option right now with my inexperience, although someday it would be nice to try making something similar to ttimo.typepad.com/blog/2013/05/es_core-an-experimental-framework-for-low-latency-high-fps-multiplayer-games.html .
Just going with Unreal 4 or Unity would not be an option, since those have too much overhead for what I'd like to make without gutting large portions of the engine.

Why is OpenGL such a mess ?
Seriously I'm trying to understand what library I should use but it's a fucking mess. Glut, Glew, what the fuck is this shit ? How do I even get OpenGL ?

Will you put in A BOMB?

OpenGL is a graphics library written in C. There are a number of wrappers available for almost any language including C++ and C#, among others. You download one of these .dlls and invoke the code normally.

Most people use a framework like SFML or perhaps SDL which builds on top of OpenGL and provides additional game related functionality such as reading hardware inputs, audio, and networking. I believe these also have bindings for many languages as well

OpenGL is implemented on a driver level. You don't "get" OpenGL.

MoM user here, I wondering if you guys could help me with a little somthing; I am trying to do research into copyright code so that some one like Bethesda can't pull a Starcraft 2 on me if I ever release my mod.

doom.wikia.com/wiki/Doom_Source_License

I am slightly confused since it seems to use a duel license, but it seems like I should be fine as it falls under the GPL code?

Glut/freeglut is an unsupported legacy framework: generally you don't want to start a project with it and there's much better alternatives like SDL and GLFW. Supposedly libepoxy is better than glew but I haven't tried it yet.

Oh yes, this is specific line of the copyright agreement that worries me

(including all characters and other images generated by the Software
or depicted in the Software) are owned by ID and is protected by
United States copyright laws and international treaty provisions.

It makes it sounds like anything that is featured in iD's engine belongs to them; does this include modded content?

You're being taken for a ride.

If you can't even figure out how to "get OpenGL" then it's probably not for you.

I remember trying OpenGL years ago in highschool there were only a handful of intuitive tutorials out there and they tended to be written in a poorly structured manner, crutches and all. The fact that he's even mentioning glut likely means he's stumbled onto this lion's den of convolution and is understandably confused.

Of course one could always try to grok a feel for OpenGL's overall design by reading the API reference but that is an impossible endeavor for someone who isn't used to reading documentation and hasn't yet cultivated the intuition needed to escape the rote formalism presented to a programming novice.

opengl.org/wiki/Getting_Started
tomdalling.com/blog/modern-opengl/01-getting-started-in-xcode-and-visual-cpp/
Ignore GLUT entirely. If you want to get started I'd recommend cloning the following repositories:[ GLFW, GLEW, and GLM (OpenGL Mathematics) ] from git and then going from there. These are all libraries that abstract a lot of the boilerplate needed to build an OpenGL application from scratch. I don't use GLEW personally but most tutorials use it so I'd recommend starting out with it. Pick another library that suites your needs once you understand what GLEW is/does.

obviously call your relevant system API to get an openGL context. like win32.

go read the first tutorial on nehe.gamedev.net

Enjoy your nonportable fixed function pipeline shit, faggot. Learn GL3/4 from a decent tutorial site and then come back for GL2 if you want to support Direct3D 9-level GPUs at the expense of flexibility and neat shit like instancing or geometry shaders.
ogldev.atspace.co.uk/
open.gl/
opengl-tutorial.org/
learnopengl.com/

I never got into shaders or whatever else modern openGL has. Actually I never got past textured polygons in 3D. You're probably right though.

How about no? He should just download a precompiled thing that works out of the box, so he doesn't have to fuck with versioning, compiling it, setting up an environment and all that other fucking retarded stuff. You said yourself that docs are "too formal" for him, yet you expect him to compile a dll just to hellotriangle?

I'm tired of being a procrastinating ideaguy. What I want to make an srpg. My idea basically started out as a bare-bones Fire Emblem clone before it became its own thing. What engine/language would be best for something like this?

Literally every statement in this post is false.

Correct.

Too low-level and too much information.
Too brief and dense, introduces too many things in too few paragraphs.

Stop recomending this fucking site. Are you actively trying to sabotage other user's attempts at learing shit?

All of these tutorials are excellent and this user knows his shit.

I'll give you a short rundown.
Forget you ever head about this. Use exactly one of the following:
They all give you the same thing: A window to draw on and user input. Some, like SFML has a bunch of other stuff, but it's mostly irrelevant.
A specification. OpenGL is not software, it's a pdf that say what an implementation should contain. Your graphics driver is (or rather has) such an implementation. You use OpenGL by doing #include (or some stupid wangblow equivalent), calling the functions it defines, then you link against the shared library your driver provides, which I hear is usually at c:\system32\opengl.dll. If you don't know what including or linking means, learn C/C++. If you don't want to, find a binding for a language you can stomach. I hear libGDX is used for java and I think there's at least two for python, though if you do venture this way few people can help you. I can only vaguely remember an user talking about libGDX and I don't think anyone else uses bindings for other languages.

Do you have any proof for those wild allegations?

afaik OpenGL itself is a standard for relating APIs, and was maintained by a review board before 2006, but is now controlled via the Khronos work group for relating APIs.

So that's why u see stuff like the red/orange book, and why it's so well documented; is because that's the purpose of OpenGL is to maintain standards, implement new things into the API, and such.
Which is then used by various graphics/rendering libraries (actual graphics libraries containing the implemented standard of the OpenGL API); i.e. what commands must be implemented, pipeline structure, etc.

Correct me if I'm wrong here, but I think that's right.

Anyone here understand the math behind taking penetration, location of impact, and a normal and calculating where the player would move from there?

Tell him you think it would be a good idea to do a port of E.T. from the Atari, it would definitely be a commercial success.

buhaha, nice

There isn't anything wrong with that website. It explains how to use openGL 1 very well.

Do you think that learning openGL 1 will hurt someone rather than help them? Is there something wrong with the information on that site?

Right but you're saying that everything was wrong.

People DO use SDL and SFML for devving and there ARE tons of wrappers for both them and OpenGL

You should look at a modern tutorial because its likely that certain optimizations, paradigms, and requirements have changed over time. You dont want to get locked into one pathway of thinking of its completely wrong and deprecated

The APIs completely changed from 2 to 3.

Those tutorials work fine on windows with no problems. Just about every graphics card is compatible with legacy openGL. The modern versions are different, but the older one can still be used without any issues.

Learning how something works is different from being locked into a pathway of thinking. Anyone who uses legacy openGL knows the limitations it has. There isn't much more you can do than textured polygons, alpha blending, and a z-buffer.

>Right but you're saying
nah user look at my ID and posts thus far.
I didn't say anything along those lines, heh, u just new or lacking reading comprehension?

Yes. If someone wants to learn how to drive a car you teach them how to drive a regular car, not a fucking roundabout. OpenGL 1 has a completely different design philosophy than 3+ and the fixed-function pipeline is considerably dumbed-down compared to the programmable pipeline, offering much less room for optimization and interesting rendering techniques on hardware produced in the last decade.
Don't steer a newcomer into learning a graphics API from the 90s without basic shit like shaders just because you're too autistic to learn an 10 year old API instead of a 25 year old API, even if you're free to do that with your own engine.

To anyone who isn't familiar with the differences between OpenGL's fixed function pipeline and the programmable pipeline, see opengl.org/wiki/Fixed_Function_Pipeline and opengl.org/wiki/Legacy_OpenGL .

...

I think its easier for a beginner to learn how to use a fixed function pipeline. Its true that its much simpler and has less functionality, but people will also use a special middle-man library to set up their openGL context, because its easier for a beginner, even though its simpler and has less functionality than regular winAPI.

What I mean is, a beginner is not going to need to know about shaders- they just need some crap on the screen. Its not very diffrent from something like SFML dumbing down modern openGL for people.

What that means is that all characters and images are owned by iD. You can't put in a Doomed Space Marine.
The actual engine itself was made open-source long ago, and Graf Zahl and Randy declared their modifications to the engine to be GPL-compliant as well.
As long as you include all of the proper license information, you'll be fine.

The only questionable factor is Ken Silverman, but if you're not selling then it shouldn't be an issue. And if you ARE selling, then just use GZDGPL.

I was working on some voxel shit in Unreal and I came to the realization that it wasn't the right package to use. I'm going to switch to Unity now. I haven't dev'ed in like a week or two at least, been dreading starting over again.
I just need to bite the bullet and get in there and start coding again.

Wish me luck anons.

>Want to make a VN
>On RPGMaker
all I need is to tripfag around and I'm the perfect gamedev :^)

So, I've been making this.

Made my own thread so it wouldn't take up space here.

Horseshit. You would have to completely ignore every prominent piece of OpenGL literature written in the last 8-11 years to not know the huge role shaders play in modern OpenGL. Even the most dumbed-down wrappers around OpenGL will have the user interacting with shaders to some degree if they wish to do anything beyond preset behaviours.
Also, stop pushing the winAPI so hard when it only makes any future cross-platform support more difficult. Everything you've said in this thread has been shooting yourself in the foot and you belong in pic related right now for trying to force your unironically terrible advice on a newfag.

