/agdg/ + /vm/ Thread

I cannot into agdg

I guess these are something I find good, But regardless I want the opinions of others as well and see if they would work as well.

Your way of thinking is wrong. There's no one set way to make a shooter. What you listed works better for an arena shooter.
For instance, iron sights work better for slower paced games, that doesn't mean they're bad.

Well I guess I should have said it would be more fast paced. I dont know, I'm still mulling it over and wanted other peoples opinions on the matter.

just like make game and experiment until you find what's fun

Everything from RNG to regenerating health has it's place, you should never rule out any feature just because it's overused or doesn't work in a certain game.

Similarly, IMO you should avoid picking features straight from other games. Think about how you want your game to play, and then decide what kind of gameplay features would support that.

As a random example, you could have all guns take ~1 second to charge before they shoot a bullet, that would significantly alter the way the gameplay feels, and would probably pave the way to making it more unique than the vast majority of the 63460334 other shooters out there.

Asking whether you should have reload or fast movement just seems like such a pointless question when you don't even have a game yet.

Just like, make a game dude.

ok that works

Time to buy a rubber duck.

Happened to me a few times too

Hey guys. Anybody know how to correctly import animation from blender to ue4?

my idle animation is frames 1-240. But for some reason it includes the next animation. from 241-480. I don't have these separated as separate animations in blender. Why would it pick up both?

I can also import the entire sequence but I don't know how to separate it into individual animations

I always save my animations as separate blender files so I don't know what to tell you.

You will, however, wan to set the skeleton to the player skeleton when importing animations.

alright guys i've been thinking about something recently
It has to do with spatial indexing so be warned i guess.
I'm making an rts game and
I've been reading up on spatial indexing methods recently but a lot of them see overly complicated

My thought was what if just make a 2 dimensional array that turns the entire map into a grid.
So you would use the index multiplied by the grid size to get the position.
That way collision would be a lot simpler but I'm not sure how much exactly.

first of all creating this data structure is simply O(n) in all cases no matter the grid size.
ok so lets look at things i will use this for
finding:
a unit in a certain area worst case is O(s + l * a) where is is the number of squares inside the area and l is the number of squares bordering the area and a is the number of units in those bordering squares
finding a unit colliding with another unit:
so this will just be bordering squares so worst case is O(l * a)
and if you want every unit instead of just one in either case the time complexity is the same except every case instead of worst.
but from what i can tell this is better than most complicated spatial indexing data structures
for instance quadtrees
which promise insertion performance of O(log(n))
and searches of O(log(n))

but with this homebrew solution searching is based on the size of the area searched.
So scaling is basically O(1) which would be much better with a large amount of units.

Am i missing something here?

man you're smart

Really?
cause i feel pretty stupid
Since my stupid solution couldn't possibly be more efficient than algorithms people have worked very long on

though maybe it has to do with the scale of the game

since with my system searching in an area would be roughly O(a) where a would be the amount units in that area
whereas a quad tree would always be O(log(n)) where n would be the total amount of units
so i guess it depends on the size and scale of the game which is more efficient.
though creating the quad tree is always slower.

I just wanna make sure I'm not doing something completely retarded before I start refactoring my game

I cant even get my character to idle correctly.

Take your time man. Eat some food and put some good study music on.

Just don't do what I did a few years ago and put tiles in a list and you'll be fine.

If you ensured the cells were at least as large as the largest unit you could get away with collision checking against the 8 surrounding cells. Though, if there's a large disparity in unit size you could still get trouble if there's a ton of small units packed together in one location.

Nah. Most of the complicated tree shit is intended for 3D where allocating enough cubes becomes a problem (and you don't necessarily know how big/small they have to be in advance either).

right now the largest unit will be 300x300
and the smallest 50x50
though that is still subject to change
Also i thought making the grid size as small as the smallest unit would actually be the best.

Depends, then you'll have to mark large units in several squares (but on the other hand, small units will need to check vs fewer units if they're tightly packed. Probably a good idea if there will be a lot more smaller units).

Threadly shilling for Pico-8

drive.google.com/open?id=0B_5ZkqdDFIVdZklYbVpYYTcwNzg

i guess i just need to experiment
also the way im programming this i can actually change the grid size at any point during the game so maybe i could write some function that sets it to the optimal value

Rubber duck debugging is a thing of beauty.

The last two days have been good. I implemented hit detection, a charged attack and some hit effects.

I will never know the truth

Spatial data structures are highly dependent on the distribution of your data. The more clusters you have, the worse they perform. I once wrote a boid system (flocking) and used a kd-tree. While profiling you could literally watch the performance tank as the time progressed. The boids clustered and the queries (all-k-nearest neighbors with k=3) started to take up more and more of our processing time.
The biggest strength of grids are their simplicity and how suitable they are for dynamic data. If your points move, it's extremely easy and fast to move them from grid to grid. There are various ways to store the grid (some more efficient than others), but that's not important for now. A grid is the most efficient, if you have evenly distributed data. The worst case happens, if all the points are clustered in a single cell. Then your grid does nothing at all and you're back to O(n²). Furthermore, you should keep in mind that grids can be rather memory inefficient depending on your data and grid are stored. If you have a very fine grid, it's entirely possible that the countless empty cells end up using more space than the actually used ones. Just something to keep in mind. The last thing is that grids are discrete. You create "steps" of a certain size (your cell size). The bigger the cells, the more points will be within a single one on average, the more your runtime efficiency scales towards the worst case. The smaller your cells are, the more overhead you get, the more memory you use and the more cells you will have to look into. The problem with that is that this will fuck up your locality of data. Getting data from the heap is very slow.
Spatial trees have different characteristics. They are more complicated and expensive to build. This also makes them more difficult to use for dynamic data. However, querying them is a lot faster on average. Their performance doesn't degrade the same way the performance of grids does. Simply put, they handle clusters better and clusters are way more common than even distribution.
What kind of data are we talking about here? Moving units or stationary buildings? Both of them should be stored separately, since the actual data has nothing in common.

Draw a list of drawn elements, draw a grid of elements larger than your screen, see if the amount of drawn elements corresponds to the amount of elements listed.

Was going to try and re-save animations one by one. Now it wont let me delete key frames.

Why cant anything just work

actually the way i've created my game there is no difference between buildings and units
since some units can't move and some building can

Then stop doing Object Oriented Design and start doing Data Oriented Design. Have separate indexing structures for stationary and dynamic entities.

Don't think I can do that without changing the engine code

I would rather not
one of the reasons i like this solution is because it's simple to understand and easy to program

shiggy

So are the basics of DOD. The difference in question here being that OOD is a top-down approach (design code first, then put your data in), whereas DOD is a bottom-up one (what's the computer dealing with? Then design the API around that data). In this particular case it means that you stop thinking about the concepts of "units and buildings" and rather think about "stationary and dynamic entities". A moving and a stationary unit are nothing alike, even if they might look like it in real life.
Still, the most important thing is that you actually get shit done. If you want to stay in your comfort zone first, by all means do it. It's better than trying to do too much at once and ultimately producing nothing. If, in the end, you find out that your approach is flawed, it'll still have been an invaluable learning experience and you will be able to do better next time.

i've been defiantly homebrewing almost everything since i started programming
and it hasn't been a problem yet

also
I don't think that many people would recommend data oriented design

it's probably just 2 or 3 people, but they do a lot of talking

Put a fat bunch of performance killing shit, then look away and see if your fps increase.

yes, I also don't want to rewrite my entire game based on a design principle that might not even be all that beneficial
all probably try my hand at DOD once I'm done with my current project

also I don't think this project actually needs it, I'm not trying to make supreme commander here.
and if it does need it i'll just give up and make something else

How do I into level design for a multiplayer shooter? Should I just ape elements of what I remember from older games and make frankenstein monsters?

