859f6e No.14794392
"Marching cubes" edition
Resources
>>>/agdg/
>>>/vm/
>#8/agdg/ via irc.rizon.net
>Dev resources: http://8agdg.wikidot.com/resources
Links
>Wiki: http://8agdg.wikidot.com/
>Beginner's guide: >>>/agdg/29080
>Previous thread: >>14771391
Announcements
>QUARTERLY DEMO DAY SCHEDULED FOR August 8th
>Please contribute to the wiki if you can
0b4b57 No.14794404
Why marching cubes? DC is superior in every way.
859f6e No.14794407
>>14794404
I don't know, I skimmed the last part of the thread for memes and saw anon was working with his cube mesh shit
0b4b57 No.14794427
>>14794407
If you're talking about my project, I'm using dual contouring. The difference is that DC supports sharp edges, dynamic LOD and doesn't need a faggy ass lookup table.
sorry if i'm coming off a bit autistic, it wasn't my intention. Thanks for making a new thread
0b4b57 No.14794648
here, this should explain things pretty well
ff43d7 No.14794650
>>14794427
>No LUT
>An advantage
With how expensive DC is, you'll fucking wish you could just do a LUT lookup.
Besides, we all know naive surface nets is the approach.
3c17f8 No.14794668
>>14794648
>>14794427
can we actually get two pictures, one for each one of the algorithm being used in the same dataset? Comparison is impossible otherwise
fa96cd No.14794715
>>14794708
>not marching tetrahedrons
look at these smooth fucking curves
859f6e No.14794721
>>14794708
MC is the deformed shitty one that has degenerate cases where the planes intersect?
0b4b57 No.14794722
>>14794650
maybe you should stop using virgin hardware
just kidding. it really is pretty fucking expensive and i wish I had thought of that. scrambling pretty hard to opitimize, especially looking for a replacement for QEF
>>14794715
i personally use sprinting tesseracts, but i wouldn't expect someone like you to have heard of that exceptionally advanced algorithm
ff43d7 No.14794726
>>14794715
The fuck is going on with those slopes?
0b4b57 No.14794729
>>14794721
yes MC is basically a joke because a lot of newbie programmers just copy/paste its' code and show off. It's a valid algorithm, all jokes aside but it's probably the most copy/pasted piece of code in the universe and it's also outdated beyond belief.
fa96cd No.14794731
>>14794726
I never got around to modifying the formula to fix that shit
3c17f8 No.14794742
>>14794708
the sphere one looks better
0b4b57 No.14794745
d3c1f9 No.14794793
Redid my first level partially and added voice lines. Think I'm going to have to leave it for now and start work on localization.
7c5d9d No.14794804
>>14794793
>localization
Good goy! Wouldn't want your game to be as poorly translated as Persona 5, would you? :^)
d3c1f9 No.14794816
>>14794804
What the fuck are you on about?
7c5d9d No.14794839
>>14794816
It was a joke, anon. Localizations are the liberal SJW version of game translations, where they force politics and memes into their games. Remember Xenoblade X and nu-Fire Emblem? That's a localization.
0b4b57 No.14794875
>>14794839
Thanks for putting that distinction into my head. Translation/localization difference has never actually popped up in my brain before.
d3c1f9 No.14794882
>>14794839
Right, my bad then. I guess implementing translations would be the proper way to express myself, that's all I'm doing for now. If I could it would be cool to localize the set of math problems generated per grade, but I think that's way too much work for one man.
859f6e No.14794922
>>14794882
>Hopoo will never make that skeleton necromancer game
7c5d9d No.14795211
>>14795051
I've mentioned bloat, and other discontent with the three big engines in previous threads, but there's three other reasons why I'm working on my own.
First, it's a phenomenal learning experience. I'll learn quite a bit along the way, and that extends beyond just gamedev to computer programming of all sorts.
Second, I can't just say "this isn't perfect, but I can't really work around it, it's an integral flaw of the engine!" If I fuck up, I know it's my fault, and this should result in far better performance in the final product.
And lastly, this is a personal point, but my work's quality will only be attributed to me. No one can say "this game looks amazing BECAUSE it's running in Unreal 4", and such other bullshit, which pisses me off perhaps more than it should and I hear it far too often. But this way, no one can pull that shit.
3c17f8 No.14795247
I've managed to fix my shader, the issue was so obvious I don't know why I didn't think of it before.