If you're using RPG Maker 2k or '03 you may want to look into easyrpg.org/

VNs would be perfect for RPG Maker, sarcasm or no. Very simple, not terribly demanding, and you could even add a small spot of actual gameplay.

good luck user

What type of voxel algorithm you going for?
I may be able to give u some recommendations for tutorials/resources relating to it, or a pointer in the right direction for Unity related stuff.

Actually, during the day you will be able to perform certain actions, explore areas, interact with people and solve some simple puzzles that will reward you with information and items. There will be lesser events here and there and some huge events you have to deal with.

A bit like Long Live the Queen meets Tsukihime with some added puzzles and plenty of customizable main character's clothes
Any suggestion would be welcome.

Doesn't seem that appealing since I'm using VX. It already does all I need to do even if I can't code.

good enough for now, I will probably make an updated version in a bit when my skills improve.

A Bomb?

Surely there's better vn software than RPG maker?

I recommend you make a VN in Idtech 1

I know what shaders do, I just don't think you need them for anything simple. I don't read OpenGL literature because I don't spend a lot of time writing OpenGL code. Sure, if you want a extra rendering features, then yes, you should use them, but if you just want texture-mapped polygons on the screen then you don't need them.

I wasn't saying that you should use winAPI- I was just giving it as an example of how you can use something with less features instead of something with more features.

I am not forcing anyone to take my advice- its just a suggestion, like everyone else who responded to his request for advice. I don't know why it looks that way from your point of view. I'm sure if the user who asked wants something with more features than openGL 1, they will use somebody else's suggestions.

Better use C v1.0 because it's more elegant and bloat-free than future versions!

Holy dick shits you can make a VN in Doom , do it scrub, do it now, or I will and release it on Steam and make like 70 cents on the dollar.

techraptor.net/content/you-can-make-a-doom-visual-novel-with-this-mod

top kek the tablet looks like it says cuck.

I said "can", not should… If I didn't use the features in C99, then I wouldn't be using C99. The same goes for anything else.

Fuck. It would actually be great for a detective VN.
if I finish this project I will try this

See ogldev.atspace.co.uk/faq.html

...

Everyone's a female when it comes to Valve and their 30 percent cut

False: we both bitch but women don't always take it up the ass anyways. It's not like we have much of a choice or the option of alimony if our asses get shredded.

That's not my point fam.

Like women we get 70% when the deal is finished?

In the newer version which uses shaders for about everything yes- but not in the older version.

Apparently Godot only detects collision for the thing that initiates contact. Am I right or no? If I am, then how do I get around this? If there is no way to do get around it, then how can I do what this guy did, but with a kinematic body instead of a rigid body?

youtu.be/GncaMNv7Fso?t=22m20s

Here he makes it to where when the ball hits the bricks, the bricks disappear. I want to do this, only the ball is a rigidbody and I can't figure out how to get that code to work for kinematic body. The bricks won't delete themselves if they're hit by the rigidbody because they are not the ones hitting the other object, or atleast that's what seems to be the problem.

actually I didn't have a point, I was just making a bad joke.

Not sure what algorithm I'm going to use, still considering my options. Any links to tutorials or voxel specific stuff for Unity helps and is appreciated.

What? You don't have to version shit. Git clone to grab the stuff you're already going to include anyways. It's not any different than getting a tarball if you stick to that basic command but you have the option of checking out the latest master later in the game when you actually learn git.

If either of those are too difficult then you really need to focus on getting the basics of programming down before diving into building a game engine. If you can't handle basic makefiles or google to learn git then you aren't going to get very far.

If you're using a C/C++ interfaces for OpenGL then you're already knee deep in shit so you may as well start wading.

As soon as AGDG strays outside of GM/Unity/UE4 or 2D, you faggots go full retard, every single time.

Fuck off

for
obviously

may of found a bug in unity… using functions that aren't used much so makes sense.
Editing the submeshes of a mesh in a vert shader using separate materials/shaders per, using the graphics.drawMesh function in an empty scene.

kek, gets me every time

U need help deciding with one? I've coded quite a few different approaches.
As the main reason being that it'll be a lot easier to help u w/specific info I have compiled as there's various styles to it… and lacking specificity in this wide topic results in just too much info to splurge haha.
What u thinking of atm, or do u not really know the options?

I made a dank meme to commemorate this thread's theme.

also the aspect ratio is terrible bcs i cropped out my discombobulated scene hierarchy

Thanks for all the replies.
To those of you wondering if I'm completely lost, no I'm not I just did not understand everything about all the libraries coming with OpenGL, also I'm not on windows I'm on Linux and since my laptop is old (Thinkpad T410) I'm stuck with OpenGL 2.1 and some people say it's bad (too old/not portable). Anyway, my goal was to make a 3d engine that could run on the worst toaster ever and the best machine as well for my dungeon crawler project. I wanted to make a 3d grid and randomly generate caves & dungeons with algorithms who would put tiles where they're supposed to go.

I'm starting over, so I just need simple tutorials for now. I'm new to C# (but I learned C a long time ago) and have coded other languages as well. Just something simple to get some cubes on the screen would be nice.

If you can only develop for 2.1 and lower, and want it to work on really old cards, then you cant use the shaders anyway, so just go for legacy openGL.
nehe's site will be fine for that. It really wont hold you back for any of the things you described. Pretty much all cards can run legacy openGL.

The unity website has good official tuts, and especially so if you already have programming experience.
Anything specific that isn't too low level, just search for a vid, and you'll more than likely find a detailed tut somewhere.

Also, docs/general resources to use (bookmark all of these):
wiki.unity3d.com/index.php/
docs.unity3d.com/540/Documentation/ScriptReference/index.html
docs.unity3d.com/540/Documentation/Manual/index.html
answers.unity3d.com/index.html

official tuts:
unity3d.com/learn/tutorials

If you want more practice, and of higher quality than your average YT tut look at their live demos. They're pretty comprehensive, and the general tut guys go into a lot of detail, tips and tricks, and lower-level stuff if it's pertinent.
Or you can use a linda, coursera, or any of those other online course providers.

to be clear, I mean the unity's official live demos

also in reference to unity staff doing unity's live demos…

sage for clarification

give me an idea for a game that's fun, replayable, simple enough for me to not give up on it due to a large scope, and also profitable

So anyone could press play and go right ?
I also checked other 3d libraries and one that stood out was Irrlicht, what do you think of it ?

how about we give you a million dollars while where at it

A game where you control gravity.

it''s been done too recently

I'm going to make an MMORPG

A game where you control magnetism.

A game where you play as a person playing the player, who is a game.

What a waste of double dubs.

Epitome of fagatron, right there.

i have ideas, just not the kind that i'm capable of finishing

Like losing your virginity lololololol

TBH try end boil your ideas down to one or two mechanics and make a game around them.

alright, i'm gonna make a rock-paper-scissor type rts

If you wanna be toaster-compatible, 2.1 with vbos+ibos and shaders is good. My toaster is even older than yours and I can do vid related at 60fps (recorded @24) and I'm not even culling verts between chunks, nor ordering chunks in front to back order, both of which are on the list. If you want your shit to run well on toasters, shaders are the way to go. Unless you run into a software-emulated feature you'll have much more control and will be able to optimize much more. Heck, just the fact that all your data is resident on the GPU instead of re-sending it every frame will give you a huge performance buff. See, that's the thing with the fixed-function pipeline. It works by you calling glVertex3f() every single frame to manually send all your vertices to the driver, one by bloody one. The slightly more modern way (which 2.1 supports) is to keep all your vertex data in a buffer on the GPU and calling glDrawArrays/Elements() every frame, then calling glBuffer(Sub)Data() when you need to update the data (which in the case of static map geometry is never, or rather on level load). What a shader does is instead of relying on the fixed fuctionality (and the goddamn matrix stack), you tell opengl what you want your vertices to look like. Then you write a vertex shader that transforms the vertices into device coordinates and a fragment shader that computes the color of the pixels. This is far easier than it sounds. Look at opengl-tutorial.org/ and you'll see.

Irrlicht is an engine. There you've moved away from grapics api territory and into engine territory. The two are different beasts entirely. You use opengl to write an engine, while Irrlicht is an engine. You can probably interact somewhat with opengl though it, say by writing custom shaders, defining custom vertex formats, etc. Also, I see you keep using the word "graphics library". Let me explain again. OpenGL is not a library, it is an abstract concept, a specification, a set of rules and regulations on what graphics drivers should expose to programmers to allow them to perform hardware-accelerated raster graphis, i.e. turning coordinates into pixels. It is not itself a library. You already have an implementation, i.e. a library if you have a working graphics driver installed. Irrlicht, like I said is an engine. Amongst other things it has a 3d renderer, which probably uses opengl, but could be using other apis as well, like DirectX Direct3D. As an engine, it provides a bunch of other stuff as well, like input, a scene graph, file io, asset importing, physics, etc. It's a much bigger thing, while OpenGL is just low-level graphics.

pick one

well i asked for ideas, didn't i?