Use actions, baka
Change the dope sheet mode to action editor.
They get exported as FBX takes.

That's perfectly fine. It's just something I like to repeat, because it's relatively easy (literally the easiest part about DOD). The whole OOD approach is basically "designing-by-analogy". It's the developer acting as though his PC is doing something it is not. That shit needs to stop. It's one of the reasons I started writing the /agdg/ guide. Having even a basic understanding of what your code actually does can get you a long way.
There is nothing wrong with taking a break for a minute or two and ask yourself what exactly you are »actually« dealing with. And you don't need to be a CS graduate to understand that either. Well, unless you want to take this to autistic levels.

you're amazing thank you

You have to think of what's visible from where, elements you can walk past to block that line of sight for others, connections from area to area, etc.
To start it all depends on your movement though.

at what point will you realize that no one is interested in what you have to say?

Hold up, you can make different animations in one .blend file and export them separately?

Is there a reason your frame indicator looks blue instead of green?

Export actions separately? The FBX format has 'takes' which are more or less the same thing as actions, and engines will import these as individual animations. You can also uncheck "All Actions" to only export the current one.
Keep in mind export and import times though, you don't want to keep all of your anims in one file.

My workflow is as follows
>link the models and skeleton from the master file
(caveat : you might want to remove the _proxy suffix after you do that)

I've changed some UI colors so maybe that. Timeline's indicator is still green.

Maybe some day I will. In the meantime I will keep posting. Sorry, if that upsets you. :^)

Interesting. I'll have a look.

I'd say we've been pretty gentle about it

Is DOD the new blender?

Can I just copy the frames over into the new action?
Or do I have to re-do the entire thing

have you been introduced to the concept of 'aggressive autism'?

...

Thank you.
Sorry, I feel like I should know half of this already.

Well, I've read your posts.

Think about a standard computer science class. Most of the people there have no idea what they're doing, but their autism will compel them to talk about shit no one actually cares about while trying to out-compete everyone else.

It's a special form of autism that leads to people talking about "data oriented design" on game development thread.

PANIC

I think you're making a much bigger deal out of it than it actually is.

Maybe I should write a wiki on it

Seems like it needs to have the same channels. Try this.

you could at least tell us where to find out more about it. when i google it i find nothing that could be right.

forgot to name it cocks.

got it

what is DOD?

you're welcome

Your approach, since the grid is static in size, seems like a better approach if you can implicitly limit the amount of objects inside your squares at once, and also if the total possible units in that grid square won't cause dips in performance (where maybe a quadtree, w/specific unit amounts, would perform better… as it's fewer comparisons for say collisions).

Quadtrees are useful when you want to have a specific amount of units per quad (square), dynamic splitting (splitting rules, if this & !this -> split), and reliable lookup/search/splitting times since you can just limit the depth of your quadtree (i.e. once at depth 3… do not split, as the quad size is at minimum size… as recursive splitting + re-insertions are the main issue).
Also, they're not too difficult to implement once you get your head around how recursion works (then u can implement a specialized quadtree specific for your uses), and source code really helps (don't look at generic ones, look at purpose specific ones).

It will more than likely be slower w/a generic quadtree, most definitely, and this is the main issue since you use recursion to create the quadtree.
However, limiting depth, and having smart split rules will get you around the brunt of this slow down (have complex split rules to prevent doing needless recursive splitting, also a specialized quadtree for your uses).

Not necessarily.
With quadtrees you can have that large amount of units separated into smaller quads, hence having fewer comparisons; although the implicit overhead for re-instertions, and splitting is incurred (so, would you rather compare 200 units all against eachother, or 20 units against eachother + splitting and re-insertions?).

Data Oriented Design. It's a code design paradigma/philosophy/call it whatever you want, that started to gain traction in the games industry some years ago. Some user posted it in one of these threads some time ago, and it now comes up every now and then. (There is also a thread on our board.)
It's the typical situation where something new comes up (in this case OOD - Object Oriented Design), then gets hyped as "the solution to all our problems" and everyone blindly jumps on the bandwagon. Then, a couple years later, people realize that it's not perfect and that there always were drawbacks to it. Stuff like that happens from time to time.
The basic ideas aren't anything new really. The name originates from a developer from Insomniac who made a presentation on a C++ conference. The goal is to write code that is easy to maintain, but also fast. OOD certainly gets the former right, but sacrifices the latter. How much, depends on your code. Whether it even matters, on your line of work. Being game devs, performance matters a lot to the guys from Insomniac. Especially since they target consoles.

Daily reminder that yes, any retard can fork an engine and shit out C++ like there's no tomorrow.

Where did you find this gem?

Tesseract's forums.
This faggot is working on a proprietary fork of Tesseract that kludges new features onto the engine and thinks he's creating the next Unity, while in reality he's just making a shittier Octaforge with AngelScript instead of LuaJIT and a really terrible method of adding Bullet physics that completely goes against everything Tesseract was built around.

Is there any open source version of Pico8?

What is with people and thinking they all can make the next great engine? Oh well, it's not like anyone will want to mess with it seeing as it's proprietary.

I don't think there is one. As far as I know, the dev has been pretty obstinate that it stays closed.

What are some of your favorite FPS'es?

Think about what makes them so fun for you.

For me, the best games with guns are all about audio and visual feedback. The way a gun reacts when the player clicks the mouse, the way it sounds, how the enemy reacts depending on the gun and where they are shot, gibs, texture splatters, muzzle flash, etc. All these small individual elements happen in the span of seconds but combined well can lead to a fun experience.

Then you can add shit ontop of that. What mechanics can I implement to make everything feel more natural or rewarding to the player? Do I want fast paced, running and jumping off walls with slow-mo feature that lowers the gravity and slows everything down through a sharpness filter? Or do you want slow, methodical, and tactical movement and shooting where death can be as instantaneous for you as it is for whatever you're shooting?

Then you have level design, enemy placement, etc. on top of that. There's a ton of shit that goes into any game genre design.

I've been procrastinating getting the next set of features in my game. Once I get back home, I'm gonna make it happen

Learn blender. You can do anything with it. It just takes time to learn and git gud.

youtube.com/user/cgboorman/videos
Start here.

I believe in you user

how do i maek geam?

you don't, get out now, it'll take years before you have anything remotely presentable and you won't have any marketable skills

even if you do work for a AAA company you'll be abused with constant crunch time, zero respect in and out of the industry, and a shortened lifespan

might as well play the fucking lottery if you're an indie developer

Do a kickstarter and then pay other people to do it.

Gaem can be maed easily using Younitee or ArPeeGeeMaker.If ur srs bout it, mak sure u'r first game vry shrt, less than 10-15 mnutes.

Strting first, you hvnt got displin. But, moaer gaems u devlop moer displin.

No wonder the rest of your posts this thread have been so shit.

this is the problem right here
retards who think that having the opportunity to waste your life working on games 80 hours a week should be cherished

anyone who actually wants to get paid for working their asses off is an entitled bastard who should fuck off

you are the pajeet

right?

You can copy digital stuff infinitely, it doesn't cost anything and some people want money for it LOL

Aye.