I JUST HAD TO FUCKING ADD 55% OF THE ALPHA VALUE TO THE GREEN/BLUE TO GET WORKING NORMALS
What the fuck is wrong with normal maps? At least now it works
Now I have to solve the lines that appear thanks to mipmapping non-power-of-two texture atlases. I can either expand the per-texture-border, so that it only appears when it is too far to see or shrink the textures so that the texture+border is a multiple of power of 2
859f6e No.14795249
I'm still dicking around with random helper functions, so I can do point picking on arbitrary geometry
In order to point pick a triangle, we need to pick one vertex as the root, and then construct vectors U and V towards the other vertices, such that U+V+W=1 (with W being inferred from the sum of the other two values), basically a barycentric algorithm.
I can do this just fine with pic related, but it needs to go through the loop an average of two times per iteration. I can't figure out a faster or deterministic way to do so, since things no longer have uniform distribution (eg if scaled). How can I fix this?
859f6e No.14795265
>>14795247
>What the fuck is wrong with normal maps? At least now it works
I forget the exact math, but I believe the vector (0,128,128) normalizes into the Z axis pointing straight towards the camera on an orthagonal plane, which is why the default normal color looks like pic related. Or something, I can't remember the details
0b4b57 No.14795341
Hey guys, would you mind helping me out? I'm trying to understand the method described at 4.2.1 to approximate a point on the isosurface, but I'm having a hard time understand the bit where they calculate the 8 forces at the vertices of the cell. Would you mind helping me a little bit by typing out the meaning of the formula describing the points in pseudocode? As I've said in the previous thread, my current understanding of written math is rather poor, despite being able to follow the concepts rather well when ingesting them through illustrations and other demos.
http://www.inf.ufrgs.br/~comba/papers/thesis/diss-leonardo.pdf
The entire paper is actually thematic to the thread, since it's pretty much all about optimizations of contouring algorithms, so I hope you find it useful for yourselves, as well!
e7e839 No.14795354
What's your reason for using Godot, /agdg/? Me, I use it because Godot is the only engine that plays nice with Linux.
ff43d7 No.14795375
>>14795341
That big ∑ is summation. https://en.wikipedia.org/wiki/Summation
Basically:
n
∑ Pi
i=1
Translates to
float sum = 0.f;
for(unsigned i = 1; i<n; i++) {
sum += P[i];
}
// sum is what the "∑" summation operator returns
The rest should be closer to usual vector math (dot, cross, scalar), although I have no idea what that comma is supposed to mean.
859f6e No.14795378
>>14795249
https://math.stackexchange.com/questions/538458/triangle-point-picking-in-3d
Nevermind, found it out. Rather than discarding, the trick is to inverse the picked points if they exceed 1. Doing this made it from 20-40% faster on average
ff43d7 No.14795379
>>14795375
wups should've been i<=n, not i<n
5d2e4c No.14795381
>>14795195
I love Mr Bean he is so funny.
>>14795354
Unity plays just as well on Linux and is a more robust and feature full engine.
98a3e3 No.14795386
I just want to confirm, selling 3D assets is bad?
7c5d9d No.14795414
>>14795386
No, and if you can make money off of retards, then more power to you.
Mostly, buying asset packs is what's scoffed at, and even then, if they actually look good and not like Nu-lowpoly, most people don't care. Whether you hire an artist or but a pack, you're paying for the same thing either way. Though, if you can make due with just premade asset packs, I'm not sure how original your game really is. Personally, I prefer having my own style and doing everything myself. It might be average, or bad, but it will be mine.
3e0653 No.14795425
>>14795386
only if you overcharge for shit
98a3e3 No.14795426
>>14795414
A few threads back people seemed really upset when I commented on how my current 3D project would hopefully be sold at an asset store or two.
859f6e No.14795436
>>14795425
>Make generic furry lowpoly chibi model with built-in facial animations and swappable textures/patterns and have them all use the same model and animation, but as separate "characters"
>Have them dork about a generic grassy hillside as a preview video
>Sell for $50 because of the perceived "quality"
>infinite money to actually dev
98a3e3 No.14795453
>>14795425
For that matter, how do you know is too expensive? specially when it is a textured and rigged model or the like?