Yeah you asked for the perfect game idea which everyone wants and nobody has
It's like asking for a miracle, it's not gonna happen

alright, forget the money part, just give me an idea for a simple/fun/replayable game

Why the fuck would anyone give you an idea they think is actually possible to complete by some random autist on the internet AND is able to make money?

If you can't come up something to make, pick a classic arcade game and put a spin on it. At least you're doing something then, and it gives you time to improve yours skills and come up with a "proper" concept.

Vid related, simple as shit yet satisfying to create.

RPS RTS is probably feasible, though they have a reputation for being a bitch to balance (still less than RPGs).

2.1, and anything pre-3.0 is considered bad because its pre-standardization/cleanup.

But really, it's fine. As says, vbos+ibos and shaders is all you need for the most part.

The main feature you're missing out is 3.0's vertex array objects. These objects will contain the vertex array bindings you've made. (That is to say, every glVertexAttribPointer call you've made.)

2.1, although ancient, will remain compatible. Do you really think Nvidia or AMD want to ship a GPU incapable of running older OpenGL2.1-reliant games? That would be some pretty terrible marketing.

alright, just one thing to ask
if i'm doing this in unity, do i have to make my own pathfinding algorithm or can i just use navmesh? i'm asking since i'd imagine that navmesh might be too much when doing hundreds of units, and baking maps could be a pain if they're too big

The problem you'll run into is that making things run great on toasters is a lot of fucking work, because old toasters differ from "modern" (very loosely used here) ones. It's something that the Blender developers had to deal with. Blender used to be backwards compatible to OpenGL 1.5 up until a version or two ago. They recently dropped support and require 2.1 now. The reason for that is that OpenGL 1.5 is so old that some features aren't implemented in modern graphics hardware anymore. As a result using OpenGL 1.5 functionality will still work, but rather than being calculated on the graphics card, those functions will be emulated on your CPU, undermining the entire point of using OpenGL and absolutely killing performance. Please bear in mind that "modern graphics cards" also includes the absolutely low-end cards from a couple of years ago. On these you will run into trouble. The only solution to this dilemma is to have two renderers – one for 1.5 and one for 2.0 – which means twice the work.
What you will need to do is decide what exactly "toasters" are. Do you mean very old hardware that only a few people still run, or do you mean very weak hardware that may or may not be recent? If you mean the former, OpenGL 1.5 will be supported on more target hardware. If you mean the latter, OpenGL 2.1 will be the better choice.
Usually I'd strongly advice you against targeting something as outdated as 2.1, but since your development machine is restricted to it, it can't be helped. I don't see the point in learning something that is already outdated. 3.3 would be a much nicer version to learn.

RTS games are not easy to make
I guess it depends on how far you wanna go with it but to me it sounds like you probably wanna do something like suggests
If you really make a RTS you better be able to contain your autism or the project can become quite complex and time consuming
Just consider things like how many factions, how many units/buildings per faction.
What kind of abilities do you want units to have.
What kind of economy system will the game have.
If you try to make supreme commander you'll spend a decade on it as a one man team

tbh, i'll just simplify it by making it mount and blade as an rts
units are all the same, with the one thing making them different is their stats and items
for everything else i think i'll just clone age of empires 2

Yes I'm stuck on the 2.1, my laptop's graphic card isn't compatible for anything higher than that. As for my definition of toaster, it's anyone who buys old laptops from 2007/2010 and runs Linux on it like me and my Thinkpad T410.

It's true but I don't want to be caught with my pants down after spending a few months coding.

Thanks for explaining everything.
In the beginning I wanted to do a raycasting engine like wolfenstein but render multiple levels and use sprites like pic related but ended up deciding against it because of the sheer amounts of 3d animations I would have to do for different weapons and armors (I want to make a more simpler Kingdom Come Deliverance type of combat system).

I planned to record footage of my main project this past weekend, but my screen-recorder is being a shit.

Decided to start up a side-project just because it helps me keep a work ethic. Began experimenting with random world generation. Right now just meaningless rooms and paths, but hopefully soon they'll actually have some meatiness to them.

Pics related show generated area, one just started and the other partially explored. The red dot is the room the player is in. More rooms are generated by player proximity.

The bigger problem with making an RTS is AI. Unless it plays in a very unconventional way, AI is going to be an absolute bitch to code in.

bump

I meant 2d sorry

Has anyone ever tried to rewrite sdt in something that isn't flash? Performance is atrocious.

Someone really should. If you add furry stuff you could get some of those patreon shekels.

Progress with sound

Don't furries already have their own 3d game that is lightyears ahead of sdt?

Tempted to. Doing it in something like löve would be fun, so it could be completely modded or extended.

furries have everything because they all seem to have a gorillion dollars to throw at their shitty porn

its probably because they are actually willing to pay for their porn or at least enough of them.
from what i hear you can also live of making furry art if your halfdecent and couldnt live of other art stuff.

what am i doing wrong
i'm trying to get units to move in a formation, in this case a square
first i tell the first line of troops where to move, and then i tell each column of troops behind them to follow each unit in front of them
but the ones that are following never seem to follow at the same distance, as you can see in the video
the navmesh stopping distance is set to 0.5
public void moveTo(Vector3 pos) { UnitSelection.Instance.StartCoroutine(Move(pos)); } IEnumerator Move(Vector3 pos) { while (Vector3.Distance(units[0].pos(), pos) > 1.5f) { Vector3 mainDir = (pos - units[0].pos()).normalized; Vector3 targetDir = Vector3.Cross(mainDir, Vector3.up);//to the right of the main dir switch (formation) { case Formation.square: if (units.Count >= 4) { int rowNumber = (int)Mathf.Sqrt(units.Count); for (int i = 0; i < rowNumber; i++) { units[i].SetDestination(pos + targetDir * i); units[i].moving = true; for (int j = 1; j < rowNumber; j++) { units[i].moving = true; int p = rowNumber * j + i, q = rowNumber * (j - 1) + i; units[p].SetDestination(units[q].pos() - mainDir.normalized); } } } else { for (int i = 0; i < units.Count; i++) { units[i].SetDestination(pos + targetDir * i); units[i].moving = true; } } break; } yield return new WaitForSeconds(0.5f); } for (int i = 0; i < units.Count; i++) { units[i].moving = false; } }

Aren't you guys worried that someone might pass by one of these threads and steal all your original content?

that just means that they'd have to suffer as i do

Don't user their official tutorials. Many are out of date to the point where even the corrections are out of date and no one's bothered correcting the corrections yet.

alright, i improved it
public void moveTo(Vector3 pos) { for (int i = 0; i < units.Count; i++) { if (units[i].squad != null && units[i].squad != this) units[i].squad.units.Remove(units[i]); if (units[i].squad == this) { UnitSelection.Instance.StopCoroutine(movement); } units[i].squad = this; } movement = UnitSelection.Instance.StartCoroutine(Move(pos)); } IEnumerator Move(Vector3 pos) { Vector3 mainDir = (pos - units[0].pos()).normalized;//main direction Vector3 targetDir = Vector3.Cross(mainDir, Vector3.up);//to the right of the main dir switch (formation) { case Formation.square: bool canbesquare = units.Count >= 4; bool finished = true; int rowNumber = (int)Mathf.Sqrt(units.Count); do { if (canbesquare == true) { units[0].SetDestination(pos); units[0].moving = true; for (int j = 1; j < rowNumber; j++) { units[j].moving = true; units[j].SetDestination(units[j - 1].pos() + targetDir); } for (int i = 1; i < rowNumber; i++) { units[rowNumber * i].SetDestination(units[rowNumber * (i - 1)].pos() - mainDir); for (int j = 1; j < rowNumber; j++) { int p = rowNumber * i + j; units[p].moving = true; units[p].SetDestination(units[p - 1].pos() + targetDir); } } } else { units[0].SetDestination(pos); units[0].moving = true; for (int j = 1; j < units.Count; j++) { units[j].moving = true; units[j].SetDestination(units[j - 1].pos() + targetDir); } } finished = true; for (int i = 0; i < units.Count; i++) { float dist = Vector3.Distance(units[i].pos(), units[i].destination); if (dist > 0.5f) finished = false; if (i != 0) { units[i].agent.speed =Mathf.Clamp( 3.5f * dist / 0.5f,3.5f,8); units[i].agent.acceleration = Mathf.Clamp(8 * dist / 0.5f, 8, 16); } } yield return new WaitForSeconds(0.25f); } while (finished == false); break; } Debug.LogError("finished"); for (int i = 0; i < units.Count; i++) { units[i].moving = false; } if (Instance.selectedSquad != this) { Destroy(this); } }
i tell the first unit to go to the destination, everyone except him in the first column to follow whoever's infront, and everyone in the rows except first column, i tell to follow their neighbor's side
the issue now is that every next column is slightly behind it's neighbor. i tried directly telling them their end destination, but then they don't move as a formation
and because of that i gave them a speed boost depending on how far they are from their target (except the leader). sort of fixes it, but they still don't move like a square
and my other issue is that this thing doesn't work if the number of troops doesn't have a round root
so with 4 or 9 troops it works, but with 6 it will give me an error, since there's nobody in the third row
wat do

