/v/ - Video Games

Do people still use >>>/agdg/ or the >>>/vg/27335 thread?

What about the wiki? IRC? Please don't ever use Discord

>>15186883

for example i was thinking of doing a metroidvania because my favorite games come from that genre and i honestly love the idea of exploration and great bosses. how would one start off in order to make a good metroidvania. do i have to play the classics first to understand how a metroidvania works.

>>15186894

Make a character that's fun to move.

Make enemies that are fun to fight.

Make abilities that are fun to use.

Make a world that's fun to explore.

Done.

>>15186901

sounds simple enough. but im worried that my execution would be shit. i had an idea of a game where you play as a psychic boy called oracle who would go around and kill deities. i was also thinking the setting would be urban fantasy

FAPPING IS THE ULTIMATE DE-MOTIVATOR

>>15186880

When I have progress I'll probably post more on those two than the /v/ threads. For now, the progress and questions are here so I'm mostly here too.

>>15186883

>>15186894

>>15186913

If you are going to use Godot, definitely go through the tutorials at http://docs.godotengine.org/en/3.0/ before trying anything fancy. You should probably check the 8/agdg/ wiki too for some useful advice and links.

>>15186880

I only go here, sometimes edit the wiki articles.

>>15186913

>im worried that my execution would be shit

You haven't even learned to just like make game yet, it's too early to worry about game design.

>>15187217

don't do animations for characters in unity, the support is generally halfassed

>>15187256

you mean, do them in blender and export?

>>15187282

yes

>>15187466

I like the first one. It sounds hopeful in a weird way, like the world is falling apart but it's going to be ok. It's reminds me a little of Circus music.

The second one sounds like a mess, which may be the point but I don't really like it. It's a bit much for my taste. This is coming from someone who loves Death Grips and Street Sects.

>>15187814

Use c++

>>15187814

What the hell are you trying to do with them?

>>15188342

He could be trying to use a memory buffer to prevent mallocs. That's the only thing that would justify 1.4k lines of code for fucking strings.

Speaking of optimizations, what's a good way to store a byte[] that you can write to a whole bunch without necessarily reallocating it? I'm familiar with the concept of a flexible array where you resize it and double the size, but I don't know if that's ideal. A List<T> does this under the hood in C#. Can you write to an arbitrary bytestream without an output target? Like

>>15188663

what?

you want a byte array then just allocate a byte array

ded tred

>>15188746

I don't want a byte array, because I have to resize it, which is a waste when it gets larger and I have to increase the size. I was looking for alternatives.

For some reason, memory streams are twice as fast; I usually get

114344us for the first test, 200849us for a byte[]

>>15188821

Whoops, my timer code was fucked please ignore my retardation

>>15188821

>>15188851

BTW the fastest way to do dynamic arrays, is to allocate a gigantic pool of memory. Like all the dynamic memory you'd ever use in the entire program, and then every time you need more memory section it off of that pool.

>>15188821

https://en.wikipedia.org/wiki/Dynamic_array

https://en.wikipedia.org/wiki/Linked_list

There is no silver bullet for your problem. You have to choose between being able to resize your data structure and your data structure being fast in any scenario.

>>15188903

Linked lists are the slowest memory container.

>>15188929

Is that a challenge?

Has anyone here tried using libgdx and what did they think of it?

This is for my input recording/playback for a level.

I think the optimal solution here is to make a List<byte[]>, and instead of resizing it, it just creates another buffer as needed and throws it in the list, tracking the length as it goes. When it's done, it just writes them all to a single byte array for playback or storage. This might actually be worth implementing as my own IList<T> implementation later; I'll have to check on my day off.

However, a good old-fashioned List<byte> will do just fine, since it handles all the resizing under the hood, and it can still copy to an array pretty quickly. I doubt any particular level will have more than 10,000 inputs to record, anyways.

>>15189047

People wouldn't use proprietary shit if blender wasnt so shitty

>>15189074

I thought log (n) was better than n, because if n is less than 10, log10(n) would be between 0 and 1 which is less than 1

>>15189110

It is. Where are you getting that it's not? The key on the site has says log(n) is "good" and n is "fair".

>>15189074

Big O doesn't account for the speed of cache or system calls, which are 2 areas the Linked List suffers heavily in.

>>15189087

It's not just blender. It's the Khronos group and Autodesk. Collanda was a fucking mess. I've been trying to do skinning with GLTF 2.0, which is Collanda 3.0 and it's a fucking mess. It's wasted a fucking month of my time. I really REALLY do not want to switch to .fbx, but GLTF is just so fucking shit I might have to.

>>15189124

>finish work and had to go back three times because i forgot to return shit like the keys

>forgot my sunglasses on the bus

>made two obvious typos in my code and fucked things up

>cant fucking tell the difference between yellow and green on a chart

Okay time for sleep, I'm too retarded to stay awake right now

>>15187466

First sounds like it'd make great music to find diegetically, like on the PA system of an abandoned mall or something. Shit screams decay.

The second one feels like two separate songs that just play over each other for a bit. I think the transition could use more work.

>>15189368

>I hope at some point in your programming career you think back on your post and feel embarrassment.

I'm sure you know more than google engineers who manage code that runs on gigantic data centers.

>>15189368

That looks really nice anon.

>>15189368

That's just a fact of how caches work, on an array you have a guarantee of spatial locality that a linked list doesn't have, since the allocator can just put the memory in the list anywhere it wants, meaning that your cache is worthless when the next element in the list is somewhere totally different in your RAM and it has to clear all the lines for that. This isn't any kind of new information.

>>15189515

>>15189381

Yes you have worse cache misses but that's not what I'm arguing about. There is NO such thing as a better data structure. To dismiss something from ALL scenarios on one account (cache), is retaded. There are scenarios where linked lists, doubly-linked lists, etc, are better than arrays. Don't evangelize programming with that mentality.

>>15189538

Can you give us an example where it's better?

>>15189538

>Yes you have worse cache misses but that's not what I'm arguing about.

Lol. I said they're the slowest containers, and somehow that has nothing to do with cache misses? The typical linked list, allocate a new element each time you insert one, has no speed advantage over the typical dynamic array, allocate a new buffer and copy everything, unless you modify it like >>15189046 is doing.

>Every time elements are inserted it makes a syscall much slower than cache miss

>Every iterate to the next element I have a cache miss

Its fucked on both ends. Sure there's no best data-structure but Linked List has virtually no advantage that another structure couldn't do better other than being easy to implement.

Your biggest mistake though, aside from trying to pull the centrists bullshit, is you think in terms of containers.

>>15189598

Theoretically it's better when you want to insert something in an array that's huge, because you just create an element go to the spot in the array, and just link it in. Theoretically, it's better for appending to a fuckhuge array to.

Unfortunately, every time you create the element it calls malloc which is extremely slow because it has to call the OS. It'd be faster to just malloc everything at once. Also, iterating through the array to get to the spot you want to insert into is ridiculously slow because you get hit by a cache miss (which is slower than doing trig functions). So if you insert 500 elements into the list, you are doing 500 trig functions. If you absolutely need to insert in the middle of a gigantic list use a tree.

It's really hard to dig up exactly what the benchmarks for all of this are since it varies between hardware. If you know the concepts of how hardware/OS can slow programs down though it's easy to see how linked list is extremely slow in every regard. If that's not enough for you here's a Google engineer who works on phones and data-centers saying "never use them".

>>15189538

>There is NO such thing as a better data structure. To dismiss something from ALL scenarios on one account (cache), is retaded

Fucking MIT niggers.

>>15189604

>Lol. I said they're the slowest containers, and somehow that has nothing to do with cache misses?

Yes. Because there is also insertion, deletion, etc., not just reading.