0b4b57 No.14795473
>>14795426
There are folks on here to whom the whole premise of selling anything means you're a kike. I've seen people unironically mad about a yesdev on here selling his game on Steam. If you make an asset and sell it, and it sells - you should be proud of yourself. It's honest business.
>>14795375
Thank you very much, anon. That's one mystery solved. Hopefully someone will have an idea as to what that comma is meant to be.
98a3e3 No.14795510
>>14795473
I see, is kind of hard to look for feedback when some people who are very vocal give you the most negative feedback, be it your particular artistic direction or the fact that you want to profit from your work, sometimes when the stuff is not of their liking despite the artist only doing what they were commissioned to do.
a8ca51 No.14795528
>>14795354
Some anon here did a much better Godot-chan
0b4b57 No.14795546
>>14795510
It's a bit of a popular meme to say these days, but negative feedback inherently sticks in your mind better than positive. It's why you should be extremely careful when ingesting it. You must, obviously - but you should absorb it in very measured doses.
>>14795528
That's much better. The other one looks like she smells like piss and gluesticks.
91efcc No.14795558
>>14795381
>Unity plays just as well on Linux
Unity's Linux editor and OpenGL support are an afterthought.
a8ca51 No.14795606
>>14795546
That's the official design, though looking like an 8 year old isn't.
5d2e4c No.14795626
>>14795558
That hardly matters. It works very well and I've been using it for months. In fact the only problem that I've had with doesn't have anything to do with the graphics. It's just code completion with Visual Code.
>>14795386
The only guilt you would get is fueling the asset flippers that never get anywhere with their games.
98a3e3 No.14795634
Well, here is my faux facehugger which was actually planned to be a giant "my innards molest people" thingy, it will be quite versatile once finished.
>>14795626
When you have something more or less unique it will heavily contrast with all the other stuff they bought, assuming they bought it because we know asset flippers only use free assets.
c0d2f4 No.14795636
0b4b57 No.14795662
>>14795606
it might be official but I stand by my original statement.
3c17f8 No.14795867
TLDR: Anyone here have any idea how to access the current mip level in fragment shader? (Unity)
The issue: At distance, the mipmap is making me take part of the normal texture as albedo and creating pink lines
Current Atlas technique: I take a texture, replicate a 1 pixel wide strip and use it as border, and reduce the texture to texSize-1 (if the texture size is 1024, it becomes 1023 + 1px border. Then I read half a pixel of the bottom left, and half a pixel of the bottom right.
Why the pink happens: At smaller(bigger? dunno) mip levels, half a pixel becomes 7/8, or 15/16 and so on, and bilinear/trilinear filtering will make me get part of that.
My solution would be to change the half pixel thingy to less and less (in the picture, making the blue square closer and closer to the center of the yellow square), so that the filtering/float error never takes a pink pixel
016fc2 No.14795898
>>14795662
Yeah the official one is crap in comparison. I'm the anon you're replying to I just changed ip
5d2e4c No.14795969
>>14795867
Why are you using texture atlases in Unity? Or rather, why are you mixing normal maps with normal albedo maps? That partly defeats the purpose of saving texture space because now your textures won't receive their respective texture compressions (DXT5 for albedo and NORMAL for normal maps).
7c5d9d No.14796090
>>14796043
Does this count?
4cea06 No.14796129
3c17f8 No.14796131
>>14795969
because I can't make textures marked as normal maps inside other assets
However, I can use two colors for normal, two for shiny/metal
and on the albedo, remove alpha and add height.
So it will probably be almost as good
9df1ff No.14796159
>>14796090
Ey yo bitch wat dat mouf do
98a3e3 No.14796160
I am beginning to consider making some low poly stuff but instead of anime style I think more like these dolls, well, more or less.
7c5d9d No.14796212
>>14796129
Sweet.
>>14796159
Aww, you made her sad with your nigger speak.
98a3e3 No.14796255
>>14796212
Speaking of these, is there any known process where you animate only the parts you want like the eyes or the mouth without using a whole set of rotating textures? I mean, having the whole face texture changed with tons of variants to cover it up.
c0d2f4 No.14796265
>>14796255
make the eyes and moth seperate textures?
98a3e3 No.14796295
>>14796265
So basically different materials in those faces? OK.
0b4b57 No.14796299
HookTube embed. Click on thumbnail to play.
fuck it im rewriting my fucking octrees this shit's making me mad
a2e179 No.14796406
>>14796376
Is this in UE4? With a little more texture work, that would be baller as fuck to use in a shooter.