nevermind, i'm retarded and figured it out
i just had to do it in one loop rather than two

bool canbesquare = units.Count >= 4; bool finished = true; int rowNumber = (int)Mathf.Sqrt(units.Count); do { if (canbesquare == true) { units[0].SetDestination(pos); units[0].moving = true; for (int i = 1; i < units.Count; i++) { if (i % rowNumber == 0) { units[i].SetDestination(units[i - rowNumber].pos() - mainDir); } else { units[i].SetDestination(units[i - 1].pos() + targetDir); } units[i].moving = true; } } else { units[0].SetDestination(pos); units[0].moving = true; for (int j = 1; j < units.Count; j++) { units[j].moving = true; units[j].SetDestination(units[j - 1].pos() + targetDir); } } finished = true; for (int i = 0; i < units.Count; i++) { float dist = Vector3.Distance(units[i].pos(), units[i].destination); if (dist > 0.5f) finished = false; if (i != 0) { units[i].agent.speed = Mathf.Clamp(3.5f * dist / 0.5f, 3.5f, 8); units[i].agent.acceleration = Mathf.Clamp(8 * dist / 0.5f, 8, 16); } } yield return new WaitForSeconds(0.25f); } while (finished == false);

Doing Super Deepthroat in Love is inherently a funny concept, if only for the name.

THATS FINE I DIDNT WANT TO BE PRODUCTIVE ON MY DAYS OFF ANYWAY

So anyways, in a city builder, how do you properly manage the grid of values and shit? Like say building A and B in pic related both set the land value to 4, with the effect tapering off. You place A first, everything in your LandValue[] is updated appropriately, and then when happens when you place B?

Should it compare values between A and B to set the higher one? What if there are other sources of land value?
Does it just look at the land values without context and update them correctly, or should it take other buildings/sources into account?
What happens when I remove the building? Should I look at ALL sources of land value modifiers and reapply them? I get the feeling that since you don't necessarily know where land value is being sourced from when given a particular tile, all the building in the city might need to be looked at.

Because of special cases where things might limit land value or take the highest value amount a type of source, rather than adding it, simply subtracting the values on removal might not work.

does it look like an rts yet
so far i've got square/2 row formations, and you've got regular/shift/control selection (selects new squad/adds to current squad/removes from current squad)
so next i'm gonna make some basic structure for all the different buttons that are supposed to appear depending on what you've selected

Switch to the Tomorrow theme?

Looking good user.

wow, unity finally added actual features

not bad
how did you do the formations?

i explained it in the previous post
first unit is told to go to the end destination
everyone in first column follows whoever's in front, until they reach the first unit
and then every unit in each row, except the ones from first column, follow whoever is on their side
for the square formation, i just set each row to have the root of the number of troops in the squad
for a formation with 2 rows, i set the number of troops in a row to be half of the troops
and then after the first unit i just cycle through each unit, checking if he's in the first column or not and telling him where to go depending on that
once each unit has gotten 0.5 or less meters close to their target, the movement has finished
aside from this, i'm thinking of adding pyramid/curved/spread out formations
and i still need to change how the units rotate, the way the navmesh does it kind of fucks some of them up when they finish a movement

so what happens when you path formations through narrow spaces that they can't fit through?

simple, we make sure to never put walls in the game
but in reality i guess they'll just try to rush through
because of the way i made it, they will always make sure to stay behind the first unit, so he's definitely gonna manage to get through without a problem
no clue what will happen to the rest though

they manage to get through, although a bit slower
also, what's the ideal camera angle?

DORITOS

The one that changes when you scroll up and down. More vertical when zoomed out further.

If you have a minZoom and a maxZoom, you can normalize the current zoom to be a value between 0..1. From there, you can transform it into a sine value, resulting in a smooth transition rather that a linear one.

I'm working on a 1:1 copy of a game I adored in the past mainly because it was written in shockwave flash and noone really supports that anymore

And I finally got enough working to post something!
Fonts are still wonky as hell, but that's not my priority right now.
Does anyone know how to properly structure my information files? I want to make this as modular as possible – to allow anyone easy level/unit/story editing.
The second and third image are my current set-ups for units and levels; the first one is a total mess but I feel like I'm getting somewhere with the second one. I'd love to have some resources on that.

wew

That should almost be a feature. An RTS where the soliders have about the same level of intelligence as the average cod kid, getting stuck in doorways while everyone is trying to go through first.

bump again

Haven't gotten much done for space pirate on the code front, mostly have been writing and working on misc design front–I've been plotting out levels, specifically.

So have one of the artists' drawings of Silver, since he seemed to be fairly popular.
She's working on providing clean drawings to use as sprites and rotations for him.

wew

This is what I have been doing tonight. A level that takes place on rocks floating in a void.

Not sure where im going with this, but its diffrent from the rest of my levels so I like it.

https:[email protected]/* */[email protected]/* */ch-except-all-the-speech-is-from-81e927a417e5#.bi5s40xcu

FUCKING MADMAN

I forgot how to link across boards [spoiler]no bully pls[/spoiler
8ch.net/agdg/res/22087.html

you making a decker clone?

Good luck.

That guy hung out on agdg? Huh.

If you had cared to read the thread, you'd have seen that that's not the dev of that game but someone using their assets without permission.
>>>/agdg/27597 (that's how you crosslink, btw)

i got camera movement to work when you hold down the scroll button or when the mouse is within 10 pixels of the border of the screen
disabled collision between units, since then they take forever to restructure when moving
added a button system. the buttons show up when you select a squad, and you get buttons depending on what the units in the squad have in their settings
in this case we only have one troop with two options for a square/rectangle formation
button position depends on it's key binding (QWER/ASDF/ZXCV)

I can't believe weapons are giving me more trouble than dresses, on the other hand I think I am finding better solutions to dresses.

added unit windows
if you click any of them, it selects a squad with only that unit
if you shift click, it excludes that unit from the current squad
new units continue their last movement, however, since they're technically in a new squad, and each of them is just a single unit, they end up going to the same destination rather than thinking about where they should be in a formation, so you end up with all the excluded units bunching up in the target destination of their old squad
this is probably why i should have just told all the units their end destination rather than who to stand next to

nope, it's the lego game "spybot: the nightfall incident"

...

whats a good starting tut then? It's hard to get into it because it changes so much so fast that everything seems out of date instantly.

they're fine

That user just doesn't understand the point of docs, and actually looking at the warnings when they say "deprecated function" (which tells u the updated function anyways).
The only thing you need to avoid is anything relating to OnGUI, as that's the old UI, and the new one is way better in terms of performance and usability (called canvas/UI).

You will run into deprecated functions, but as mentioned above, docs are nice, and look at the warnings.

You need to learn OnGUI() if you want to extend the editor, and believe me, you do want to extend the editor.

sadly this
in my case i used it to make a shitty node-based dialogue editor, like unreal's blueprints
although for simple stuff like the inspector you could just use GUILayout, which is a lot simpler than GUI

yes, but doing it for a few editor tools vs doing it for an entire game's UI is another beast; hence me warning said user for using the preferable canvas/UI for an in-game UI.
The whole onsceneGUI is useful for custom editors, but isn't exactly something you go diving into when you first hopped into the engine; it's something you use when you need it as there's not another option.

I'll be in a plane for 12 hours soon.
Which books should I read?

Masters of Doom is great if you haven't read it already.

Does it have any insights that could help with coding in general?

i put a placeholder model for the unit, from the store
also added the option to change the spacing between units in the editor
i'm also saving each unit in a squad's final destination, so that if you exclude troops via shift clicking their icons, or selecting them as a group with Control, they definitely won't all group up in 1 spot, but will go to their original formation destination.
i'll probably do healthbars or something next
also any clue why shit isn't dynamically batching? all my units use the same material, including their weapons, but i'm still getting around 80 draw calls

harry potter and the methods of rationality, it's very autistic, for autists like myself

Not really, mostly motivational. If you want to do some coding exercises bring lots of papers and draw yourself some architecture, design interfaces and such. Game coding complete, Game engine architecture and Practical rendering and computation with direct3d 11 are all good. Last one isn't just D3D11, it teaches a lot of general rendering concepts. If you're an autist you can read the OpenGL superbible, Lua reference or C++ Bible.

I don't think I can bring pen or pencils to the plane, though

Why not? Is the TSA that retarded?

A good exercise anyway is to design either a component entity system, scripting system, event system, graphics interface or similar.

Nice try Mohammad

I can write up the logic for a UI's simply elements (buttons, cursors, etc) no problem. Rendering it is the biggest pain. What about different styles and such?

That explains this then.

...

Does sharpening a pencil from above to below allowed length count as destroying evidence?