>Every time elements are inserted it makes a syscall

Only if they are implemented to allocate memory on insertion. There are many variations you can even mix it with arrays.

>Every iterate to the next element I have a cache miss

That's flat out false. There is no guarantee that it will be in the same cache but there is also no guarantee that it won't.

>Your biggest mistake though, aside from trying to pull the centrists bullshit, is you think in terms of containers.

This isn't politics. There is no left and right wing on data structures. If you dismiss a solution forever without even considering it, you loose a solution. Making you a terrible programmer.

>>15189648

I already know why there isn't any real example beyond that, I am just asking since when someone has to come up with a reason for when it is better, they have to consider the arguments about why it is worse and can maybe agree with them.

In that one situation, with lots of random inserts on a large set, I would end up doing a combination of both, essentially deliberately fragmenting my array by having a linked list of arrays that would be accessed with some kind of getter/setter function. Probably would over-allocate the arrays too so they have room for those inserts, and I have to call realloc less… so it's not a totally wrong approach, if my dataset is extreme. But the kind of linked list they teach in say first year CS is just cancer, with one value per list element, it's useful for learning about pointers and how malloc works but you should never actually use it outside of learning to be comfortable with those things.

Maybe I am disagreeing with him since I am still thinking of that strawman, but he says >>15189696 a much better argument, about how you realistically implement it with arrays. So I can't criticize that post if we are just talking about it in a generic way. Although I do think that he has moved the goalposts a little bit.

But who knows what the context >>15188821 wants to use his data structure in? Most likely he can get by with a array, but we don't know what he is doing.

>>15189598

There are probably more examples where it's better to use arrays than linked lists. But here is a guy who breaks it down with very specific scenarios: https://cppexperience.wordpress.com/2014/10/05/c-linked-lists-are-considered-harmful-arnt-they/

>>15189648

I've seen this video before but most people cling onto is because of ad populum and it's a dude from Google. Linked lists are far from perfect but under the right circumstances (in particular when you don't have to traverse the whole list), they can be better than traversing an array.

>>15189826

Writing a custom allocator is not really a hueg task (and you'd do it for trees too) bro.

>>15189841

The best way to go about that is to allocate a batch of nodes at a time, which is essentially what >>15189046 is doing.

>>15189826

>Which suffer…

Second point addressed this

>The system will allocate memory together…

I've worked on phone and console hardware. It can still be rare but there are certain configurations that make exceptions. Never trust your odds in programming.

>seems like you're just saying whatever you can

Don't project Anon.

>They're only better if heavily modified, in which a tree would serve better.

That doesn't dismiss it from being better at all. You said so yourself, you just have to implement them appropriately which is my original point. Trees and linked lists are also not mutually exclusive.

I saw in some video (I think by NVIDIA) talking about how rainbow six siege uses a method to speed up performance by only rendering every second pixel then generating some approximation of the surrounding pixels.

Does anyone have any idea what this technique is called?

>>15190174

Oh nevermind I'm such a dumb cunt.

It was a talk at GDC called "rendering rainbow six siege" and the technique is called "checkerbox rendering" if anyone is interested

>>15188342

Putting things in the middle of a string, compositing multiple strings into one, putting numbers into a string, extract a section of another string, create a string of unknown length, things like that. I need them mostly when dealing with files and text input/commands.

The main problems are that C's variadic arguments are a piece of shit, and that every time you want to make a new string you need to allocate memory for it yourself and then free every single one of them afterwards (thus temporary memory system). A lot of that code is redundant though because I want to support both my custom text struct and C's null terminated strings, and possibly have a third function for a single char rather than a string.

>>15189368

Where's the toilet?

>>15190244

Bottom left corner. Ain't no time for pajeet.

>>15187536

awesome environment modelling, great detail

>>15188219

what is your engine? I end up too burned out after work to do much myself. modelling in Blender is relaxing for me though so I pretty much quit coding my opengl dungeon explorer RPG and started working with Armory and a simple convenience store simulator/adventure. Armory is pretty rough, but looks like enough to make a simple sim and conversation adventure. Armory's skinnable UI needs work tho.

>>15189047

FBX is poorly supported in Blender. From my personal experience. For static objects OBJ works well. For animated objects .3DS and .X seem OK. I'm using AssImp as an intermediary tool to convert to my internal binary format.

>>15189125 Sorry to hear about the wasted time man. Thanks for the tip on GLTF 2. I was seriously considering exploring it.

>>15190589

On FBX, Blender Guru interview with Ton Roosendaal, he describes how Autodesk makes changes to the FBX format yearly making it very difficult to keep the export/import code up to date and working. He also talks about the Autodesk license and how it isn't that friendly or compatible with reasonable people, making it very difficult to include Autodesk's DLL for FBX handling.

What's the best way to lock down my game so users have absolute no rights to modify the software on their own computer that they purchased and have to spend potentially thousands of dollars just to change an in-game texture on their weapon?

I'm bilingual so you can answer this question in either English or Hebrew, whatever suits you best.

>>15190929

Joking aside, the only true way to do this would be to have a game that takes place on servers, have the server check the validity of the user's in-game equipment, and only have official servers. At best you could then mod the game to give you premium textures locally, which defeats the purpose of having the skin in the first place.

>>15190929

There is no foolproof way on a user-controlled piece of hardware. Which is why everyone is hard at work subverting these systems and demanding corporate/copyright holder control of your OS & processor. It sounds like you want to be a console or iOS developer. For a limited time, the PC is still mostly free.

>>15190929

The only way to ensure this is through cloud gaming. You render the entire game's image on your computer and just stream it to them. It's as unhackable as it gets, even client-side.

>>15190589

I have a thread for it here: https://8ch.net/agdg/res/31318.html

I have finished reinstalling windows and I also now have a partition for Slackware, so at least once I get these operating systems set up correctly, I can add vulkan support to the Linux version and have an easier time developing it. Eventually I want the Linux version to have the same features, it just takes more effort than windows does.

>>15191584

How do you render text? Do you just use a spritefont?

>>15192667

Why though?

>>15192677

Because I want to make money off of Neofags, clearly.

>>15187466

Loving that second one

Any musicfags know what technique or instrument this is?

I'm guessing it's just a tremolo but I'm not sure.

https://youtu.be/mHFOLqWgpEY?t=97

>>15193946

Here's another example

Saging for doublepost

>>15193958

Fuck

https://youtu.be/HgQ6Cttubto?t=17

While I work on my current game, I am trying to think about mechanics for another one.

In an RPG, I want it to be a dungeon crawler, with an ATB-style combat system, with each skill having different wind-up and recovery costs. I just can't think of how to approach the class/guild/skill system. I know I want each monster to give a fixed amount of XP, but beyond that I'm not sure what else. You should only be able to participate in one guild (job) at a time, and use what you learn there with other abilities to make a build. You should be able to freely change abilities around, but I don't want to have characters be a snowball of everything, but at the same time I want to encourage experimentation

>>15193946

>>15193961

>Instrument

I'm guessing you're referring to the cello/double bass? It's pretty difficult to get a convincing solo cello sound without an actual cellist.

>>15194010

It's the plucked instrument that uses tremolo.

>>15194004

>You should be able to freely change abilities around, but I don't want to have characters be a snowball of everything, but at the same time I want to encourage experimentation

I would imagine the simplest way to handle that would be to give each skill from every class a point value, and give characters a maximum value of points for skills that they can have equipped. Like, if the Thief class has a Dodge Up ability, it can cost 2 points out of a character's 15 total skill points. He has 13 left to spend on whatever skills he wants to equip from whatever other class.

Or you can go the FFT way and just have different types of skills (e.g. active skills, movement skills, reaction skills, etc.) and characters can only equip so many of each type.