Games are a waste of time and money and destroy lives

bostonmagazine.com/2012/07/38-studios-end-game/

It's a hobby, half the people here are on NEETbux anyway, and the main goal is to make a game you want to play. Also

...

PROGRESS

anyone know any good profilers for windows
i tried very sleepy and codeXL
but very sleepy doesn't load debug symbols and codeXL can't show me incluse function samples

The thumbnail better load for this one

Damn, that's just sad. It's always depressing seeing someone with the best of intentions fucking everything up like that.

Just a heads up, I'd be very happy if you'd do that guide.

thank you user. I was able to get home and code in one new thing before I was so tired that I had to crash. It was kind of a big thing, so I'm okay with only getting that one thing done.

I've been working on a voxel system, and before, it would just output randomly generated data. I got it up to the point where I could save, so I figured it was time to start having it output some meaningful data was the next logical step.

I was able to get it wired up so that it has a database of different 8x8x8 "chunklets" and also a database of "slices" (8x8 blocks)

from that, I hardcoded a couple values temporarily to make a house (3 different 8x8 "slices") and then programmed it to make 8x8x8 "chunklets" out of either nothing, just cubes, or to build a house.

The screenshot here, is my first step in getting it from just random data, into a sort of "scene". At this point though, the scene doesn't consist of anything except big blocks of blocks, and simple house structures.

In the future, it will generate houses and other structures procedurally from parameters, but for now I just needed to get these systems hooked up and talking to each other.

...

I got IK sort of working.

Question is, how does one attach one character pawn to another and still have its animations react to its environment? Because when I just bump into it, it walks whenever it's forced to move.

Just not when it's attached.

idea guy here pls tell me I'm useful

Does that kick have the impact it should have? Maybe she should duck her upper body a bit more during the windup to telegraph it harder.

You're the least useful member of the team you'll never be able to be in.

at least learn BP in UE4

i have a task for you
write an interesting and engaging story related to boxing that isn't rocky/hajime no ippo

its hard to write sports stories for sports when sports are the main focus
it usually falls onto character relate-ability and their convictions to sell you on it
of course since this is how the vast majority of sports movies go, and boxing especially you could always give a "genre-breaker/deconstruction" thing a go, make fun of common conventions in sports films, the unexplained high pressure envoirments and so on
writing isn't as hard as you think
if you're having trouble writing something from a specific kind of genre, go consume a bunch of media from it and look for it's weak points and try to draw inspiration from others ideas
there is no original thought, only good thinkers who can mask their thievery well

Maybe you could copy Butch's story from Pulp Fiction. Mafia involvement or something.

Having asked other devs how they wrote their games, most seem to write the beginning and the end first, then fill out the middle as they go.

Great read, but holy shit.
What's with those fucking astronomical numbers? If I could get my hands on 0.1% of those "just" $22 million in assets, my life would be a lot less complicated.
As for the guy himself. The more I read, the less sympathy I felt for him. Calling it a train wreck doesn't even begin to do it justice.

Learn to turn your ideas into design documents and you'll be useful. Though it'd be advised to still learn some kind of scripting to help out anyway.

why not have a boxing game set in the Victorian Era when Bartitsu came about?

docs.google.com/document/d/15ixwxD-Z0g_eb_jmtQjdfBzJ1IUO05fAn0QB5wqE9bs/edit?usp=sharing

no

i'll take it
sadly i'm a starving university student, so i can't really repay you in any way

don't worry about it

do you want me to keep writing that doc up or just give it up to you?

go wild, i'll try and exploit as much of it as i can
assuming i ever have the time to do so

alrighty

pajeets also expect quality work done for pennies

finally

time to piss off /k/

...

you might be useful
give me some good ideas on unique mechanics for an rts game.

Do you have a team of at least one code monkey, artfag and sound guy working with you? If not then no, you're not useful. Idea guys are only useful if they've got the manpower to realize those ideas

unit combining

you select two units like a buggy and a tank inorder to form a tank which can use roads to travel across a map faster or a buggy and a heavy troop to make a tankette which is early game anti-infrantry

ball crush

there's a few massive balls on the map that you have to make your units shoot inorder to move and crush your opponents units with as you can't directly destroy them but merely hinder them in certain ways(immobilize, make them unable to shoot) them

scrap magnets

you have infantry with magnets that pull components off of vehicles to use in vehicles of your choice

drag racing

maps are actually two or more massive vehicles which each player owns and as such have to assault others and defend their own by controlling important parts of the machine that effect how the massive vehicle is performing in the race and if your vehicle wins the race you win

I've no man power, just broken dreams

My OS arch linux inexplicably broke after years of faithful operation and for the first time since I installed it years ago I wasn't able/just don't have the goddamn patience to fix it. It pissed me off so bad that I took my entire computer apart and cleaned ever square inch and routed the cables autistically last night before a fresh fedora install. Feels okay man.

my issue is I don't really know what one looks like

i was actually already planning on adding something like this

I actually like this idea but more as an one off game mode not something i would base my entire game around.

so units that steal stuff from other units?
not bad honestly could do some pretty cool stuff with this.

???

I ran out of anything interesting and just basically gave a reason for a point control mode

but feel free to whatever else

Same thing happened with Debian after using it for two years, Granted, I think it was my fault anyway. I was fucking with partitions and one thing led to another and it wouldn't even boot. Wiped that, installed Gentoo, and restored a backup of my home directory after that. Surprisingly, devving with linux isn't as bad as I thought it would be, most things just work.

Somehow everything in /usr/bin got masked during a system update. I blame yaourt and systemd.

...

Absolutely crazy.
I dig it.

That's something alright

Every time you move a unit, an object spawns on the map that multiplies every turn. Your goal is to win the map before the game crashes.

You had my curiosity before but now you have my attention.

That whole concept is literally the Eternal Silence mod for half-life 2: Two massive space ships that attack each other automatically; their health decreases over the course of the game; players can either try and reduce the health of the enemy's ship from the outside, or capture and disable internal subsystems; altogether providing a big enough advantage to win the match.

A bit sad that the multiplayer died long ago, but great fun nonetheless. Also the dog fighting in that game had the best controls (6 axis), and the infantry battles weren't as generic as you'd expect.

I love this idea.

Spainfag here. I'm quite new at coding and I've been trying to make a simple RPG without RPGMaker, I hate how every game is so samey on it written in C just for practice. It's still too early to call it anything other than a bunch of code though.

Today I've been trying to write a random name generator but after 5 hours I think the names don't feel quite right, even if they're supposed to seem exotic.
The function basically picks the amount of syllabes and the amount of letters per syllabe using rand(), then creates the syllabes by randomly picking letters according to conditions/rules loosely based on Romance languages (spanish, parts of english, etc) and depending on each syllabe's length. Result:

What do you anons think? How should I improve it?

I've written a random name generator, what I did was have a list of prefixes, mid segments and suffixes. Each name was a randomly selected prefix, randomly selected mid segments * n, then a randomly selected suffix.

It's obviously not as random as your approach and the lists of word segments can get prety big but this approach worked well for me. You can easily tune it so you get similar sounding names if you want location based names, or have them as varied as you want.

I just so happen to really enjoy constructed languages (which extends into names). The long and short of it is if you want your names to sound like they come from an exotic, yet internally consistent language you are going to need to teach yourself something about linguistics, at the very least how the IPA chart is laid out and works, and how syllables work. Give me some time and I'll see if I can dig into some resources that may help.