Is one of the prefabrication stages of pencil factories cornered off and under much stricter regulation due to the manufacture of dangerous arms?

Will they start banning pencils below a certain hardness as dangerously pointy and start forcing 9Bs on everyone without a pencil license?

alright, now all the units rotate the same as their leader
also if the units are close enough to their final destination, they'll just go to it rather than spinning in circles because the formation is rotating
so at close distances they don't spend forever going back in forth, while at long distances they still move in a formation

are units just clipping through eachother or does it just look like that?

I'm looking for a text editing tool, and I'm not sure if such a thing exists or what it would be called.

I desire a text editor with hierarchical/outline or tree based organization, that has fully programmable output that allows me to process and spit it out in whatever arbitrary formatting and glue I choose. If you've ever used tile studio and its code generator, that's the kind and type of output control I'm looking for. And I'm looking for something not specifically tied to one engine or language.

KISS

Don't give them ideas, I beg you.

Because every game is going to have the exact same icons for buttons

so fellow programmers, did you finish the game menus, while all the artists do the real work?

i intentionally did that, otherwise it takes even longer for them to get in formation

math

Fuck my life into pieces.

Hey

Hey Term

Update Demonsteele

...

...

Here comes the ugly part, finding a point on the surface.

And there they are, the patch mesh verts.

I think I'm starting to get the hang of sculpting models, but does anyone know any good tutorials for modeling hair in blender?

They look like lumpy potatoes

pls help

Oh yes

make a half sphere, model it to look like hair, then duplicate and scale it 2-3 times
then just use a texture with transparent areas for each layer of hair
i'm sure the witcher 3 did it like that

...

I reckon they would, they aren't rendered, cleaned up, or finished. Though they may just not be that good. I'm only about two weeks in and I'm still kind of learning the ropes.

gotcha, I'll give it a shot.
Thanks

Yeah wash and cook those potatoes you fuck.

You don't like curves?

I have trouble even tesselating my own mesh, purely programmatically

well, if you can articulate why they are bad I'd love to hear it. Pointing out why something is shit can only help me improve faster.

what a shame

I'm still awaiting a playable demo of that.

I'm clearly shitposting user.

Nice user, that's looking really nifty. It's a great touch imo.
Also, I'd recommend adding the ability to use shift to define multiple moves in a row (hold shift, click to define a move here, then click again somewhere else to define a sequential move, etc). Supreme commander uses this for movement, building, etc, and is extremely useful/convenient imo.

Make sure when you're setting the material that's it's defined as a "sharedMaterial".
That should enable dynamic batching

i'm almost certain that age of empires 2 had some waypoint system that i just never bothered to figure out how to use
is there any way to set a material to be shared from the editor or is it in code only?

afaik it's code only, but it's pretty simple.
Just use the unit's mesh renderer component, and when setting the material do it as sharedMaterial instead of setting it as material.

Mesh renderer component:
docs.unity3d.com/540/Documentation/ScriptReference/Renderer.html

setting shared material:
docs.unity3d.com/540/Documentation/ScriptReference/Renderer-sharedMaterial.html

oh no wait, nevermind, apparently only non skinned meshes get dynamically batched
so the spears are getting batched, the bodies aren't
for some reason i thought they had figured out how to do this with skinned meshes by now

You should be able to use this technique to make good poly hair.

They are pretty potato headed but you have a firm base clearly, keep practicing and you will be good enough to work as the character artist for Deus Ex HR or Falout 3. Try polishing up your work.

ah, skinned meshes

The best bet you have now is making different LoD models (so shaders aren't working as hard), and icons when zooming out far (like supCommander).

How do these look so far? I'm thinking of making several modular "props" to make Roman buildings out of

Like here's an example of a forum or something I guess.

My palette isn't very good, unfortunately.

For creating fonts:
Torrent photoshop. forums.adobe.com/thread/1208358http://www.creativebloq.com/typography/create-your-own-fonts-photoshop-or-illustrator-101517254

or use GIMP. Despite the fact that GIMP is clunky as shit and lacks many of the specific features and refinement that Photoshop has, you can still make it work though I don't think you can make TTF/OpenType fonts unless you use a plugin or write one yourself.

Editing text:
As far as a text editor with programmable output is concerned, I think you're going to have trouble that completely meets your requirements. Text handling, despite whatever library you may be using, will require at least some boilerplate on your part due to most text implementations being context specific. If you had the autism you could probably kludge together an interface library built on top of LaTeX and use that. But I usually just format text by means of linking the output of a UTF-8 character stream to a streaming parser that operates on tag symbols for switching formatting contexts. After a certain point you have to be willing to ask yourself if you need all the features you think you need or if a simpler, custom solution would be more pertinent to your endeavors (from a time perspective).

For packing bitmap fonts w/ information on the structure of the output format(s) for building your own parser. Take a look at libgdx or Slick2d for code examples. They both use Hiero for generating glyph sets/bitmap fonts:
angelcode.com/products/bmfont/images/bmfont.png

For zipping together or creating ttf fonts:
fontforge.github.io/en-US/

For editing text:
Any editor that supports the specific encoding (ASCII/UTF-8/16/32) you want to use and some tagging system for formatting. Another thing you could do is port Lua to your engine and build in bindings for text manipulation. Lua can be useful as a flexible config system or provide dynamic formatting for text.

====
If you absolutely need the functionality you describe you will need to build your own tool chain + frontend tool. For it to be flexible you'll have to design it from the top down and try to restrain yourself from writing any code until you know exactly what you want to achieve and the stages at which you go about achieving said ends.

Edit: I linked an image of the program as opposed to the actual program.
angelcode.com/products/bmfont/

You making like a new Caesar game? your art works, I'll give it that.

minor updates
now if you don't make an entire rectangle while selecting units, but just do one quick mouse click, you select whatever the mouse is hovering over. should have done this ages ago
also i made the formation making a bit faster. if the leader of a squad is more than 90 degrees from his target, the squad gets resorted based on their distance to the target
so when the squad is moving forward, nothing changes, but if i tell it to go in the opposite direction, the furthest row of troops will now be the first. that way there's no need for the squad to do a really rong rotating movement
so next i'm doing waypoints
and for some reason i've got this bug, where when i select units their icons don't seem to be scaled correctly. and it seems to happen at random. when i'm recording, i see that with all the units selected, the last line of units is going through the bottom border of the screen. but when i turn the game off and on and don't bother recording, they're scaled correctly and fit in the screen. go figure

I appreciate the video and advice. So should I just keep sculpting off reference till I git gud then? Also, how would I go about gauging my own skill level. Like, how will I know when I'm mediocre or good?

i also got waypoints working surprisingly easy, you just shift+right click
the waypoints can be reset if you just right click to a new destination

Thanks for the reply, but I wasn't referring to anything to do with fonts, so I suspect I didn't describe what I'm looking for well enough.

I want something that lets me fiddle with text in a tree format similar to pics related, and that has some fully scriptable export feature so that I can run said tree through scripts that I would write to, eg. generate a fully functioning lua file that I could load up as monster info, or a spawning script, or dialogue trees, or really anything.

TileStudio lets you do something similar with sprites, spritesheets, maps, etc. which has been fun to play around with, which is why I was mentioning that as a rough example.

It's the same as drawing, practice makes perfect, practice every day and you will still be bad, like me, but you won't be as bad.

I don't know how to gauge skill level, I think it's measured more in proficiency, like are you able to make x or y. I normally redo my last years work and compare to see how much I've improved.

also post your content to /3d/ it's a slow as fuck board, but I check it every week to catch posts and tell people how much more than me they suck

post your stuff to /3d/ as well, you need more criticism

i'm at a conundrum
if i tell my squad of units to go to a location, that they don't have a possible path to, they'll walk to it, but the squad's movement will never stop, since i'm checking if it's finished by making sure every unit is less than 0.5 m from their target destination
i could check if every unit has a path to end, but that would be very costly, for more than 2-4 units, at least with navmesh
wat do
any other way to check if they can't reach their target, but have already gotten as close as possible?
i was considering checking the magnitude of their movement delta, but that could just mean that they've temporarily been slowed down/stopped, not that they've reached their target as close as possible

im not sure how your pathfinding works
but if you are using some kind of A* variant it should be relatively easy to adjust to algorithm to always return the position closest to the target

i'm using unity's navmesh
the issue is that i need to know when the whole squad of units can't move any more.
for example, like in , if i tell my squad to go to 2 waypoints, and the first one is unreachable, they'll keep walking towards it and never be told to go to the next one, despite being as close as possible to the first

Satisfyingly tidy user.

in that case you could set a different waypoint for each unit based on their position in the formation
or use the center of mass of the formation to determine when they have reached their destination

actually yeah, i could just get the middle point of the squad and check if it has a path to the target destination, that shouldn't be as costly as checking the same thing for 72 units

yeah for your purposes you probably want the second option

What if the middle point is out of bounds?

actually not sure
is there any way i can check whether a vector is within the navmesh?

SamplePosition with a small distance?