>>15194154

>Ability points that are consumed while equipped

Oh hey that just might work. I'll have to consider the exact details ,but that gives me a very good base to work from

>>15194140

I'm guessing a zither, particularly for your second example. There are a lot of plucked-string instruments that can produce similar sounds to the one in the first example though (guzheng, koto).

>>15194929

>Schwerer Gustav

nice

>>15195754

>inb4 the whole thing IS the boss

>you gotta blow it up before it reaches the german base where you will get shot down from all sides, or defend it if you play as german

There's too many games where the goal is to stop the train entirely with units, or if you're on it, you just get to the control room and shut it down. That's boring.

Imagine like FF8, where the goal was to swap out a dummy set of cars in order to kidnap a president while the trains are in motion, only to find out it was a body double. Or what about setting a bomb in the cabin, then escaping to the rear train and releasing it while on it, so nothing looks amiss?

>>15195956

nvm figured out the issue. The actual window drawing area isn't 1920x1080.

>>15196013

someone was doing a snow/sandboarding one in godot

>>15195956

>>15195972

Is the problem that, when you try and make a window that is 1920x1080 on a monitor that is 1920x1080, the actual client area of the window is less than 1920x1080? That shouldn't be happening on windows at least, what platform is this on? I was looking at the source code to see if there was a bug in SFML, but it seems to not have this problem on windows if you look at the code here:

https://github.com/SFML/SFML/blob/ee957f342051f509634355755049ad0486dafd43/src/SFML/Window/Win32/WindowImplWin32.cpp

>>15196024

It's 16 days till demo day, bud.

>>15196042

Antergos (aka Arch Linux with nigger-friendly installation).

>>15196002

A game where you are a repair mechanic/surgeon for robot women, and while they're on the table they complain about not being loved at work or home, and you pick which one you want to bond with

•a cafe maid bot that just needs minor maintenance

•construction bot girl that needs regular limb replacement from heavy stress jobs

•spec ops Ghost-in-the-Shell type bot girl that comes to you for full body swaps and holes blown through her

•a sales assistant bot that has bad luck and gets damage way beyond normal chance and is upbeat to at least still be online/functioning despite her luck

the actual repair mechanic need only be minimal drag-and-drop shit, and waiting for parts to be delivered by drone, so you talk while you wait. This shit writes itself.

>>15196449

What language?

As far as I know, it should be a contiguous block of memory, for value types

>>15196449

>is new guaranteed to return a contiguous block of memory

I don't see how it could be, or why. It's an OS detail.

>>15196002

>what do i have to do to make this print money?

add fetishes

>>15196449

From your programs point of view, it will be a contiguous block of (virtual) memory.

There's no guarantee that it is actually contiguous in physical memory.

The first few bytes might be at the end of one memory page, and the next few on a different one, mapped to contiguous virtual memory, but in different physical memory locations.

As long as you're not writing kernel drivers this is completely irrelevant to you and you're safe assuming everything is contiguous.

>>15196551

Then why do people always argue that game devs allocate all memory up front so its contiguous?

>>15196584

My understanding is that when you allocate an object, the system attempts to find a chunk of memory that can hold the requested object. Whether it finds the smallest or nearest memory is an implementation detail by the system. What matters is that your int32's get 32 bits, etc etc.

References don't need to be contiguous, so eg a linked list can be all over the place in random spots of memory, as long as each node is contiguous to itself. If you allocate an array, then it's all in contiguous memory, large enough to fit all the values of the array. In some cases, this is a huge list of reference pointers that go to fuck knows where, but if it's a byte array, it should all be there, compact data.

At least in C#. While it's good to care about such things for performance, it's probably not something to worry about in an amateur thread before you have a finished game

>>15196584

OSes tend to allocate memory in blocks of a few kb at least. Even if you get a bunch of blocks all over physical ram you're still better off than completely random allocation.

>>15196609

Oh i was talking about C++ not C#.

People here often claim professional devs allocate a giant block of memory at the start so all of it is contiguous. However that makes no sense if new doesn't guarantee this.

So are the people that claim this just full of shit? I mean after all anons wouldn't just lie in the /agdg/ thread right?

>>15196634

Those professional devs also happen to be working on consoles, anon. And it CAN be done on PC/Windows, just not as easily as calling new()

https://docs.microsoft.com/en-us/windows-hardware/drivers/ddi/content/wdm/nf-wdm-mmallocatecontiguousmemory

>>15196584

>>15196634

That's not to be contiguous, but to stop memory fragmentation. That's a different issue entirely.

Also, "professional" devs are misnomer in the video game industry. We still have retards getting their offices broken into and being set back months of work for fuck sake.

>>15196765

Isn't fragmented memory and non-contiguous memory the same?

>>15196102

You only get C&D from western characters. Japan welcomes it. Its why they have such a big doujin scene.

>>15196794

No.

Memory fragmentation in programming refers to memory being used inefficiently.

Due to different sizes in allocations, "holes" of unused memory might appear in which no allocation fits. This is essentially wasted memory.

For example, let's say you allocate objects A, B, C and D. A, C and D are 100kb, B is 99kb. These are allocated independently and in order, putting them next to each other (A, B, C, D, 399kb allocated), occupying 399kb of memory.

Then, B is a freed. This leaves a hole of 99kb. (A, 99kb unused, C, D, 399kb allocated)

Then we decide to allocate E, which also is 100kb. But this leaves us with a problem, E doesn't fit in the place of B! So we have to allocate it after D. (A, 99kb unused, C, D, E, 499kb allocated)

The end result is that we have a program that consumes 499kb of memory, despite technically only needing 400kb.

Now, instead of talking about objects that are 99kb and 100kb, I want you to think about several-megabyte-large-textures and objects that are a mere hundred bytes. If you allocate and deallocate tons of objects, you'll end up with objects all over your allocated memory, rendering most of it unusable for large objects. This is memory fragmentation.

For preventing memory fragmentation, simply reducing allocations and deallocations is the trick. Allocating all the big stuff at program startup, and only deallocating at shut down (which often means "just exit() and let the OS handle it lmao") is a very easy way to accomplish this.

Note that all of this isn't really an issue anymore in the modern day since people no longer have potato computers with shit for memory. On top of that, C# and Java actually move objects around in memory to reduce fragmentation.

>>15196095

It makes sense that you're having issues, Linux window managers are way more messy to work with. Implementations of X don't work in the same exact way, it's a pain to do it. If you read the source code to create the X11 window, it is full of comments that talk about it in terms of "some WM's do this" and "most WM's do this" or "a reasonable WM will probably do this". With a ton of defensive programming and lots of fallbacks when something doesn't work.

https://github.com/SFML/SFML/blob/ee957f342051f509634355755049ad0486dafd43/src/SFML/Window/Unix/WindowImplX11.cpp

>>15196584

>Then why do people always argue that game devs allocate all memory up front so its contiguous?

We do it because malloc(), free(), new, delete, new[], and delete[] are all expensive operations and can not be execute per-frame if you want to maintain a reasonable, stutter-free game or simulation. The reason is because these routines call into your OS, which is 99.999999% chance not an RTOS. Making sure it's continuous is a nice benefit, and extremely easy to achieve. My hint: look at standard containers that guarantee continuous memory. std::vector and std::array if you need more of a hint.

>>15197147

>allocating in a hot loop

>not smart

amazing insight

>>15197161

Who hasn't allocated a few streams in your moms hot loop at this point?

>>15197992

Good job anon

>>15196809

>You only get C&D from western characters. Japan welcomes it. Its why they have such a big doujin scene.

>Nintendo

>>15199566

That sounds like Nintendo of America Anon. Same jew.

>>15199587

Make a game like missile command but instead of stopping missiles you have to stop migrant boats.

>>15200168

this brings back fond memories of creating Duke Nukem 3D levels with the included editor on PC.