In the mean time here's a random word generator that you can play around with.
akana.conlang.org/tools/awkwords/

Perhaps instead of picking random letters, pick random sounds instead? Perhaps add some randomness to it where some sounds ending with consonants may get vowels appended as well.

So you have an algorithm that decides the amount of syllables, then generates each syllable by randomly choosing from a list of sounds ending with consonants, and then have there be a chance that a vowel gets appended?

I dunno man.

If the camera is just a rectangle, and your function to cull objects uses a rectangle to check; pass a rectangle thats smaller than the camera to your drawing function and use a camera normally so that you can see what is drawn and what isn't.

Alright, here's what I have. Not everything in all of these videos are necessary for you to write your name generator program, but the better you understand how language works the better your name generator will be.

Also it turns out that one channel has pretty much everything you need.

Introduction to making conlangs, in particular sounds.
youtube.com/watch?v=sFWc0sBO62c

These are not absolutely necessary especially if you are just using sounds from English and Spanish. But if you do end up totally skipping them you will have to learn about the IPA through other sources. But, hey they still offer good information if you want to watch them.
IPA and sound stuff
youtube.com/watch?v=xMEFr7ghMTg
More IPA and sound stuff
youtube.com/watch?v=J3IO5K5ZGB4
More IPA and sound stuff
youtube.com/watch?v=jkfSA4_DCfs
More IPA and sound stuff (mainly Vowels)
youtube.com/watch?v=nhOhZ5HSd54
Wikipedia link to the IPA chart and an explination of it
en.wikipedia.org/wiki/International_Phonetic_Alphabet

Selecting sounds after you have some understanding of the IPA chart.
youtube.com/watch?v=3378FlHK4v0 (I apologize for the banter in this video if you don't like that)

Phonotactics. You absolutely need to understand this to make your name generation script work, and you will only really understand this if you understand the technical, linguistic difference between consonant and vowels.
youtube.com/watch?v=1Up5hSm7LYI

And now you can generate names one sound at a time.

Other sources
David Peterson also goes over the basics.
youtube.com/channel/UCgJSf-fmdfUsSlcr7A92-aA?user=UCgJSf-fmdfUsSlcr7A92-aA
This website
zompist.com/kitlong.html

well fuck. back to the drawing board I guess.

the shadows look too spazzy in the second one
maybe disable the player's shadow during that

How do you scale pixelart so that it stays sharp when it's big, but doesn't become jagged and shitty when it's at something like 1.2x or 0.8x size?

I like that Arctic Monkeys song.

you don't

Still looking for either:

A) people who need a writer, texture artist, or proofreader for a project

B) amateurs who are looking to be part of a side project mainly focused on helping people learn to work in a team environment and/or get better at their dev skill.

Fuckin ropes man, how do those work?

I mean hell, for all its faults, even the fucking Source engine had rope constraints (and decent ones at that). Does UE4 fair any better in that regard?

unity does have different kinds of joints
it's just that they're a pain to setup, and getting a decent looking rope/chain with them is even harder

Unreal engine 4 has something called a "cable actor". Here's a blog post about it.
unrealengine.com/blog/cable-component-plugin-for-ue4
And there is also a plugin that supports cables/ropes with collision, but the forums are down for maintenance at the moment so I can't check for sure.
forums.unrealengine.com/showthread.php?55332-Dynamic-Rope-Cable-Plugin

Nearest neighbor scaling.

That makes it precisely "jagged and shitty".

You can upscale pixel art only at full numbers. 2x, 3x etc. Basically, for every pixel you show 2 or 3 pixels instead of 1.
1.2 or 0.8 pixels are impossible.

No it isn't. It just becomes slightly blurry.

The problem is that bicubic/bilinear makes 2x+ a total blurry mess, but nearest neighbor makes anything that isn't full multiplier into an ugly jagged uneven shit.

If you rendered every pixel separately as a 3D quad and enabled anti-aliasing, it would look like what I want.

...

Integer scale it first and then use the smoother operations on the much larger sprite, I guess.

Or never do that at all and reconsider your entire choice of using sprites in this manner.

You've just described every aspect of at home game devving

Not my webm, but I'm doing something like this.

What's the side project?

I'm always looking for people to help out with my game
And I don't have any story at all yet
I'm making an rts game, mp4 related
also i do have some idea of what the factions are going to be like thematically

So the problem is that you want to have visible pixels at >1x scale, instead of the image being blurred?
Or do you want the pixels not to be defined?

I'm not sure, really. I've got a few ideas for relatively simple games that could be made in a span of months and just polished up to hone talents.

I'm happy to help with whatever, just yell for me in the threads and I'll (probably) see.

i've always been curious, how do you select units with the rectangle?
my first idea would be to just have a list of all your units, and when you make the rectangle, check if each of your units are within the rectangle
but if you get a lot of units, i'd imagine that it would be a costly method of doing it
is there another way?

Exactly how many units are you planning to have?

none, i'm just asking hypothetically
in age of empires, you usually have at least 100 military units per player, up to 300
and when you get like 4 players, i'd expect shit to run horribly this way

You could probably have 10s of thousands of units before you'll have problems with a simple collision test like that.

Even then it wouldn't be too different from all the culling you'd be doing in each frame anyway. (If you were doing the engine you could even save most of the effort by keeping the list of visible units around and using it for both purposes.)

right now i just have a unorder_map with all units in and i just go through them all and check if they are inside the box
but im currently implementing a grid based data structure which is a type of spatial indexing which will allow me to more quickly find units in an area
which i will use for selecting units with a box and also a shit load of other things

unreal or unity? i have a very complex rpg with and event tree that changes based on actions and time.It's honestly getting pretty complex, so im wondering what engine would make this task much easier

Reminds me of Dungeon Dice Monsters.

I made a quick half-ass RPG using UE4 in about a week one time. Honestly, it's the easier engine to use, but any game you make on it will run like shit in low-mid range PCS, no matter how it looks

Did somebody made a Shader for UE4 that it is less demanding?

Sounds like you could do what I described.

Basically, make your sprite look pixel-for-pixel what it should look like at the furthest away possible zoom level.
Then integer scale it to fit your closest, biggest zoom level. That is your sprite now. All operations are to be done on this big fucker.

If you make small nearest-neighbor scaling changes on big sprites with multiple visible subpixels per actual pixel, the error gets evenly distributed and you can still make out all the details, even if the exact proportions between the pixels themselves aren't perfect. Then you can use some smooth interpolation to shrink it down for rendering and it'll probably look like you want it to look.

If you try to do small changes to small sprites and blow it up later, you're going to lose detail and make everything looks all kinds of fucked up.

For example, everything in this is nearest neighbor, and one of them isn't lookin' too hot.

You should now know how to into mixel-roxel indie pixelshit.

Drawfag here having a tough time balancing getting good at drawing with getting good at coding. How much daily time investment should I put into learning to code for good information retention?

you should probably do one at a time

can you make me a map texture?
like right now i'm just using something i pulled of google but it's not very good

Preferably quite big something like 5k by 5k though bigger is not a problem either
it doesn't need to be super detailed and i don't mind if there is a lot of repetition
also it should be to distracting not to many crazy colors and shapes and such

And put this dual monitor setup to waste? I don't think so.
In all seriousness, does that mean that when I learn coding, I should dedicate a few days in a row to learning coding? Is that what you mean by one at a time?

I don't think there's any correct way to learn.

IMO you should dedicate a day to one or the other, not try to shoehorn in both. They're very different kinds of skills and use different sides of your brain, and I think there's a certain kind of warmup time before you really get into it.