98a3e3 No.14796435
>>14796406
>>14796376
Looks like Hammer/Source.
0b4b57 No.14796436
>>14796376
>tfw no fancy texture tricks needed because your entire world is actually deformable and persistent
I'm just being an asshole, anon. Excellent work, it looks really good.
a51a1d No.14796478
>>14796406
Unity. I'm using the screen space boolean shader from the previous thread.
>>14796436
different uses, I guess. and I'm pretty sure I wouldn't be able to achieve what you are making
>>14796435
>Looks like Source.
don't wanna get banned
bbb589 No.14796488
>>14795381
>Unity plays just as well on Linux
bull fucking shit
>is a more robust and feature full engine.
you mean a more robust and feature full store.
a2e179 No.14796500
>>14796478
For what its worth, this is a great thing to have in a game. One thing a lot of games fail to due is convey the power of weapons through showing what they do to the environment. A miss that strikes something unintentionally should show just as much power to the player as the gory results of a shot that connects.
In conjunction with good animations and audio, you give the player a feeling of power, which results in more of a feeling of agency, and generally contributes to the overall quality of gameplay.
Keep up the good work dude.
0b4b57 No.14796513
>>14796478
Anon, if I can achieve what I'm making - anyone else can.
I'm pretty sure I'm like learning disabled and every single thing I've done so far for my game was achieved by me pestering actually competent programmers to explain various concepts to me. The only credit I'm willing to take is persistence and passion.
The rest is all on the beautiful motherfuckers of /agdg/ and Google. I love the lot of you.
98a3e3 No.14796519
>>14796478
You can get banned for working on source engine?
5ad6b2 No.14796526
>>14796519
No, it's just a meme.
d47d2e No.14796596
>well, let's try this whole animation thing
>I'm a programmer, not an artist, but this can't possibly be that hard, right?
It really doesn't help that the entire body is divided into multiple pieces so they can be chopped off or replaced by something else and also act as items when they fall down, ready to be attached to someone else. The autistic limb-base inventory system is working, now I just gotta convey it visually in just 3 dimensions instead of how many this horror spawn manifests in.
1cba88 No.14796604
What programming language should I learn?
I know next to nothing about technology.
I don't want to learn to code just for games, the more versatile it is, the better.
I've heard that Java is not viable anymore, and c++ is not very efficent in Windows.
Just hoped I could get some pointers.
54caa2 No.14796611
Does anyone know a good guide for learning how to code in Ruby? My game uses RPG Maker VXA for the engine and it'd be really helpful if I could just write scripts myself for features I want.
I'd also appreciate guides for 8-bit spriting and music, because I'll be damned if I use any default assests.
0b4b57 No.14796616
>>14796604
Pick an engine and learn the engine. You'll learn most programming concepts in the process. I'd recommend doing UE4 but I'm biased and the syntax might be confusing to a beginner. So, you may want to try Unity I guess. It'll teach you C# which is basically babby-tier C++.
be1bdd No.14796620
>>14796604
>I don't want to learn to code just for games
Read SICP, you can find it for free on the Internet.
859f6e No.14796624
1cba88 No.14796646
>>14796620
I've been putting it on hold for a while now, I'm gonna get back into it.
Does it teach a specific language?
>>14796616
Thanks for the reply, but I'll hold off on fiddling with engines for now.
I'd rather have a solid grasp on the basics, since I've got no fucking idea of what I'm doing.
0b4b57 No.14796651
>>14796646
Oh, I understand but you'd have a much easier and more fun time getting results directly from the game as far as code is concerned. It really does give you a much more hands-on idea of what your code is doing.
7a4d82 No.14796656
>>14794708
This actually a really unfair comparison, because they have different grid resolutions.
It appears here that the MC grid's resolution min size is much larger than what DC grid dynamically subdivides to (giving the MC sphere a messy/blocky appearance).
If the grid resolution was lowered to match the min DC octant size, then the MC sphere would still look a bit more jagged, but it would conform to the contours of the sphere; as compared to the blocky low-res mess it is in the video.
Though, obviously DC has the edge in almost any surface contouring due to dynamic subdivision/octant size, and the dynamic vertex determination instead of a LUT; while being able to actually conform to the isosurface unlike the MC approach.