Good luck with level design friendo, I don't envy the hair-pulling you're in for in making compelling levels people actually want to play. But nice engine I guess!

>>15200438

The fuck is wrong with that vagina? Is she post op?

>>15201074

i don't know fam, i didn't make the model

>>15201502

>i didn't make the model

Neither did Blizzard if it has a vag. You can't fault them for shitty bones in an SFM rig they didn't make.

>>15200438

Eat that pickle you dirty whore.

>>15201797

>Vagina bones

Makes me laugh every time

>>15202433

What's so funny,this is clearly a real thing.

>>15202453

I know, but it's trying to imagine an actual vagina as having bones in it

>>15202459

Are you telling me they dont?

Next you're gonna tell me there's no bones in boners

This is the EBNF defining the custom language for item files in the text adventure. This will eventually take the form of a recursive descent parser.

block = ["VAR", list],
        (statement, lineend),
        {block};

statement = (list | external), ("="|"+="|"-="|"*="|"\="), expression |
          | text |
          | ifstatement |
          | whilestatement;

text = '"' {newline | '\"' | character} '"';

print = "<<" (ident | external)

newline = "\n";

ifstatement = "if", "(", condition, ")", container,
              [ "else", [ifstatement], container];

whilestatement = "while", "(", condition, ")", container;

container = "{", statement, "}";

condition = ident, ("=="|"<"|"<="|">"|">="), expression, {("&&", "||"), condition};

expression = term, {("+"|"-"), term};

term = factor, {("*"|"/"), factor};

factor = ident |
       | external |
       | number |
       | text |
       | "(", expression, ")";

external = ident, ".", ident;

list = ident, {"," ident};

Certain things to note.

>This is the parser for the "adventure mode" of the text adventure. Dialog and combat are two separate modes that won't parse item files.

>Variables are weakly typed

>There are two types of variables, item variables and command variables. Item variables are defined in the item itself (that means it's not put through the parser) and command variables are defined in the item command (that means they are put through the parser)

>command variables exist only for the sake of that command, and don't persist after the command is finished.

>To print some text on screen, you just have to put it between quotation marks.

>To print a variable on screen you put it after a "<<"

>There are if statements and while loops

>There's math up to addition, subtraction, multiplication, and division

>Variables in external objects can be read and edited

>There are no functions, classes, objects, etc.

>ident is an identifier, such as a variable name

>character is any character

>number is a number

>lineend is the end of a line

Are there any other text adventure specific features you would like to see in this? Any improvements I could make to the syntax? I'm still questioning whether or not I should keep the variables weakly typed.

>>15204181

>the last thing we want is to impact your development in any way

>I'd like to help you upgrade to Unity Plus or Pro in order to be compliant

Mafias in the game dev scene.

We live in the worse timeline.

>>15200168

It looks really good.

Do you have any books/ressources to share on how to make a build engine type of engine ?

Have you ever thought about forking already existing open source video game and making your game based on it?

>>15201846

more like egg

>>15204181

Just another reason why OSS engines like idtech3 are nicer…

>>15204882

Here I am, I have some more links too, but a lot of them are about general graphics programming background that you need to start working on an engine. So here's some stuff-:

The first part of the BSP tutorial, it explains the compiler more:

https://www.cs.utah.edu/~jsnider/SeniorProj/BSP1/default.html

This video is mostly about theory, but he does implement a Duke3D rendering engine on screen:

https://www.youtube.com/watch?v=HQYsFshbkYw

This topic is a little different from how Infernis anon is doing his engine- he is writing a sector-based map editor (I think? correct me if I am wrong) which means that he is placing his walls manually and it can divide itself into convex sectors as he places them- while I am writing a brush based map editor; where I create my map out of convex solid shapes, right now it's only blocks but it could be another shape later. So to avoid intersections between brushes, which makes BSP compilation impossible for me, I clip all my brushes against each other in a CSG step which can be explained in this tutorial:

http://www.alsprogrammingresource.com/bsptut.zip

This isn't needed unless you want to write a brush based map editor. So maybe for a build style engine you don't need it.

Besides this since all of the tutorials here assume you already know how to draw polygons and just tells you how to reason about them, these are some basic graphics programming tutorials if you need them:

http://winprog.org/tutorial/

https://www.khronos.org/opengl/wiki/Programming_OpenGL_in_Linux:_GLX_and_Xlib

http://nehe.gamedev.net/

http://www.glprogramming.com/red/

https://www.khronos.org/registry/OpenGL/specs/gl/glspec10.pdf

Maybe a lot of people have a lot of different opinions about how to draw graphics, this is one way but there are a lot of others that you could use.

>>15204181

>random accounts

Almost certainly not, only the ones that actually made around that much money from their game

>blackmail users into buying premium licenses

How the fuck is it blackmail if they tell you from the start that you have to get a premium license if you earn >$100,000 in a year?

This is like showing an email from Epic when they send you a mail telling you to send them 5% of their profits and claiming it's wallet-rape. They're very fucking explicit about it, no need to bitch about it if you're too lazy to make your own engine or use a free "less useful" one.

>>15203329

>This is the EBNF defining the custom language for item files in the text adventure. This will eventually take the form of a recursive descent parser.

You like the theoretical part more than making games, don't you?

>>15205988

>There is no guarantee that allocating memory returns a solid block of physical memory.

What does this even mean?

>>15206008

When you have a pointer to memory in your program, this is actually a virtual memory address that gets translated to a physical address using the MMU in your CPU. So this allows each process to have its own address space. In theory if you have a block of memory from malloc(), even though you are addressing it with a continuous series of addresses, these are all virtual addresses and might not be continuous physical addresses. For example if I allocate a block of memory that is two pages big, the first page and second page might not actually be right next to each other on my RAM, meaning that there is still memory fragmentation happening, with no way of being aware of it as long as malloc() or other higher level memory management systems are being used.

>>15206020

Imagine being this schizophrenic.

>>15206045

Thanks

>>15205988

Install gentoo

>>15186873

>Amateur

Your lack of ambition disgusts me.

>>15206102

Most people who go pro leave so it's fairly accurate

Swimming in videogames, how do I do it?

I've been thinking of mesh/box colliders (my water have no waves) with exclusion colliders for air tunnels and whatnot. If it's inside a water collider, but not inside a exclusion collider, it's swimming.

Sounds fine and all, but i'd rather research a big more on the topic but I can't seem to find anything about this shit.

You can easily find dozens of articles for AI, Pathfind, shaders and whatnot, but I can't seem to find anything about how swimming is done.

Anyone have a good resource for me to read about the subject?

>>15206055

It's important to keep in mind that this isn't an absolute and there are lower level memory management tools that can guarantee that there will not be any physical memory fragmentation, in kernel mode this is easier but in user mode this is still probably possible, it just requires a good understanding of the operating system you are using. So it isn't an impossibility, it's just very difficult.

>>15205973

I do really like the theoretical part but in my defense it's easier to change EBNF than it is to change a finished parser, and writing everything out in EBNF keeps me organized when I actually start writing the parser.

Also I can't get around the fact that I need some way in the game files to tell the game what to do, and the best way to tell a computer what to do is though a programming language.

This game is a learning experience for me. I've finished clones of Atari games in the past.

>>15205968

Read the article dipshit.

>>15206604

It also doesn't really matter if all your memory is physically contiguous. The cache loads memory aligned to it's own width every time it has to get something from RAM that isn't already in the cache, so as long as you allocate your variable arrays or large memory blocks with considerably easier to use methods to align it to those addresses you'll have all the cache optimization possible even without writing your own allocators.

>>15206102

Has nothing to do with ambition, you're only not an amateur at something if you make a living with it. Plus the word come from amare, meaning you do it because you love to.