Try to keep a vague schedule; each day of the week for either art or coding.

The best use of dual monitors is to keep your work on one, and movies/streams/shitposting on the other.

I mean get to know one really well before you start learning the other
It's not necessary to do it this way
but if you know very well you can spend most on the other without losing or forgetting to much of the other.
I just feel like it's easier to learn of you focus on one thing at a time but maybe thats just me
also either task benefits greatly from having to monitors

with programming for instance i have one monitor with my text editor and console
and my game and my browser on the other

Sound advice

I already do

That's fair. I had recently put down coding entirely because I felt that learning to code when I sucked, and improving my drawing when it's good enough to get me into a fairly nice art college, was just dragging down the art. I guess the best course of action is just to leave it alone until I get more comfortable with where I am with my art again.

Thanks a lot, it took me a bit longer than expected but I rewrote the function to read common syllabes present in romance and english languages from a list + some improved rules to determine when and if a letter should be added as preffix/suffix from a specific and consistent-with-the-language set of letters.

Result:

Much better than before, the names feel ok and now the rules are easier to tweak.

Coding can be pretty tough. I can't cite them right now, but there have been studies on memory and they have found that the best way to remember something is repeated exposures over time. not necessarily cramming as much in as fast as possible.
Theres an app called "Super Note" I believe, its like some kinda juiced up flash cards. Not that I'm recommending it, but if you want to look up the science behind it, that's a way to track it down.

Anyway. I would just suggest that even if you do make art your main focus do spend a little bit of time every other day or two, just reading about coding at least. Maybe one or two hours, just study your favorite language. You don't necessarily have to be programming anything, but just keep reading about the stuff. Even if some of the concepts are over your head, just keep reading them and coming back to them.

Over time, all the of the terms and ideas will connect up together better, and when you are ready to sit down and focus your attention on programming, you will have built up a body of knowledge to draw from. Programming is one of those "Use it or lose it" skills, just keeping up with reading about it will keep it fresh for you.

how

simulating a town now, still have a lot of optimization to do. It's going to take a while, but it's coming together bit by bit.

Looks pretty good. Will it be able to generate buildings with interiors or is that as good as you think it's gonna get?

thanks, this is just the start. It will do full interiors with usable items, decorations, multiple building types, styles, etc.

At this point, I just got the chunk generator working, so its either generating open land, a "house" or one of those big cubes. eventually I will program in procedural system so it can build all kinds of stuff.

I'm not sure if I should start optimizing it again, or add in new systems/features

Neat, planning to use this for something or are you just experimenting?

building a game bit by bit.

I've been wanting to make a game like this for years and years and it was kind of driving me a little nuts. I eventually just forced myself to start learning and developing it. It's actually coming together a lot faster than I expected, it's pretty exciting to see it actually working to some extent.

Still a ways off from a game, but its nice to be able to start programming the actual game rather than voxel systems for a little bit.

picture related, very early testing of random data written to a chunk.

The != is there because for visible constraints, the eye actually evaluates to false, and it's true when you disable it
thus variable != 1 means "keep this constraint disabled unless variable is 1"

Making the default value of a custom property an integer instead of a real (0 instead of 0.0000) will make it behave like an integer.

Got my chunk loader done, this time with much less hardboiled spaghetti.
green = radius where chunks are loaded
red = border beyond which chunks are unloaded
white = loaded chunks

In the actual game green would be a chunk or two outside the screen, and red would be several/many chunks away, that way the required loading/unloading would be minimized.

Next up; actually load blocks into the chunks

looks abstract and nice

is Unreal even optimized enough to run voxel like structures? bagels and balls are going to eat up valuable polycount

Not if I eat up the bagels first.

based blender user.
Do you know why I can see through my geometry?

I bet the stupid fucking $100 custom cable of theirs has died. Fuck everything.

Thanks, the balls and bagels were actually just test pieces I made quickly in blender, not really final by any means.

I actually didn't think that unreal could do this type of a project, but then I saw what this guy made here: youtube.com/watch?v=EjCandEgtHo
and I stopped learning Unity and then switched to Unreal.

Im still not sure how they get the lighting working that fast, but I'll figure it out eventually.

One thing thats kinda cool that Unreal can do is instance meshes, together as one actor. So it only needs to do one draw call for each copy of the mesh you instance. So for example, this "village" scene here thats built out of just cubes is only one draw call because its being called as a single actor. Check out highlight on the right side

sorry user, I didn't write more I was 2tired today from going outside, I get to it tomorrow and if not the day after since I also have stuff on for tomorrow

did you draw the faces on backwards? renderes are normally configured only to draw one side of a face

Flip your normals.

I think it has something to do with the clipping plane. The normals seem correct.

Yeah, near plane seems too short.

Well, let's say you have 256 tiles to chose from, that's one byte. Then other stuff… Depends how much information. If for example you only have 8 directions then you can store that in just 3 bits, then use the last 5 for something else.

N key?

Well as a huge fps fag I can list few things I like

I like having lots of enemies at the same time
Fuck hitscan, if possible to only projectiles and melee
Item pickups of course
Fast movement
Secrets
Dont hold my hand. Don't make every intractable object flash/glow
Fuck having only bullets weapons, go nuts. Bazookas, grenade launchers lasers whatever

...

dead thread

i hope that means you're working on your game

what the fugg?

did you make collision for the model you made too?

I'm still waiting for a playable build of your game dude also do you got a Steam or some such I would love to talk to you about your game, even development I could learn a bit about coding even if you think you're bad at this.

How would you fags feel about a roll-a-ball adventure game?

Because I followed that shitty tutorial like 8 months ago and I've been adding polish and depth to it ever since. Basically the end goal is to give it some personality.

fucking dropped

Compile Ardour

Now to add the weathering, but at least the base is done.

no thanks, i use DAWs to get shit done, not fuck around with hotkeys and configs

10/10

That happens when your near clipping plane is too small and you're in perspective mode. Scale your model up (say by factor 10) and export it at factor 0.1.

Fucking hell, that looks horrible. Will you get some smoothing groups and a proper normal map on that model, please?

FUCK YOU UNITY
WHY DO I HAVE TO CONSTANTLY SECOND-GUESS THE VERY BASIC FUNCTIONS OF MY TOOLS
I REMEMBER LIVING IN A WORLD WHERE BLAMING YOUR TOOLS WAS THE LAST THING YOU DO, BUT NO, NOW I FUIJKCING HAVE TO FIGURE OUT WHAT HOW AND WHY BASIC FUNCTIONS DON'T WORK

Change to an engine that isn't shit and lets you see the source code and you won't have that problem ;^)

if greentexting and good writing are interchangeable, I'm some sort of writing guy

the amount of times I've ran into unity functions that became obsolete in an update, without them giving a warning, is completely fucking absurd

If you're going to keep a function but make it do literally nothing, give me a fucking warning or something

shitheads

I want to rape

I should start decompiling the unity .net code so I can verify that they do literally nothing

check this

github.com/MattRix/UnityDecompiled

i havn't yet spend much time reading into them, but yeah - almost all parts i was interested in are just stubs linking into the closed source C++ parts

Yeah, I had the same thing happen when I tried to dynamically change the UI navigation. It's not the fault of the Unity developers though. These are the value vs. reference type shenanigans you can thank Microsoft for. They made structs and classes behave differently, but look the same from the outside, so you don't know what you're dealing with at first glance.
Transforms are classes. Vector3s are structs. The former are passed as reference. The latter as value. When you call someshit.transform, it returns a reference to the transform object. When you change something, you're working on the original someshit object. When you call transform.position, it returns the value of the position. The value is just that, a valuie which has no association to the original object. When you change something about it, you're essentially working on a copy.
Use transform.position = new Vector3(x, y, z). That's the proper way of dealing with this.