I dont really have a way to have shadows on things …. i need to really redo my palette so things have more than 3 darkness levels and theres more contrast

alright, the idea for the middle point didn't work because reasons
so i had another idea
before the units start moving, i set their final destination to the closest reachable edge
and when they're close enough to it, they directly go to it rather than following their neighbors, so i don't have to question whether they've reached their destination if it's unreachable
now this works when i tell them to go near the edge
those that are supposed to go beyond the navmesh stop at the border, and are considered as if they've reached their target
and the squad is treated as if it's finished it's movement
so if i set a waypoint to the border and then somewhere else that works
my issue is that when i tell the squad to go to the "not walkable" block, this doesn't work. they still keep walking towards it
any clue why?
for (int i = 0; i < units.Count; i++) { Vector3 dest = (pos - mainDir * ((i / rowNumber) - (units.Count / rowNumber) / 2) * dist + targetDir * ((i % rowNumber) - rowNumber / 2) * dist); if (NavMesh.SamplePosition(dest, out hit, SquadWidth(), 1

oh wait, unity is just fucking with me

alright, so for the moment i solved this by lifting up the cube, rebaking the navmesh and putting the cube back into place
is there any other way i can get unity to just not treat the cube as if it's hollow?

added a single line formation
kind of bugs out if the line exceeds the map size though

Any reason why your making everything so low res? looks like you need a few more pixels

Looking good, god knows we could use some more RTS gaem.

MT. DEW

i sort of got a wedge formation going
i have 2 issues though
1) i have no idea how to keep up the formation while they're moving. right now i'm just telling them where they're supposed to end up, nothing while they're actually moving
2) for some reason, when changing formations they may or may not all go in 1 spot. no idea why

Mostly I'm too stubborn to learn how to make 3D shit and export it as an isometric sprite. So I end up doing it by hand instead, and 32x32 is like the default size for that sort of shit.

Helps if you can deliberately turn off the analytical part of your brain user (turning this off: how can I put this into code, that's not possible w/this approach, etc).
That's where I generally come up with my most creative ideas, solutions, and concepts of an algorithm or ideas, as in, turning my pattern recognition to overdrive, and completely ignoring the "how" of an idea, and only with the "what" of an idea.
That's where the crazy idea train goes on overdrive (which is really just seeing new patterns, linking them together in novel ways, etc), and I can come up with some pretty enthralling concepts/ideas/solutions to later put into algorithms when I turn the analytical part back on.

Finished up my short and long range ambient occlusion for terrain using noise to determine intersections w/surface -> i.e occlusion value.
Turned out pretty well, and I'm happy with the results; pics provided.
I can always step up the sample amount if I want higher details as the overhead is negligible, but atm it's good enough to move on to implementing more features.

Currently rewriting my terrain generation algorithm to support CSG (constructive solid geometry), which will allow for way more customization of terrain outputs via proc gen algs, and allowing for way easier editing on the fly; in addition to being one of the preparation steps I have planned to being able to implement my procedural modeling system.

Have a bump

Also, that looks nice

fuck this

At you least you got far enough to have something playable

In 2 months, I've probably redone basic boilerplate code like 5 times and changed my mind on the project just as often

I've spent over a year at this point making my game. It's not really playable yet as I'm missing a lot of features. Going over the inventory again for about the third time and I feel so braindead atm that even writing functions for the inventory to check for space etc is aggravating.

Keep working on your game until it's fun, or cut the losses, take what you've learned and make something else.

k den tbqh smh famalads

If I want to make characters for my game should I focus on learning in a 3d package or hone my drawing skills first?

will game be 2d or 3d

3d

why would you need to draw then

practice drawing, start with shading and object construction/deconstruction

Store shelves.

Breh, didn't people tell you? Making the game is the easy part, around 10% of the effort. The hard part is making it fun.

so i'm going to guess that Voxel Farm or something along those lines are the only sensible options for voxel usage, as Atomontage is only built by a single dude, and Voxlap is fucking ancient by this point

Your mistake is thinking that making mistakes is a waste of time. Need to do it wrong before you can do it right.

Both.

this perspective confuses and angers me

...

Buttons! Will be connected to dynamic props like doors etc in the future

OH by the way my game Hypnorain on Steam is 75% off for a few more hours, in case anyone is interested: store.steampowered.com/app/495230/

So you're telling me that someone from /agdg/ actually delivered?
I don't like bullethell shooters like, at all. I put it on my wishlist though. Gotta help an user out and who knows, maybe your game will be the one to sell me on the genre. No money at the moment though, so I'll buy it at some point in the future.

Bought it to help out a fellow user at discount price. I hope you enjoy your $1.50 profit.

i got placeholders working with the Legacy gui system since it's a bit easier to manage for this
but from the looks of it i'm definitely gonna have to use the newer one even for this. drawing 3 textures, for all those units is currently putting me at around 1.7k batches

oh wait, it's only 300 batches, no clue where i got the last number from
still, without the healthbars it's only 1 batch per unit, so i still need to use the newer UI instead of legacy

Is there any way to get it drm-free? Or is even the demo only for steam users?
t. shmupper pro

...

I want to make an online game for Halloween, ideally in Love and having the server side stuff written in lua too, since I'm familiar with doing that. I've never delved beyond the usual shared hosting stuff, so I'm a little unsure.

Am I right in thinking I'd need VPS hosting so I could install dependencies via luarocks then run the server script with luajit?

If it's true, and I wouldn't know since I don't use Godot, you can still just give a function to the brick that tells it to destroy itself, and on the ball (which DOES report itself hitting the bricks right?) say other.thatfunction or whatever way Godot gives you a reference to the other object.
Long story short, learn to fucking program first.

alright, now i've got another silly issue
i'm making the healthbars work with the normal UI system, and their position and everything works
i'm checking whether they're within the window rect in which the units are visible, so that i can destroy the healthbars or instantiate them
but for some reason GameRect.rect.Contains only works for the bottom left quarter of the actual rect
UnitSelection.Squadron squad = UnitSelection.Instance.selectedSquad; if (squad != null) { for (int i = 0; i < squad.units.Count; i++) { Unit unit = squad.units[i]; Vector2 pos = Camera.main.WorldToScreenPoint(unit.pos()); if (GameRect.rect.Contains(pos)) { if (unit.Healthbar != null) { unit.Healthbar.transform.position = pos; } else { CreateHealthbar(unit, pos); } } else { if (unit.Healthbar != null) { Destroy(unit.Healthbar); } } } }

huh, for whatever reason, the rect has a misplaced position
Rect temp = new Rect(GameRect.rect.x + Screen.width / 2, GameRect.rect.y + Screen.height / 2+30, GameRect.rect.width, GameRect.rect.height); if (temp.Contains(pos))
now it works

It's only on Steam at the moment, but it is DRM free.

So when Steam says I'm not allowed to download it, how are my digital rights not being managed?

sorry man but thinking anything is drm free but at the same time requires a steam account to download is delusional as fuck

Jesus Christ, just use a throwaway account.

That would still require me to install jewish software and give valve traffic.

Fair enough. Just don't expect anyone to go out of their way to appease you. Or people like you for that matter. Not many take their freedoms that far.

but that only costs them money
Also steam has a DRM free option for devs which allows people to play the game without having the steam client installed.

You know it brudder, even gog has started implementing drm these days in the form of gog galaxy.

It also boosts their alexa rank and they likely offset the costs by forcing users to view ads and by selling user information. As miniscule of a difference it makes, I'd rather not give Valve anything else.

ok guys i need help with some code
I'm still working on my line of sight system but i think I have decent idea to make i work
I will construct the vision from triangles which will look round because of a circular texture this part i already got working
The issue is calculating which line segments block vision and which don't
I can get a list of all walls in vision range but those are all rectangles and only 2 of the faces of the the rectangle will ever block the vision so how do i figure out which faces those are in an efficient way

i've got a performance/gameplay question
i can take my rts units, and turn them into static meshes, after i've posed them
then, since they're not skinned, they'd be getting dynamically batched
i can essentially make all the units, and a few weapons/armours into a couple of draw calls, so that i can have even more units on the field
alternatively, i'd have to get at least 1 draw call per unit, but they'd be animated
considering this is an rts, would the lack of animations be worth getting extra numbers?

Are you imagining having an rts game where units just slide around and shoot projectiles out of mid air?

well chess is a thing, so yes
besides, i can still move the models, just not pose them

chess isn't an rts, dipshit
hence me asking if you're going to just have them slide around

how can you possibly be this stupid?

i can make them hop while they're moving

this sounds like nonsense to me
I have fully animated units in my rts and i only have like 4 draw calls in sfml

well this is unity we're talking about
we don't have fancy rendering magic
at least not for skinned meshes
;_;

psst use imposters

It's not fancy rendering magic at all
I just put all my vertices in one array and i can render whatever I want in draw call
It really just baffles me that this is somehow in issue in unity

Slightly off-topic, but this is a sentiment I've seen a lot of, and it's a little worrying to me since I'm a fan of dodging lots of bullets and I was making my game with it in mind.