>>15209589

Had no reason to believe that archive would be an archive and not just an archive of a tweet containing the image. Reading the actual article makes it even more of a nothingburger though, an undivulged (so probably low but intentionally vague to stir outrage and thus get more views) number of developers get a message based on an error that has already been apologized for. Big whoop.

>>15209671

>would be an archive

*Would be an article

>>15209727

World coordinate triplanar texture mapping?

>>15209778

i guess

originally i just unpacked the texture with world pos, but that got fucked on horizontally flat surfaces

then i found this to fix that

		float3 worldCol(Input IN,sampler2D tex,fixed scale)
		{
			float3 Pos = IN.worldPos / (-1.0);
            float3 c00 = tex2D (tex, IN.worldPos / 10);
 
            float3 c1 = tex2D ( tex, Pos.yz *scale ).rgb;
            float3 c2 = tex2D ( tex, Pos.xz *scale).rgb;
            float3 c3 = tex2D ( tex, Pos.xy *scale).rgb;
 
            float alpha21 = abs (IN.worldNormal.x);
            float alpha23 = abs (IN.worldNormal.z);
 
            float3 c21 = lerp ( c2, c1, alpha21 ).rgb;
            float3 c23 = lerp ( c21, c3, alpha23 ).rgb;
			return c23;
		}

the rest is just blending different textures depending on the vertex normal y coord

there's also other textures for random details but i'm not using them in the examples

the important detail is that you don't have to worry about UVs, meshes with the same material will always sort of blend

the reason for the shader was that our levels have a width/length limit, if a mesh goes over that it has to be split, and managing UVs there to match would be a pain

my artist basically wanted me to copy vid related, at 4:30

>>15199452

That's really nice. What did you use to build the collision geometry?

>>15205599

nice progress. unified coordinates are nice. even for ortho projections, using [-1,1] for view coordinates solves so many problems instead of using pixel coordinates.

>>15206045

Holly shit you're right. Some days I miss asm programming on amiga and st.

>>15209988

>third one

that's cool as fuck

>>15209988

The doom one has no business being among those.

>>15210416

It looks shitty enough.

>>15210455

I see no 1px limbs, the movement is clear and expressive with personality, and so is the face. Sure, it's an edit, but I don't think it deserves to be called pixelshit.

>>15209988

I swear to god I saw those exact 5 images posted calling them pixelshit over a year ago. And exactly like last time, people call out the doom guy to not be pixelshit

>>15210552

I remember that too. Just go with the rehash flow, man.

>>15211168

That's pretty shameful but that's not what a botnet means. And its pretty standard for a site to send shifty goddawful messages to keep using their service.

>>15212475

I also still think software is really shitty these days. You need better and better hardware just to render text and images every season. It's because of the attitude that hardware is so good now, that less time and money should be spent on making software efficient. I believe this is only offloading development expenses on the user's hardware budget.

However, the opposite extreme is no good. Ultimately, the gamer wants to play your game, and most users can't appreciate the effort you went through to add a few more FPS. If you know how you can add that FPS without delaying the game; go for it. There's no silver bullet though, so you can't always know for sure that an optimization will help much. Just look at the arguments in this thread. You're already wasting your time before you know you need it.

Look at Bob Ross: do you see him spending the whole episode painting one corner of his canvas? Like the name of the show: instead of discovering the Joy of Painting, you should discover the joy of programming.

>>15212475

Engineering as a whole is "fuck it, good enough". True optimization minimizes time and effort for practical results.

>>15213224

>True optimization minimizes time and effort for practical results.

The problem with this mindset is that it's frequently abused by shitters as an excuse to write poorly-performing code.

>decide I'm finally going to make that text adventure

>open nano

>what the fuck am I doing

This has happened multiple times. Why did I have to be a right-brained faggot?

>>15213318

That's fair, but I'm talking about the good yesdevs here: don't let your perfectionism get in the way.

>>15213448

As someone who is making a text adventure, here's the basic structure of my game.

1. Accept user input

2. Parse user input, figure out what is the verb and what is the direct object/noun

3. Read and parse the game files into data structures

4. Find the noun in the data structures

5. Find the verb corresponding to that noun

6. Execute the command corresponding to that verb

7. Go to step 1

There are a number of little improvements you can make to the above to make your code more efficient. For example you probably don't want to parse the relevant game files after every command. Instead you should parse things like the player file at the beginning of the game and the room/item files when you enter a room with those specific items. But those aren't options until you have a player file or a way to travel between rooms.

>>15213677

Oh no, you misunderstood. I have no clue how to use the editor itself. And I probably never will, given how fucking retarded I actually am.

>>15213448

>Why did I have to be a right-brained faggot

You're probably uniformly retarded, the left/right brain bullshit is mostly pushed so people can pretend that a deficiency in one area means they're naturally good at another.

>>15213784

I actually am "right brained" though. I'm more interested in writing than mathematics, a fact which I loathe.

>>15213802

Math is more rigorous than writing. You're overthinking it.

>>15213806

That rigor gives it value. Writing has no intrinsic worth whatsoever, it's all sentimental.

is there any shader for unity that makes UI text change color to be the opposite of whatever is rendered behind the text? i need my shit to be readable

>>15213996

You are surprisingly industrious.

>>15214005

You mean an inverse color blend?

>>15214005

>i need my shit to be readable

White text with a black outline or shadow.

Translucent to opaque black/dark grey rectangle with white or otherwise light colored text over that.

That's what they do for subtitles.

>>15214042

yea, i just found the outline, didn't know there was a built in one

i still think the inverse color would have worked better though

>>15214005

No. You are gonna have to make a shader for it. Just grab the screen buffer and invert the color you write.

>>15214031

Thanks. Strange choice of word.

>>15214005

>>15214059

>opposite of whatever is rendered behind the text

Not a good idea. It may look like a clusterfuck depending on the background. There's also a good chance that it'll become completely invisible if the background is 50% gray because it doesn't have an inverse color.

Using inverted colors are ok for shapes and outlines that just need to be visible, but it's not good for shapes that need to be clearly recognizable.

>>15214065

Forgive me, I'd expect most porn game devs to abandon their projects, yet you keep making progress.

If I do "#define something 100*(4+2)", does something then equal "100*(4+2)" or "600"? I'm wondering if I should put that number into a constant so the calculation doesn't need to be done many times.

>>15214440

it should not make a calculation during runtime

>>15214491

>download videos from hooktube

use genyoutube

>>15214517

>there will never be a fully controllable lewd game

Sure there will, it's just a matter of time.

>>15214517

>that hint of leg wrapping around the man

>>15214096

Good job, but oh boy, I can hear the neckbeards on /tg/ screaming in incoherent rage.

Tau looks great, I really like their mech designs.

>>15214714

Never said we'd live to see it. :^)

>>15214517

>fully controllable lewd game

Fully controllable how? There's a lot of ways you may want to move each bodypart, I wonder if it's even possible to make something like that before we have full body controllers or something.

>>15215247

like super deepthroat but more

at the very least 3d

Quick question for Software Developers in /agdg/.

I'm looking for jobs in IT, a lot of them ask for C# .Net experience. I know C#, but I've really ever only used it inside Unity. How easy is it to transfer knowledge between the two? Should I apply for these jobs?

>>15216682

Find some .NET C# projects on github and see if you understand what is going on in those projects.

>>15216962

Nice to see you back anon.

>I'm getting closer to a point where all I have to do is print maps and content out.

That must be a great feeling. Looks like it's coming along nicely.

>>15186873

Would you guys be interested in making Sanic Stages in Unity? I was thinking of starting a thread where we'd basically make Sonic Stages and then share them Mario Maker style.

>>15217388

Are you going to have a demo on 8/8? Looks interesting.