Ah, yeah. I know the distinction (or lack of), I just hadn't considered that when I was calling a function called Set on the position of a transform of a gameobject, that it was all on a copied structure.

youtube.com/watch?v=ifFD8iD976g

around 13:00 he says the vertices should be up and down as much as possible

does he mean A or B on pic related?

unrelated to gamedev, but i'm studying discreet mathematics right now
can someone please explain to me what the fuck a closed set means
everywhere i search i get this:
the set F is a closed set if [F] = F
what the fuck does it mean

the one thing that really annoys me is that i can't make a class that inherits from gameobject, because of their shitty component system
meanwhile in unreal i can easily make a class that directly inherits from the actor class

That's because UE actors are not the same thing as Unity GameObjects. As someone who switched from Unity to Unreal, I found Unreal's system hard to understand at first.

yea, but in unreal i don't have to mess with a weird component system
i just treat everything as an object
i can treat everything as code, and not as some physical thing on a scene, which in turn makes things easier to organize

You were modifying a copy of the struct.
For the same reason, you can't set transform.position.x individually.
transform.position is a property, accessing it invokes get{} which returns a Vector3 and since it's a struct, it's giving you a copy.

You could've made that mistake in a lot of engines and languages, C++ included. (getting something by value and modifying it does nothing to the original)

But Unity components are not a physical thing on a scene. I guess it's a getting used to the Unreal/Unity way of doing things. I started porting my game from Unity to Unreal, and trying to port systems made for Unity into Unreal fucks my shit up. They're similar enough so that people compare them, but not similar enough for comparisons to actually get you anywhere.

they're literally a piece of code attached to a gameObject
and the gameObject belongs to the scene
how is that not physical

As I understand it, it's a set that includes its limits. Analogous to closed / open intervals.

can you give me an example

I guess the best definition I could give you is that it's a set containing its boundaries.

I'll just use a real number example. Consider the differences between [0,1] and (0,1).
The former is a closed set, the latter an open set.

For the closed set, if you're at 1, you're at the end, the boundary. There is nothing beyond that point in the set. If you go further, you're no longer in the set.

For the open set, 1 is outside the set. You can get real close. 0.9, 0.99, whatever. But you can always get closer. It's open in the sense that you can always cram another number in between an existing element in the set and its boundary.

And for every closed set, the complementing set containing everything has to be open, and vice versa.

In the case of [F], F is inside the set's boundaries.
If it were an open set, F would be considered "just outside".

thanks, that explained it

I've implemented jogging, which is a movement slower than a full-on sprint but faster than normal walking. It doesn't need any other bindings other than the run key. If the running key is held down, the character runs. If it's pressed and released, it will start jogging, provided that, in both cases, the player is holding down the walking forward button. The jog can be canceled by simply releasing the walking key.

I've also been working on item adquisition systems.

honestly, if you just add bulge/dick physics to that the reddit autism machine will propel your nodev career ahead by light years

It's a woman, actually.

But thanks for the suggestion.

my point stands

i got another question related to my subject
can you check if this Zhegalkin polynomial is correct?
x|y -> z = z + xy
i tried using the triangle method but i don't have any answer sheet since this wasn't the main thing i needed to be looking for in my task

nevermind, i found easier tasks with an answer sheet and i'm wright about those

/agdg/ - Homework

it's not like we have anything better to do

Wot.

cheeki breeki

it's not actually russian

found the ukranian hohol

Then I have a challenge for those unoccupied. Can someone tell me where to get legally free assets to use in commercial games? I've already looked around the UE4 forums and got plenty of useful stuff, but it didn't have that much free stuff. Can you guys recommend me other sites?

turbosquid, I guess

You know what to do

wrong again

Bulgaria is not a real land, it has no right to exist there is nothing. It belongs to Russia

Bulgaria actual name is "border" and it served it's purpose as border. you will become russian clay now.

filthy tatars go home

butthurt turks

more likely butthurt bulgaria.
Bulgaria cannot produce anything of value they have only shit. The only they have is resource which should be left better to russians they can handle it with efficiency.

it is overrun with muzzys, just give us your clay then we can get rid of your muzzys problem and the jewrop union too.

Putin is the savior of he white race but he cannot help you without signing this contract first.

yeah, that worked really well in WWII
thanks for the freedom and all, but we've had one too many problems caused by the commies

I come from the extension of glorious nippon.

go eat sushi, you slant-eyed midget

ah c'mon give us another try we changed alot during the 80 years.

We have learned alot from our mistakes and we will promise to become a good cooperation partner. We have cheap gases too.

But whatever don't tell to me I haven't warned you when the NATO comes in they will abuse you in the worst possible way. And it would help only spread the cancer that is American Imperialism.

Change the white with yellow, fam

chink? isn't it illegal for you to be here

gr8 show tbh smh fam 100

I was making a shitty joke, but now you've ruined it.

Back with more stupid questions.

I've set up my animation to open the door when a Boolean is ticked, and to close when un-ticked.
How do I make collision with the trigger area activate the boolean?
Which graph am I supposed to be using?
Am I supposed to set the trigger on the model itself like a camera on the pawn?

Uh, you forgot to connect the pins for the casting nodes in the last picture.

Also, you should have Shopportunity for both the main BP and the animBP, then in the animBP cast the mainBP to grab its Shopportunity to set it to the animBP's Shopportunity, so that its animation state machine can work.

HEADPATS AND HOLDING HANDS

Have you considered that that's terrifying?

Combat. No damage system yet.

Holy shit, you are terrible at your own game. Have you considered adding an over-the-shoulder aiming mode? It is usually better for accuracy since your own character doesn't block the target.

Have you considered making her a bit slower, or is this a shootan mosou

wtf is going on in this thread

It's now possible to add blocks, and collide with them.

Also adjusted the chunk loader, you can now make maps that are pseudo-infinite or wrap around vertically and/or horizontally. Also you can load all chunks in a fixed map at once if you don't want to deal with dynamic chunks. I'm hoping to make the engine as flexible as possible so I can use it for other games too.

Next up, either platformer physics/controllers, or graphics.

You shouldn't be doing handholding via transforms. You need an AI for that.

Here's strafing mode, which I plan to increase damage/accuracy.

First version used AI, but needed to put a navmesh through the whole level. Plus she kept bumping up and spazzing out. Will try to tackle it again once I figure it out.

not like I could make anything at the moment anyway, I guess

as far as i can understand, whatever your collision is, should have an option to add an OnCollisionEnter event or something of the sorts, in it's blueprint

you really should just make the character strafe while looking forward, rather than rotate when you're walking somewhere
aiming seems like an absolute pain this way

this
Don't blame your incompetence on the engine, kek

Progress on my carrier command game for tablet (will be transitioning my carrier controls / turret / upgrade system and all that to a real game once I finish this one though, using a simple game to build a framework for a larger one)

it looks very boring

Thanks user! I just started a few days ago so I've got lots of time to try and improve it.

I kinda have to second what said.

What do you actually do besides sit and watch after you've placed your turrets? It doesn't even look like there's any strategy involved.

I haven't added that shit yet.

I will add that you shouldn't trust

Vector3 my_vector = get_vector ();
This is ok.