Here's a question for you, /agdg/.
For you, personally, how do you balance between making stuff you like to play and audience criticism about gameplay stuff?

beats me

if I'm in the mood to play a bullet hell I just want to dodge bullets and not deal with a million gimmicks

aren't i supposed to use that only under certain angles, or at a certain distance?
in my case the camera angle isn't changing, and the distance from the camera to wherever it's pointing it is almost always gonna be the same, not to mention that it's orthographic, not perspective

If they're not a fan of the genre, don't bother. At some point you have to stop trying to please everyone because otherwise you'll have a fucking mess.

are there any exercises/work sheets that I can use to practice learning programming with Unity?
it's the best way I learn things more

do you have the version where it's yandere-chan overtop yandevs colossal bool/string comparisons fuckup?

i really wish i kept that one

Bullet hell is a niche genre, but you have to make the game you want to play. Both so you won't stop working on it, and because it generally results in better games.

So im generating a vector with angles where objects interrupt the line of sight
and im saving the angles like this

for (int i = 0; i + 1 < lines.size(); i += 2)
{
anglesSet.push_back(atan2f(lines[i].x - center.x, lines[i].y - center.y));
anglesSet.push_back(atan2f(lines[i + 1].x - center.x, lines[i + 1].y - center.y));
}

Now i just need to figure out how to turn this into the angles that do need to set
Anyone got any ideas?

I'd help you out if I could, buddy, but I have no idea what you're asking

I have a unit who has circular vision
Certain parts of that vision are obscured by line segments
in AnglesSet i store the two angles of the cone of obscured vision
I just need to figure which parts of the vision arent obscured so i draw the vision area correctly

It's a taste thing and I'm going to agree with and on this one. Although I will say that not listening to people who aren't into the genre at all isn't something you should do. At least ask what exactly it is that they don't like. Additional feedback won't hurt, even if you decide not to implement their ideas in the end.
Bullet hell isn't just a niche genre. Horror games are a niche genre. Bullet hell games are a niche genre by niche standards. You won't get a whole lot of sales so the best thing to do is to make it something you'll like. As far as I am concerned, I really enjoyed the shit out of Tyrian, so I like the general concept of the vertical shooter. Besides that, I love Geometry Wars which basically turns into a bullet hell at about 150k points (fuck the sneks in that game though. Stupid pieces of shit. The developer who thought I would be a great idea to make them spawn around you deserves a special place in hell.).

Ah, I see.

I think the easiest and fastest way to make the 'circular vision' work is to just find the areas that they can't see.

I made an mspaint drawing.

First, assume everything that's too far away can't be seen. This is trivial.

The yellow circle is the unit, and a set of two red lines connect to each black 'line' you've made to block sight.

They're then projected outwards a set distance. It doesn't matter how far, as long as they're guaranteed to be off screen. Once they're projected, you can add another line "off screen" to complete the green polygons.

Each green polygon represents what cannot be seen. All of them have four points, so everything that isn't contained in those points can be seen.

You can probably do it with other shapes. I'm not sure, I'm awful at math.

I can't really understand what you're going for. Be more concise.

This is what you're looking for:
wiki.unity3d.com/index.php/MeshMerger
It combines all meshes into a single mesh, so there's one draw call (i.e. passing it once to the graphics pipeline, singular vertex shader for all verts/etc).
It moves the GOs (transforming vertices), independently, via the bones built.
Also, you need to be using the same material for all of the GOs, so it's one draw call per unit type w/the same material for ease of use.

Here's an explanation of what it does:
archive.is/tSN9M

Also, for future reference if you need nearly the same thing for skinned meshes:
wiki.unity3d.com/index.php?title=SkinnedMeshCombiner

Entirely possible in Unity, but it's not always necessary; as most of that is abstracted away if you don't have a special use case.
As you can access the low-level graphics API (openGL, directX) if you need to (easily), but most don't bother as they don't need the performance gains.
I use it for my procedural meshing system, as I have millions of verts, and need the performance gain due to it being a special use case (procedurally generated meshes via GPGPU, that stay entirely on the GPU).

They could also be used as regular polygons for a 'fog of war' sort of thing.

Here's one with circles, because I love you so much. Just assume that nothing inside the circle barrier can be seen, and then project from the edge of the circle

I'm not sure how to mathematically represent the two points on the circle that you have to project from. Maybe someone will help me out.

unfortunately excluding areas of vision won't work
I need to draw to actual vision area because of blendmode fuckery
What i need is to figure out the angles between obscured vision parts
So basically the parts besides the green segments in the drawing
I already have all the corner points of the green rectangles
though only closest points are relevant so it really just boils down to line segments

It's been a while since I've used blendmodes, but I'm sure there's a way to invert it so everything draws except the 'hidden' polygons

The issue is if im drawing the fog one units line of sight fog will be drawn over other units and fix this is a huge pain

Just use… trig
If you know the vectors of the two lines creating said angle (i.e. "I want these angles" part), you can determine the angle between each.
Look up: "angle between two vectors".

Does this help?
redblobgames.com/articles/visibility/

that's a smart approach.
Red blob always has good stuff

odd… do u have the VS tools installed, and have you set VS as the preferred IDE in unity?
I've never had that issue after doing the steps above.

anyone have torrents for FL Studio?

not a bad idea
but the issue is that i don't know which lines are next to eachother

i'm trying the skinned mesh renderer combiner, but i'm running into issues
first of all, CombineMeshes is limited to 16 at a time
so i think i fixed that in the start, however from what i can see only about a 4th of the troops are combined
and the worse thing is, that the created mesh is warped weirdly. as far as i can see all my units are a (1,1,1) scale, so i don't think that's the issue
it did reduce draw calls by 30 though
is my math wrong or something
using UnityEngine;using System.Collections.Generic;public class SkinnedMeshCombiner : MonoBehaviour{ void Start() { SkinnedMeshRenderer[] smRenderers = GetComponentsInChildren(); int RepeatTimes = smRenderers.Length / 16; for (int q = 0; q < RepeatTimes; q++) { List temp = new List(); for (int i = 0; i < 16; i++) { temp.Add(smRenderers[i + q * RepeatTimes]); } CreateMeshGameObject(temp.ToArray()); } } void CreateMeshGameObject(SkinnedMeshRenderer[] smRenderers) { List bones = new List(); List boneWeights = new List(); List combineInstances = new List(); List textures = new List(); int numSubs = 0; foreach (SkinnedMeshRenderer smr in smRenderers) numSubs += smr.sharedMesh.subMeshCount; int[] meshIndex = new int[numSubs]; int boneOffset = 0; for (int s = 0; s < smRenderers.Length; s++) { SkinnedMeshRenderer smr = smRenderers[s]; BoneWeight[] meshBoneweight = smr.sharedMesh.boneWeights; // May want to modify this if the renderer shares bones as unnecessary bones will get added. foreach (BoneWeight bw in meshBoneweight) { BoneWeight bWeight = bw; bWeight.boneIndex0 += boneOffset; bWeight.boneIndex1 += boneOffset; bWeight.boneIndex2 += boneOffset; bWeight.boneIndex3 += boneOffset; boneWeights.Add(bWeight); } boneOffset += smr.bones.Length; Transform[] meshBones = smr.bones; foreach (Transform bone in meshBones) bones.Add(bone); if (smr.material.mainTexture != null) textures.Add(smr.GetComponent().material.mainTexture as Texture2D); CombineInstance ci = new CombineInstance(); ci.mesh = smr.sharedMesh; meshIndex[s] = ci.mesh.vertexCount; ci.transform = smr.transform.localToWorldMatrix; combineInstances.Add(ci); Object.Destroy(smr.gameObject); } List bindposes = new List(); for (int b = 0; b < bones.Count; b++) { bindposes.Add(bones[b].worldToLocalMatrix * transform.worldToLocalMatrix); } GameObject tar = new GameObject(); tar.transform.SetParent(transform); SkinnedMeshRenderer r = tar.AddComponent(); r.sharedMesh = new Mesh(); r.sharedMesh.CombineMeshes(combineInstances.ToArray(), true, true); Texture2D skinnedMeshAtlas = new Texture2D(128, 128); for (int i = 0; i < textures.Count; i++) { } Rect[] packingResult = skinnedMeshAtlas.PackTextures(textures.ToArray(), 0); Vector2[] originalUVs = r.sharedMesh.uv; Vector2[] atlasUVs = new Vector2[originalUVs.Length]; int rectIndex = 0; int vertTracker = 0; for (int i = 0; i < atlasUVs.Length; i++) { atlasUVs[i].x = Mathf.Lerp(packingResult[rectIndex].xMin, packingResult[rectIndex].xMax, originalUVs[i].x); atlasUVs[i].y = Mathf.Lerp(packingResult[rectIndex].yMin, packingResult[rectIndex].yMax, originalUVs[i].y); if (i >= meshIndex[rectIndex] + vertTracker) { vertTracker += meshIndex[rectIndex]; rectIndex++; } } Material combinedMat = new Material(Shader.Find("Diffuse")); combinedMat.mainTexture = skinnedMeshAtlas; r.sharedMesh.uv = atlasUVs; r.sharedMaterial = combinedMat; r.bones = bones.ToArray(); r.sharedMesh.boneWeights = boneWeights.ToArray(); r.sharedMesh.bindposes = bindposes.ToArray(); r.sharedMesh.RecalculateBounds(); }}

That doesn't work you faggot. Did you even browse previous threads? And I think I explained this enough in my post.

oh i'm retarded, it's temp.Add(smRenderers[i + q * 16]);
now it gets MOST of the troops
the mesh is still fucked up though

alright, seems like the mesh is rotated for some reason, and that fucks it up during animations
rotating the gameobject back to (0,0,0) doesn't really help though, the mesh has already been set to the bones
wat do

oh dang, that makes it a lot more complicated.
You could always pre-determine this when adding them, and reformat what you have to work with so it's easier to find a solution.

use torrent project or some shit u lazy nigger, it's one search heh
Make me some legit synthwave tunes though, and I will be forever grateful.

haven't actually used that script btw, but will try to help
haha holy shit, what happened there, verts are all over the place, same with frags kek

good stuff, progress

Try only having 16 skinned mesh renderers, as you're currently truncating the result (with the int cast) if the .Length method result isn't divisible by 16.