>>15217397

Don't give him any funny ideas, not until the camera and combat system are in a better state.

>>15218636

Once you have a few years of programming experience under your belt, you stop feeling.

>>15218636

I don't care at all.

What part of "amateur" did you not understand?

>>15218896

>Change degree to focus on more math

I know that fucking feel. I thought it'd do me good to get a good math background before "doing more fun stuff".

>>15218958

You do the fulltime one a while until you buffer enough of that way-too-much-money, then quit and do your passion fulltime.

Although fulltime jobs still leave time to work on your own stuff, but only if you're one of those people who doesn't do much aside from sit behind you PC in his spare time, and not lazy enough to just sit on 8chan all day when you do.

>>15218968

He will be unemployable if he's out of work for like half a year after his first job. Even if he did get an interview, this would be a guaranteed question and he'd have to come up with a legit reason. If he said he was on a hiatus because of gamedev, he'd get laughed out of the room. Also, having money is addictive. He probably won't want to go back to having no money and being anxious whether he will have overspent by ordering something that's not directly tied to his survival.

>>15218982

This is essentially my fear

>>15218988

Find a job that is 3/4 full time instead. Either that or dedicate a day of the week, like friday, saturday or sunday for your gamedev. Do not quit for a long time after your first job, this raises a lot of red flags especially in a field that is fast becoming oversaturated. It just shows you cannot cope with pressure/overworking/responsibilities to the employer. Quitting after a year and getting one month off before looking for another job is also kinda iffy, but more acceptable but you have to convince yourself that you wanted to take on new challenges in IT and say so in the interviews for your second job. I think you start getting proper vacation days after roughly a year of employment as well.

>>15218958

>get full time job

>get buffer money

>switch to part time

>semi-retire until game is just like made

>>15219000

>Find a job that is 3/4

I don't think that exists anon

>>15219005

>>15219000

>>15218988

>>15218982

>>15218968

>>15218958

how about you just do whatever work you can find and gamedev when you aren't starving on the streets

>>15219020

They are very rare, but I've seen jobs that require you to work about 30h a week with variations ie. 32h, 33h.

>>15211494

No. (((Unity))) literally tracks your usage and sets up an RDP server so they can have direct desktop control.

>>15219025

Where I come from that's just called a part time job that asks too much of you.

>>15218636

Pretty good. I'm working to build my AI and machine learning skills. Current employer is even paying for classes.

>>15218982

>If he said he was on a hiatus because of gamedev, he'd get laughed out of the room.

Depends on how you say it, and if you can speak with confidence.

You can word it in a way as to show that you're an industrious, independent, risk-taking, reliable, and hard working person.

Or, you can word it, "yah and i like took a break bcs i love bidya gaymes and i failed so i'm here for a jerb", and seem like an idiot for doing so, and won't be staying at the company long.

It really depends on how you say it, and what perspective you allow them to take on your own experience; as you're the one conveying your experiences, and can word in anyway you want to give off any signals you want.

Interviews are easy as fuck along as you have a way with words, and can read the interviewer's body language.

>Also, having money is addictive.

If you have a lot of money, but no time to spend it; what's the fucking point?

Anyone who says they "love money" is also someone who, in most cases, doesn't have a lot of money, or is a hedonistic fool with a low IQ (or at least from the people I've talked to, who have that world view of, "green is life nigga") or is just your average hedonistic woman/feminized man.

Anyone who only actively works for money is also someone who isn't forward thinking.

Invest. Build streams of passive income into a river of passive income, and live off that.

>He probably won't want to go back to having no money

Thus the last points.

>>15219000

>Do not quit for a long time after your first job, this raises a lot of red flags

truth

>fast becoming oversaturated

If you're willing to relocate, and live in a large country such as the US; then it's not a problem.

>>15216682

Usually if you're going to school with a specific focus (web dev w/javascript, python or backend w/java, C#, php, etc) then they should have term projects using the popular APIs as to prepare you for jobs.

If not then those teachers have probably never had a traditional programming job, and went straight from getting their Phd/masters to becoming a professor.

The general C# usage you get from unity is pretty transferable in terms of the "general knowledge".

Though, not being familiar with the specific APIs used in the field can be a hindrance; in addition to the specific design patterns that are common.

So, they will probably ask you what you like from X/Y/Z APIs, projects you've worked on/contributed to, and to give them a brief overview of W design pattern that they utilize (in a real world scenario, or ask you about a specific part to test your knowledge).

Also you don't need a github, if you say something like, "I intend to make profits from my work, and do not publish my work online".

This is common enough to not raise any eyebrows, and as long as you have samples of your work at hand (if they request it); then you're fine. Though it's generally easier to have a github w/samples of your work.

>>15219031

>sets up an RDP server so they can have direct desktop control

Bullshit. Proof or gtfo.

>tfw discovered Elona and suddenly my game won't be ready for demo day

>>15219025

i'm not familiar with developing/programming at all, but in many jobs in the US they'll have employees work 34 hours or less a week, in order to classify them as part-time and thus avoid needing to pay the same benefits as full-time employees.

Is it worth it? Is it worth it to focus all this time and effort on graphics programming when something like security has so much more potential to give one a better job?

has anyone tried doing modding in unity, particularly with new code?

looking at this, https://github.com/Hello-Meow/ModTool

is it a decent option or is there better?

Why are ground textures so bad?

Is this logic correct for how engines handle motion of multiple distinct objects:

>Vertex data is loaded for each model into a buffer (or multiple but for this example just one buffer to make thingssimpler)

>Essentially, a uniform buffer is created for each object and translations/transformations are made to each uniform buffer's MVP matrix every update which positions the object correctly in the world.

How does collision work then?

>>15223793

This is how I am doing it in Vulkan. In OpenGL I just use the matrix stack to do basically the same thing.

Collision and rendering are different things. I use a BSP tree for collision detection with the world, but there are a lot of ways to do collision detection. If I wanted to have a moving box the player can collide with, I would just translate the coordinates on the CPU with a matrix as well as the GPU.

>>15223817

Im using Vulkan too.

That's precisely what I'm worried about though. I don't want to calculate all updates on the CPU for collision only to repeat this process in the vertex shader.

There has to be some way to use like a Compute shader to calculate the positions of all objects then it outputs a list of all indices for all objects that collided to the CPU for collision handling. Wouldn't this be quicker? I'm not familiar enough with shaders to know if this is preferable or if it's even possible to implement continuous collision detection…

>>15223836

Also, please tell me if I'm thinking too much about pre-optimizing.

>>15223758

Have you thought about adding an animation for when you character goes in another direction on the ice like vid related at 0:37 instead of keeping the normal running animation ?

>>15223836

I calculate all of my matrices on the CPU already to set the uniforms- the vertex shader just applies this matrix. Maybe that's not as fast as it could be… I don't have a problem in repeating some effort on the CPU- it seems that in the time you can do these calculations in a compute shader, it would be slower to do it that way unless you needed to do collision detection against a very large set. I think it would be faster to just do the calculations on the CPU instead of waiting for all your vertex data to go up and down the PCI bus.

It's possible for me to do fast collision detection with a dynamic object like this, because I can use the BSP tree to reject objects that the player couldn't be touching. So, if I am in leaf 5, and the moving box is in leaf 6, no collision check needs to be performed. I just have to walk the BSP tree a few times to reject it. I might actually not do this because the collision test itself is probably faster than proving I can't collide with it, though.

I don't think that you're thinking too much about pre-optimization, you want to do it correctly the first time or at least set it up to be done correctly. Maybe the issue is, you have thought about how to render objects a lot, but not yet how to do collision detection. You need to look into collision detection algorithms and data structures and think about which one you are going to use in your engine. I think that collision detection is a much more difficult problem than rendering stuff.