Vector3 my_vector = get_vector ();modify_vector(my_vector);
This is not ok. The struct my_vector is copied when it's passed to the function.

Vector3 my_vector = get_vector ();modify_vector(ref my_vector);
This is ok. The struct my_vector is directly passed to the function.
Though you will need to put a ref in your declaration:
void modify_vector(ref v)

By the way, if you are using a struct in your functions (vector3 or something else), remember that the whole struct is copied if you don't use ref. It can lead to memory leaks, if you have custom big struct.

What will you do, then?

There will be 3 turret types - light, medium, heavy. Light works almost like a TD, defends from fighters and missiles that do damage over time.

Medium can target enemy capitals to do passive damage over time or defend from fighters, inferior to heavies and light at their respective jobs but able to do both.

Heavies are active-targetted towards enemy capitals and can target Hull (pure damage), Engines (can reduce enemy ship speed increasing aim, as well as disabling enemy ships for capture later alongside weapons), Weapons (can disable them to decrease damage taken, but deals less damage) and Critical systems - can crit to deal more damage but deals less otherwise.

The direct gameplay stategy comes from deciding which target to prioritize and which subsystem.

There will also be a power distribution system to buff various aspects you can dictate.

Then obviously the metagame will significantly impact the direct gameplay by impacting Crew (perks), resources for purchasing, and upgrades for turrets.

The main reason I am building this is to have all of the systems in place for carriers to do their thing with minimal control, because they will eventually be put in a 3D space combat game (of which I already have the framework built) where you can directly control fighters while capitals do their thing.

Also the combat phase will be a small part of the game, it's mostly about generating the solar system, and basically all of the pieces of the overworld I need. The overworld will be where most of the gameplay takes place

Vid related - it's what I will be combining with this project so that I have a (semi) open world space combat game with carriers that have basic commands, and direct 3D space combat between them.

Obviously the overworld then becomes the important part so that's what I'm building next, once I have a passable combat system for my mini-game.

You should add some kind of loose aim assist. Like a 15-25 degree cone in front of the player character that auto directs your shots towards the target. Something small but useful for those free camera movements.

Nice, very Vanquish. I like it. Also, get OBS studio. It's superior to free version bandicam in every way and is also free.

wew lad

OBS Studio is open-source. Those things you mentioned already have user made plugins.

PROGRESS!

I extracted data from the feedback people sent, using the small survey at the end of each level. Pic related, the closer the values are to 100%, the better.

(it's for my game God is a Cube, which is a puzzle game with currently 51 levels in the alpha youtube.com/watch?v=3VrRv0nVz-g )

I'm gonna give a few explanations to explain why surveying tools (by asking directly people) are great to make your game better.

Checking the data from the last 6 months, it looks like people really like the levels, except for a few levels too difficult (high red spike) or too easy (low red spike). I created a "satisfaction" indicator (bottom chart) by averaging the fun, interest and right difficulty ratings for each level. The overall satisfaction is really good with an average of 87%.

As you can see, the first 21 levels have less good values, mainly due to the fact that developers find those levels boring.
But I added a panel at the start of the game asking people who already know programming to start at level 22. The next version should have better values, by letting only beginners play this part.

Why the 3 curves? It's really important to have more than one factor to judge the gamers satisfaction.
By example, Level 31 is really difficult (123% difficulty) but people find it interesting (100%) and fun (97%).
On the other hand, Level 41 is even more difficult (135% difficulty) and people find it as interesting (100%), but far less fun (80%). Typically, that's the kind of level where I will add a few runes (the solution will be partly completed) so people can more easily find the solution. It should make the difficulty easier and the fun better, while not impacting the interest rating.

On the values themselves, the interest and fun ratings were a simple boolean (fun / not fun) with average values ranging between 0% (not fun) and 100% (fun). Because it wouldn't bear sense to ask people if a level is "too much interesting" or "too much fun".
On the other hand, the difficulty rating needs to check if levels are too easy or too hard, so there are three values (0, 1 and 2) ranging between 0% (no difficulty at all) and 200% (insanely difficult).
The satisfaction is calculated by taking an average between fun (0-100%), interest (0-100%) and how far difficulty is from 100%. Which means that a 150% difficulty (too hard) will be considered as bad as a 50% difficulty (too easy).

On the values themselves, the interest and fun ratings were a simple boolean (fun / not fun) with average values ranging between 0% (not fun) and 100% (fun). Because it wouldn't bear sense to ask people if a level is "too much interesting" or "too much fun".
On the other hand, the difficulty rating needs to check if levels are too easy or too hard, so there are three values (0, 1 and 2) ranging between 0% (no difficulty at all) and 200% (insanely difficult).
The satisfaction is calculated by taking an average between fun (0-100%), interest (0-100%) and how far difficulty is from 100%. Which means that a 150% difficulty (too hard) will be considered as bad as a 50% difficulty (too easy).

Here is the formulae:
((100 - ABS(difficulty-100)) + interest + fun) / 3

Examples:
- difficulty 50, interest 98, fun 99
( (100 - ABS(50-100)) + 98 + 99 ) / 3
= ( (100 - ABS(-50)) + 98 + 99 ) / 3
= ( (100 - 50) + 98 + 99 ) / 3
= 82.33% satisfaction

- difficulty 125, interest 98, fun 99
( (100 - ABS(125-100)) + 98 + 99 ) / 3
= ( (100 - ABS(25)) + 98 + 99 ) / 3
= ( (100 - 25) + 98 + 99 ) / 3
= 90.66% satisfaction

In conclusion, when working on a game, rather than just asking "do you like it?", it's far much better to ask several questions, especially because you want to know what could be deceiving for your players.
This kind of survey allows you to find what is wrong in a level or a part of your game (the difficulty? the lack of fun? no interest?), while still allowing you to have a way to know if people are globally satisfied.

Note: I would like to say that putting a survey at the end of each level is an idea I took from Infinifactory (by the guy who made SpaceChem), with its author's blessing.

Fuck that, just record with ffmpeg. OBS is just as slow anyway.

Surveys tell you a lot about the kind of people that take surveys.

That's just a few boxes to check at the end of each level. Pic related.
I got feedback for 5 300 levels played this way.
Of course I have more feedback for the first playable level (312 feedback answers) than the last level (13 feedback answers).

Though I should calculate what percentage of people gave feedback, if they stopped giving feedback (but continued playing) at some point, etc.

Like tears in the rain

This reeks of focus testing and other corporate bullshit. You might be making a big mistake here.

Well what's stopping you?

learn ECS, ffs user, it'll save you from future headaches.

Components are attached to game objects (entities in the scene), and are physical in the scene.
Is a joint component physically represented in the scene, and does it use the physics system; hence physically existing?
U betchur ass it is.

Something that's not physical in the scene, for unity, is say scriptable objects. Which aren't required to be physically attached to in-scene entity to run, and are usually data containers or templates which don't physically exist within the scene unlike components attached to entities… which - by definition - has to exist in the scene to be used.
However, they're referenced/instanced via a monobehavior script, but if referenced exist separate to the per scene instanced data (i.e. data isn't erased via a scene reset, if referenced, hence separate from the scene; therefor not a "physical thing" in a scene).

Looks like it could be a lot of fun with polish + other systems you're still working on, as u know of course.
Imo, keep the speed, and general feel of fluidity of motion transitions, but only if that's what you're going for of course.

Loose aim assist, like this user mentioned may be a good idea if it's going to keep the "gotta go fast" feel.
Unless… you do something dark souls esque w/no auto aim (orientation correction after movements); due to the need for rapid + repeated movements, and the inevitable disorientation that occurs (this is similar'ish to what I plan on implementing).