Haven't looked at the rest yet… lmk

To make it more dynamic (if this is actually creating one of the issues, and u need it to be dynamic… already typed it out so yeah…):
Have the repeatTimes evenly divisible by the amount of smRenderers.
So the int used for repeat times isn't inaccurate, as, f.e. if you do 40 / 16 it truncates the results into: 2 (actual is 2.5), but we want 2, 3 or 4 evenly divisible here).
So, what you'd do is, say you have 40 smRenderers, do:
int len = smRenderers.Length;int divisibleBy; //outerloopif (len % 4 == 0) //can have a few to divide by divisibleBy = 4;int innerLoop= len / divisibleBy; //10 //amt of skinnedmeshRender per combined, not truncating value, hence not improper combining initial step

have, temp.Add index as:
[innerLoop * q + i]
//as you're just going from 0-9, then 10-19, etc.

you could just sort them all into an array by their angle, then you can find the neighbors for any angle in constant time. The sorting might be slow as shit, though.

for posterity, here's a mobile game I was working on that never got finished

well the main issue is still that the mesh is all fucked up for whatever reason
but i'm gonna give up on merging the meshes, at this number, there's a bit of a delay for it to happen, not to mention it will be a huge pain to constantly monitor which meshes need to be combined
even if we get 800 units on the screen, 800 draw calls is bearable, although framerate might have been lowered by a bit. if AOE2 can get away with it, then so can i

Yeah this is probably what i will do

alrighty

I'd try out that mesh merger script though (the non-skinned one), and test it on only a few meshes at once (like 3 or 4) to get it working, and to figure out how it works (break it, fix it, etc).
You can also do stuff like this over multiple frames too, and only do it if you really need it (f.e. over 100 units on screen at once, and have pre-pooled resources so it's fast, and you're not creating whole new arrays and using slow as shit lists).

do i really need it right now?
non-skinned meshes get batched anyways
and if i combine non-skinned meshes, the result will have to be a static object that never moves, or else i'll see that it's combined
like i can't combine the spears right now, because they're each parented to a spearman
as far as i understand, i should be using the merger script only for the static decor
in other words, i won't need it for a while

look at the first link, it works for moving non-skinned meshes too.
As "moving" is just transforming the vertices, which uses bones in skinned meshes, which can be replicated for a non-skinned mesh; which is done in that script.

You don't necessarily need it atm, but eventually you will if you have plans for a large amount of units; although premature optimization is the root of a lot bugs.
It's a pretty fundamental optimization step, but it is best to only do it when you need it; it'll also add more motivation to actually doing so too.

oh, so it makes one big skinned mesh, and tells each submesh in it, that their bones are whatever they used to be a mesh to
in my case, if i have a GameObject with a meshfilter for the spear, the transform of that gameobject will now become the bone for the new combined mesh
i guess that could work, i'll try it tomorrow

If it happens to be tile-based, you can use the "my algorithm" section here
adammil.net/blog/v125_Roguelike_Vision_Algorithms.html

I have completed a new level for my game, and released a new version with all the improvements that I made since the last version.

Right now I have a lot of in-development levels that have gotten more complete, and one level that was in development last version is
now complete. I have also full implemented the "major" enemy, which was in the game for several versions now without a full implementation.
You can fight him on the level "forum" and the level "outside" even though he is in that level by accident (I meant to place a "captian" enemy in the editor)

My development website is:

ika.neocities.org

If any of you feel like giving it a try, I would like to know what you thought of it, and what I can be doing better.

neato, I will be checking it out soon.

Damn what did you do with the movement? Now it feels really sluggish.

Make it at least smooth like in Unreal 1/Unreal Tournament 1999

starting to work on getting units to attack
currently i have no idea of how to get them to stay in formation while attacking. do only the first lines attack? do they move to the middle of the enemy squad and then spread out to attack anyone in range?
right now each unit just checks in a surrounding bounds, if there's any enemy unit there, if so, he goes to said unit and starts attacking

well i got units getting destroyed to work
and if they don't have an enemy, they try to find something within their range
so right now, both groups can successfully wipe out one another
now i need to figure out how to get them not to clump up in one spot
i tried setting the navmeshagent avoidance to high, priority 50, i tried changing the radius of the agent, but there's always a group of units that are way too close, while others work as intended

kind of improved it by getting them to attack the closest troop, not the first one they find
also got double click selection to work, on screen units only
still not quite sure what to do with the formations when they're fighting
maybe i should just give squads the option to either attack at random like they do right now, or to stay in formation and only for units that are in range to attack, without moving

a way to fix this would be to turn on collision when units are fighting
Also when you are unsure about something being one way or another you can always make it an option

now where getting somewhere

DAMN this is a really nice FOV effect. Much better then how Spring RTS does it.

btw why are you not lurking at #8/agdg/ IRC channel anymore?

I dunno it just seems a little pointless
maybe I should though

Not sure what you mean, do you think you can elaborate?

Do you want the player to accelerate faster? Decelerate slower? Move faster?

I have tweaked it but im not sure what to change about the movement to make it better.

Which versions did you play that you think had better movement? I can copy the movement parameters from the earlier versions into the new version.

why don't you just make so you hit top speed in like ten frames or 0.2 seconds if thats easier

Unless I missed something but my 0.7.5 play through I accelerated very slowly and I have to jump constantly to move faster.

mhm I can't exactly remember it anymore because I always replace your game with the newer one, so sorry about that one.

The movement is something that everyone has a different opinion about, some people dont want the game to use acceleration so you just snap to and from top speed, and other people want tons of friction so you cant slide around, but I want to make it smooth like doom. There isnt really an easy answer.

Okay, so its a speed problem… That makes sense because movement speed was slightly faster earlier. I will make the acceleration higher and send you another exe that you can check.
If you want to check older versions of my game, every release can be found on ika.neocities.org/.

Okay, here is a test where instead of accelerating to maximum speed of 12.something, you go up to 13.9999…. The acceleration value was changed from 25 to 28. Friction was not changed (-2).

p.trumpshare.com/h0gd/lbt3.zip

The top speed of 0.6.6 was ok, the acceleration could be faster but it's alright.

The only gripe I had with that the sliding is way too strong, it is like the floor is made out of ice instead of concerete, wood whatever.

Your test version improved it a bit but not much and the sliding is just as bad.

it kinds of works
but now i've got the issue that a lot of troops end up attacking a single enemy, and many in the back can't reach him
technically they should just go around, but for some reason they ignore the fact that someone's infront of them and just keep trying to run forward
if this was grid based rather than navmesh, this wouldn't be an issue, since the units themselves would serve as an obstacle, and they'd just have to find a new path

you could try adding units to the navmesh as obstacles once they are attacking
though it might be quite expensive im not sure how unity handles it's pathfinding

unity has NavMeshObstacle component that's supposed to do that, used for moving targets
but adding that to a gameObject that already has NavMeshAgent component to it, is only gonna fuck shit up

I like the sliding, its not even as much as the sliding that doom has. I CAN make it so that all the sliding is gone and the accel is really fast, but I don't like it that way, because it wont feel as smooth as the slidy movement.

The top speed in 0.6.6 is about as fast as the top speed in 0.7.5 which you complained about being too slow, and the acceleration is much slower than 0.7.5

so, im not sure what you mean… the accel in 0.6.6 is slower than 0.7.5, and the top speed is the same, but you complained about the accel in 0.7.5 being too slow?…
So it sounds like a contradiction, so im not sure what to do. The friction in 0.6.6 is much higher than 0.7.5, which tones it down a lot. It sounds like 0.7.5 would have fixed
all the problems you have with 0.6.6, so im not sure why you dont like 0.7.5. Maybe you could elaborate more?

made a new thread