>>15223917

The next step for my engine is to simulate as many thousands of spheres bouncing and interacting with one another as possible. I want my engine to be very focused on efficient physics simulation.

The tranformation matrices are calculated on the CPU but I'm thinking about collision which requires the updated position of each object which you do once for collision then repeat the same process in the vertex shader for objects that did not collide. I'm just spitballing about how that seems redundant.

You are right about potential problems with bandwidth (is that the correct term here?). Perhaps though, in a scene filled with simple spheres, all that would have to be passed is a single vector for each sphere representing the centre as well as a radius. Then collision detection is hugely efficient and could possibly make a compute shader worth it? Maybe I could even reuse the same vertex buffer data then for every object…

I'll just try it on the CPU first. The compute shader does seem like something that could be implemented later, same with the BSP trees.

>>15223942

>Are you the anon working on the particle interaction thing

Yes.

>>15223793

My experience is with OpenGL and Bullet3D.

You create separate Bullet3D collision objects. They really should be convex. You can use vertex data from the actual render object and build a collision mesh object, or just use a capsule, sphere, box or whatever as an approximation of the render object.

Each frame Bullet3D may move the physics object (if it's a rigid body object and not just a static collision object). It's your job to make sure the graphics object 3D transform is updated to track the physics object. This can get pretty tricky if you are using multiple frames and substepping in the physics simulation, and if your physics world and render loop are not synced in lock-step (multithreaded - in different threads).

So yeah, the render mesh and transform is not used in the physics simulation directly. It has it's own copy, and the mesh is usually a far simpler mesh. just an approximation of the shape of the actual graphic object. The transforms have to be synced each frame. Bullet at least provides a simple mechanism to do this - fill in your update code in the btWorldTransform class function.

You don't have to update each object each frame as most objects are not moving and Bullet will "put them to sleep" after not moving for a short period of time.

Bullet also has a version that runs on the GPU with OpenCL. I haven't tried it though and I don't know of the graphics and physics transforms are handled exclusively on the GPU and avoids the CPU memory bus.

I think in your case for max sphere count, doing direct sphere physics on the GPU would be the way to go.

All people saying entity component systems are good.

Is this true?

>>15225773

I've heard it's all the rage in gamedev.

>>15225871

>22nd century communist canada

>23rd century fascist socialist britain

>some complicated thing that's difficult to discern

I decided to take a little break from writing the parser and work on the story and setting. I would appreciate any thoughts and critiques on this.

Humanity creates a state of the art space station that is capable of generating it's own gravity. The space station is huge and meant to be able to go long periods of time without re-stocking, at most 5 years. The station is put into orbit around the sun between earth and mars. After not even 3 months in service the space station vanishes. It doesn't respond to radio communication, and when people look where it should be it's not there. The gravity tech is deemed a failure and shelved.

One year later an alien scout ship arrives to earth. Humanity makes first contact successfully and establishes friendly foreign relations with the aliens.

Three years a later an alien capital ship arrives. They bring technology and a greater understanding of the universe. Humanity, when asked to show off their most advanced science brings forward the old gravity tech among fusion reactors in their infancy, silicon and quantum computing, and so on. The alien scientists see the gravity tech and go "what the hell?" because apparently we had stumbled upon some freaky stuff. They say that it's too dangerous to use, we go "yeah we figured that out", and the tech is shelved again.

The aliens share their tech and knowledge, including their own safer, more compact, and more efficient gravity tech. There's a boom in humanities development. Humanity starts exploring mars, and eventually the stars. Space travel becomes mundane. Energy becomes ridiculously cheep.

Ten years later a human space trucker is drifting out past Pluto when he seems something on his instruments. It has a really bizarre and noisy signal, but when he draws closer there's a flash of light from the object and moments later a piece of his ship breaks off. He looses control and starts falling towards the object. As he draws closer he sees that part of the object resembles an alien ship, and the other part of the object resembles what humanities space tech looked like before the aliens arrived. He eventually collides with the ship and finds out that his ship is too damaged to be able to take off, and the radio signal emanating from the object is too strong and would drown out any signal he may try to send.

Now he has to wonder the space station and try to find the parts and tools to fix his ship.

>>15225871

Here's a sight where you can see what the flag looks like in the wind.

https://krikienoid.github.io/flagwaver/

>>15225871

The animals on the flags would be difficult to recognize at a distance.

>>15218982

>He will be unemployable if he's out of work for like half a year after his first job. Even if he did get an interview, this would be a guaranteed question and he'd have to come up with a legit reason.

Nah. The whole process is just a simple funnel to make sure the HR monkeys don't hire retards. The cover letter is a quick litmus test to make sure you are in line with what they expect, and an invitation for them to look at your resume. The resume makes sure you have the proper skills, and the interview makes sure you have the proper personality.

If you can get that far, it just becomes a matter of not dropping your spaghetti. If you worked your first fulltime job, stick with it for at least two years so it doesn't look like you're a job-hopper. It's all how you phrase it and present yoursefl

>>15225943

The second flag is your best, IMO. The black bars in particular evoke some futuristic-like fascism, like a dreadful imposing order. The angle makes them look futuristic. The gold streaks land the communism connection. I'm not sure about the bird though - it screams of freedom, flight.

True that coat of arms can be complicated, but I'd suggest keeping you flags simple for the sake of clear identification.

In certain situations, I may intentionally give multiple volunteers the same task, in anticipation that one of the volunteers will fail to complete the task.

You must accept the possibility that your work may be discarded in favor of superior work.

yandere-dev yes

>>15226034

I'm having a few issues with my puzzle game at the moment. Mostly I haven't worked on it in a bit because I don't know how to approach it. Main problems are:

>Need to consistently update all actors in reverse order (from bottom right to top left), ideally maintain actors in sorted array(?) and update their positions as needed

>Maintain a list of which tiles are occupied by an actor, bool[] seems to be ideal here, rather than maintaining a List or Hash of <Position>

>Somehow actors need to be able to define an interaction, eg a "rock" on "water" will sink, but a "boat" would not, ideally without introducing state flags like IS_FALLING

>Somehow define common behavior that objects can reuse (eg rocks, keys, gems all fall)

>>15224178

Lookin good.

>>15212449

you should just post

[AUTISM INTENSIFIES]

this freak discuss about nonsense and endup becoming trannies doing suicide onlive

hipsters are making games while you autist discuss how to optimeze the nothingness, you don't have completed a game, nont a single one, yet you pour countless breads on shit talking, so noobs coming here think it is important, it is not

by the way i've already delivered 4 games, but not alone, as a part of a team

keep talking future trannies

>>15226034

Yeah I thought so. I'll have to look into it.

>>15218982

I took a 6 month hiatus after saving up money from my first job. The question did come up at my next interview and I just said I was exploring avenues of self-employment but it didn't work out.

Got the job.

>>15225904

Sounds similar to the plot of Event Horizon, great movie I assume you have seen it. It would make a super interesting game in the vein of Dead Space.

>>15227706

If you have a single missing triangle, then the potential causes are:

>You are importing 1 to 3 vertices too little

>You are importing 1 to 3 indices too little

Incomplete triangles aren't drawn, so it really could be a "+1" or "-1" kind of problem.

>>15186873

Hello guys neet here. Currently learning Unreal engine 4, already know C++ fortunately

Is it a good idea to get into game dev for PC and consoles for a 1 man indie developer? Or will success be easier to find on android/ iOS?

>>15227975

>10,000 collision balls

>no webm

Don't be a tease, anon.

>>15228125

Mobile games don't pay, unless you're one of the 3 that make all the money.

>>15186873

>>15228196

Yeah figured. Have you made any money from pc/ console games?