For a game heavily based on puzzle solving I'd say it's an outlier case where this is warranted, and should even be encouraged.
As that user isn't changing the core mechanics, but I'd assume they're just altering the layout of the levels if somethings too hard too soon (which could be implemented in a later level instead, if it's too difficult f.e.).
Also, what types of challenges do people find the most fun, and hence what mix of challenge equates to not just the most interest, but to the maxima of fun; etc etc.

Same, been working and attending uni.
Though, I've been working on the theoretical parts of my game mechanics (core system that makes a lot of mechanics possible); that I'll be diving into once I finish up the current system I'm working on (ECS style, so a complete system used by entities w/their own components specific to this system).
Any progress is good progress imo.

Yeah, the purpose of this survey is to correct mistakes made when designing levels. The first feedback batch helped me to see several cases of a too hard level followed by a too easy level. In this case, just switching the two levels made the two difficulty pikes disappear.

It also helps a lot to see when there is an unnatural difficulty. By example, people didn't understand clearly a specific gameplay mechanic, so I added a tutorial panel at the beginning of those levels. And the difficulty pikes disappeared.

The hardest levels to correct are the ones deemed too easy and not interesting at the same time. I have to see if those levels are really necessary, if people think they have less interest.
("less interesting" is 80%+ interesting, but it's still a low compared to other levels)
By example, Level 49 was designed to look good and I placed it towards the end because it's big. But this level is far too easy at this point, so I will surely put it sooner.

Those are the kind of modifications I'm doing following the survey. I'm changing the game itself when I have human feedback (such as changing controls, adding features, etc).

I don't even know, I just sit on my pc and shitpost all fucking day.
The only thing that I have the passion for now is hating other things, nothing gets me madder than people thinking that RPG Maker is a legitimate engine, or when people praise shit games.

If only I was making any progress

Makes sense, and smart planning.

As long as you stick to your vision, imo, that's legit.

Do shit nigger, be dedicated, and push yourself to actually do something, and I mean anything game dev related; especially if you're not motivated.
Dedication is key here, and trumps motivation by leagues (as motivation is born by dedication, not the other way around).
You can do any number of things… planning, refining your design doc, making pseudo algorithms for your planned mechanics, reading a game dev related book (programming, arting, etc etc), and anything that at least feels like progress.
Then, once you feel confident, and that dedication has solidified into motivation; then work on your game.
Also, having a solid "vision" of it in your mind helps a lot, and helps form said dedication to your vision/dream after enough time solidifying it in your mind's eye.

Man, I'm kind of reading a programming book right now, but I've been stuck on the same page for 2 weeks now.
I just went on it wrongly, I did all the exercises for the easy stuff that I could pick up in minutes(syntax) and now I'm struggling to continue on to reach the actually hard parts(memory management)

Have you ever made a game in a garbage collected language? If not, you might consider starting there. Love2d is a good option that lets you do everything yourself but lua is a very nice and simple language.

And by "lets you do everything yourself" I mean you'll put together a game file by file (not garbage collection/memory management) Lua is a nice scripting language, and definitely not just for beginners.

Well, if the book method isn't working, then go with a different method of learning.
In my experience, an online course with a professional instructor helped me get over that initial learning curve of programming (coursera, linda… u can get some of these via torrents).
I had a difficult time even then, u know wanting to play vidya or shitpost, but with set sessions where I could divy up my time while doing at least one session a day I knew I would eventually get through everything.
So, dedication, scheduling, not overwhelming yourself (pace yourself, and what topics you will broach… i.e. a professional instructor will implicitly do this for u), and pushing yourself to at least make progress daily is what will help here.
Then, with dedication + a schedule, it's only a matter of time before you finish at least completing one of your goals (like finishing an intro to programming course).

TL;DR
Find what works for you, test a few approaches to learning.
Then, be dedicated to a schedule, and then you will complete one of your goals in due time.

how the fuck do I enable those plugins? I see it in the plugins folder at the source that it has OBS-ALSA. Yet for unknown reason it doesn't even detect 1 fucking audio devices.

It won't let me select anything it just shows "Disabled"

While mumble detected fuckloads of devices.

/agdg/, I'm going to be productive for once.

After trying to edit some MUGEN sprites for the first time, I've decided to try out making sprite arts and (eventually) WILL try to make full-blown sprite porn.

Help a brother out, are there any good documentations online for sprite art ? What about the good tools ? I want to make lewd dot arts as soon as possible.

I agree. It's just that I took the strict definition of a Unity component. They aren't physical themselves (except the transform component), but have to be attached to a GameObject which is (since every GameObject has to have a transform). Let's not turn this into a pointless debate about semantics, since the outcome won't change. Whether you say that components are physical themselves or that all of them are physical by extension, doesn't matter, so in the end you're right.

user, I am very familiar with Unity's component system, believe me. All the workarounds I had to come up with over time. All the bullshit I had to deal with.

What the fuck is going on here? Why am I getting a new ID every few hours? Is this some kind of new feature?

i got another question related to discreet math
"how many bool functions are there, with 5 arguments that save the constant 0 (functions that return 0 if their arguments are f(0,0,0,0,0), not exactly sure how this kind of a function translates to english)"
i know that the answer should be 2^31, but i'm not quite sure why
as far as i can see, i get that by doing 2^((2^5)-1)
the first 2 is because we have a possible result of {0,1}
the second 2 is because each argument can be {0,1}
5 because we have 5 arguments
but why do we subtract 1? that's somehow supposed to be related to the fact that the function saves the constant 0, but i don't get why

oh, wait i got it, i'm subtracting 1, because any function that has the arguments (0,0,0,0,0) will always have the same answer, while with other arguments it might not

Alright, so, update on the whole "Looking for beginners who want to get some experience working as a team" thing.

Idea 1 would be a 3D survival horror game with a hyper-stereotypical 80's slasher feel. When I say "survival horror", be aware that I mean Resident Evil and not fucking $2 Unity games.

Idea 2 is a more standard 2D turn-based strategy RPG a-la Fire Emblem or Ogre Battle.

If anyone's a relative beginner and either of those ideas sound enterprising, then please get ahold of me. I'd love to build a ragtag team of high-functioning autists for a project.

i don't want to participate, but please do idea 2

I would just like to say your game is cool as fuck and i hope it turns out well

I would pay for a cool puzzle game like yours

Okay so im working on a model in blender, but when I go to the side view to move the planes at the appropriate distance they wont move on the z-axis. What did I do wrong?

This is a new version of my game, what do you guys think?

mega.nz/#!BFwh3ABI!14VL1g_DnUjx3Pr7HQPJJt8Up0A2qVkWwgkF_wT7QE0

vid related is the last gameplay video I made, but the level being shown has been changed since then.

neato. I will check it out soon.

btw do you have any plans for modding support?

This is an ametur engine, and so the only way to mod it is through digging around in the source code- also the map editor is very badly made, and crashes a lot. It only has support for rectangular prisms. The main bottleneck with development will be the map editor, because it is extremely crude and limited.

I wouldn't recommend using my engine to make mods, because basically any other engine could do what i'm doing better.. like some of the fan-upgraded doom engine variants.

I will release the code when i'm done with the game, so if you really really want to make a mod of my game you can.

When im done with this game, I plan to go right on to making a better engine, because I have learned a lot about engines from this project.

oh alright, cool very cool.

Psst, new bread