>>15227706

What file format are you using, because I've had a similar problem and it turns out .OBJ files index starts at 1 instead of 0.

>>15228233

I'm a nodev.

>>15228474

It's common knowledge, anon. No bully

>>15228233

>>15228474

>Have you made any money

>Jew thinks he can't be outed if he posts a Nazi cat meme fresh out of imgur.

You're not fooling anyone. A nodev ideasguy can learn to make game, but you will always be a shekel-chasing failure. If you want to make money, go start a patreon and cry about the anonymous meanies on 8chan who bullied you.

>>15228555

If you want to sell your game then you too are doing it for money. Stop your holier than thou bullshit pls.

>>15228571

based

>>15228633

How's it feel being retarded, 4cuck?

>>15228637

he btfo you so you reply to me? lol.

>>15228655

I'd tell you why you're new and retarded but where's the fun in that.

Try opening the board settings and see if you can figure it our yourself, samefag.

>>15228678

>implying i wasnt roleplaying a 4cuck refugee

lmao

>>15228682

>I was only playing g-guys.

>>15228474

>>15228571

>>15228633

Dude just upload the balls webm.

>>15228824

nevermind.

>>15227975

Tell me more anon. Which language, is it 2D or 3D? When they bounce are they using simulated physics, or just arbitrary behavior you gave them? They only collide with the walls?

>>15226584

>first

Build an array of tiles in proper order upon initializing, and use that

>second

Use a position component, each actor gets one, in the component store which tile/thus position (index in ordered array per Q#1), and use the position system to update its component per tick.

For positions on a limited/known tiles per row/column you can use a spatial formula similar to how threads are ordered in compute shaders, like x + y * 5 = index position

X is column position, y is row (5 tiles per row).

>third/fourth

Instead of defining each actor as an “object”, it would be easier if you define it in an ECS sense.

Each actor is an entity, with a set of aspects defining it (components).

A rock and a boat both have sprite renderer components, and water system components, but their component data is what differentiates them.

A rock and boat have different sprites, and different flags defined for the water system’s component (sinks/floats); though both use the same components.

Systems are there to transform the data of the components.

F.e. To move an actor, you would use a locomotion component (defining speed, what tiles it can move over, valid movement directions), the locomotion system would use the input component (polled input data, done earlier via input system) to apply user input, the locomotion system would then use valid input (as defined by locomotion component) to apply a change in position, and would write the results to the position component.

There’s a few good articles on ECS if u do a quick search, look at the T-machine article and the gamedev.net articles.

I highly reccomend to use an open source library for this, as building your own fully featured ECS framework is a huge task (a competent framework is huge, and should have code generation).

If u have specific questions or if I missed a detail specific to your usage case then lmk and I’ll do my best assist.

>>15229833

Looks like it already is. Here's some suggestions.

1- Keep a counter of missiles shot by the player: low amounts can give a score multiplier (maybe 1:1 duck/missile ratio is 2x or something). Alternatively, limit bullets (there can be ammo powerups to shoot).

2- randomize duck placement.

3- more enemy bullets to make it more challenging

4- add a start and end screen so it feels like a cohesive thing.

Curious that you went for the same gameplay as my own first game prototype/experiment.

>>15229465

What is this for?

>>15229833

Put an effort into adding more juice/game feel.

Look at "juice it or lose it" on yt

>>15229465

Is there a difference between both or is it just the background? The cyan one separates the foreground better.

>>15228125

>1 man indie developer

>Unreal engine 4

You're gonna carry that weight.

>>15229905

Oh. I like the right one though I think the left one is more proper.

>>15229889

So I'm learning "juice" is apparently an accepted term for game feel/little details.

>>15229477

I'm curious about getting into ECS systems after seeing your posts. What open source libraries would you recommend looking into? I'm interested in lightweight and minimal dependencies, if that matters. Thanks

>>15230006

Actually, juice is just a particular type of game feel term. Although it gets used as a blanket term quite often, I guess.

A few types of game feel are juicy, sticky, snappy, and jerky feels.

>juicy

Something moves, and its motion has follow through. Imagine a button sliding down, then overshooting a little before going back to its final destination. Or Samus bouncing a bit when in morph ball form in the original metroid. Was showcased very well at GDC that one time, with a pong game.

>sticky

Impact followed by a pause, mostly. Think that little pause that happens when you hit enemies in some beat'em-ups, or fightan games. Or when your character falls form a larger height in some games and sticks to the floor a bit when you hit it, like in 3D Zelda games or Hollow Knight.

>snappy

Instant reaction. Think when you click a virtual button. It doesn't slowly slide down, it snaps down. Preferably with a clicky sound.

>jerky

Delayed or even opposite reaction. Can be terrible, such as lag, but also applies to the kick upward when you shoot a shotgun in an FPS.

Sadly there's not much discussion online about other types of gamefeel, aside from one shitaku article by a guy who has a serious problem with getting to the fucking point. http://archive.is/nFnUE

At least ctrl+f "THE FRICTIONARY" if you read it, that skips his 5000+ word prelude about his balls.

Got entity and wall collisions moved over to a spatial hash and happy with how much it dropped checks. Ten idle entities spread out around the level went from 117 checks down to just 1. Next up to get projectiles back in with the rendering and collision changes and add swept collision checks for fast moving objects.

>>15230006

If you're not familiar with the principles of animation (easing, anticipation, mass and so on) look those up. The general premise is the same that you're trying to make something more satisfying and alive feeling.

>>15230313

I think they call "sticky" hit stop in fighting circles. I remember some dev showing a video that used different lengths from popular fighting games and it was surprising how recognisable the feeling was just from that.

>>15230454

>>15230313

I have seen that pong video and another related one. Juicy doesn't mean a specific tween type, but all little details that include tweens, particles, etc.

Blizzard has always been good at this sort of thing: units die satisfyingly and hits feel punchy.

>>15230131

Good to hear

I currently use the entitas framework for ECS.

It has support for quite a few languages, mainly C# (what I utilize), but also languages like C++, GO, java, and so on.

It's fully featured + production ready, and has examples/wiki for extending the framework.

https://github.com/sschmid/Entitas-CSharp

Entitas has code generation (though, only for languages that have some capacity for reflection), and code gen takes awhile to get used to (also a bit of pain to initially set up). Although, the structure, ease of use, and rapid prototyping it provides is amazing when you're acclimated to the workflow.

I'd recommend to watch sschmid's entitas ECS vids on YT, and then review the wiki/example projects.

I've been away from game dev and programming in general for quite a while. What's the status on FOSS engines? I'm assuming Polycode is still nonexistent. There's Godot, which seems pretty great and I'm trying out now. But I remember hearing about some other FOSS engine that was fairly obscure and still early in development but I can't find it again. Anyone have a any idea what it is?

>>15229871

>>15229893

>>15229905

Posted it here and promptly forgot about it. A while ago someone told me to switch most of the black on the inside of my sprites for dark colors; looks great, although I've been leaving some black to give the illusion of depth.

The black the is supposed to indicate that the bun is on the other side of her head, but I also like the way it looks without it, so I was wondering which one y'all thought looked better. The background was added on a whim.

>>15229976

The outline is there so the characters pop out more ala Mother 3, but I'm willing to give that style a try after I'm done with my current sets.

>>15230639

Polycode hasn't seen a commit since 2015, and there's still loads of decent 2D freetard engines no one uses because muh Game Maker and muh Unity.

>>15230778

https://twitter.com/ivansafrin/status/1024005167077896195

Ahhh, I see. He has moved onto 'more important' things.

>>15230808

>rainbow flag in username

>bitching about Trump

Checks out.

>>15230820

Is that even bitching about Trump? I can't really see anything I disagree with in that post. Seriously. Imagine showing people 20 years ago that headline. It would provoke so many questions.