>>14794715
Definitely a superior primal contouring algorithm compared to MC in terms of appearance/applications, but not in speed (for a naive implementation that is).
>>14795247
Make sure you do "decode normals" in your shader, else they're incorrect normals.
>mipmapping non-power-of-two texture atlases
make sure they're power of 2, don't try to reinvent the wheel here.
>>14795867
>access the current mip level in fragment shader
yeah, calculate the current mip level
<https://www.opengl.org/discussion_boards/showthread.php/177520-Mipmap-level-calculation-using-dFdx-dFdy
then use tex2Dlod to access the current mip level
<http://developer.download.nvidia.com/cg/tex2Dlod.html
>>14795341
There's a few implementations online if you do a search (source code on github).
Also that nick DC guy I linked to awhile back uses a particle decomposition approach in one of his blog posts.
>>14796376
Looks amazing anon, nice.
Creative application of the ss boolean.
>>14796436
The only issue with small deformations like that is rapidly recontouring the surface, and the depth at which you need to refine the octree as to define such small surface features.
Therefor you need realtime contouring of the entire chunk you're working with; at a high LoD.
That in itself is quite difficult, and also has the same results of a SS boolean if you don't update the colliders; otherwise doing it via contouring is only necessary if you need to update the colliders themselves.
0b4b57 No.14796677
>>14796656
Thanks, anon. I'm gonna look through his blog posts.
859f6e No.14796724
>Make INode interface with only links to parent and children
>Make extension methods that support queries, like GetRoot() or a.IsDescendentOf(b)
>No idea how to visualize it, shit out raw text and put it into a website that visualizes it for me (pic and link related) https://pastebin.com/raw/rs8GRSkK
I mean, I guess it works? I need a more concrete add/remove parent method, though. Not sure how to guarantee or find the root without it looping and crashing. I can use a list to track visited nodes by reference, but this seems cumbersome
https://hastebin.com/raw/ojunegaxul
7a4d82 No.14796762
>contouring thread
i have a webm a made a few days ago after reorganizing my DC source code.
<warning: my recording software amplified the bass of the music i was listening to
For reference the highest LoD chunks (the small ones in the middle) goes to a minimum resolution of 1/4 units in unity; in addition to contouring in real time if I have the data organized (actual contouring of a cell w/max detail takes ~1 to 2ms per chunk w/old approach shown in video, and ~.3ms to 1ms per chunk with new approach).
Also I ended up deciding to completely redo the front-end systems like the loading system, so how I'll be loading chunks, and how I store/load the data. As it's a naive approach to just demo it atm, and it'll add a lot more speed w/my new refined approach.
859f6e No.14796793
>>14796724
Oh I see, I confused the requirements of trees/graphs together.
69ffec No.14797002
>>14795211
>my work's quality will only be attributed to me. No one can say "this game looks amazing BECAUSE it's running in Unreal 4"
No one's going to say this. People might say "wow, he did all the art/music, etc". They're not gonna give a shit about how it was made. Didn't ZUN make his own engine for TH6+? How many people do you see praising him for that compared to everything else about the games? I mean if you find it fun then fine, I just hate seeing anons waste their time when they don't have to.
0b4b57 No.14797129
>>14795211
imo developing your own engine is a phenomenal learning experience but developing your own engine when you're alone for the sake of making your own game is
OPINION WARNING
unrealistic and overly ambitious.
>And lastly, this is a personal point, but my work's quality will only be attributed to me. No one can say "this game looks amazing BECAUSE it's running in Unreal 4", and such other bullshit, which pisses me off perhaps more than it should and I hear it far too often. But this way, no one can pull that shit.
only an absolute retard would say this sort of thing, I think.
7cdbdb No.14797281
>>14796646
No; it uses a simple version of Scheme, but what it teaches is universal. On that note, I wouldn't recommend Scheme as a language for any big project. You end up having to implement tons of simple things yourself that other languages and other Lisps like Common Lisp already have.
56f3ff No.14797547
>>14796478
Someone should put that shader on the wiki because I'm too retarded to figure out how to find deleted 8chan threads
5ad6b2 No.14797662
HookTube embed. Click on thumbnail to play.
>>14797547
This wiki page stores the archives, but they are not complete.
http://8agdg.wikidot.com/agdg
adb70d No.14797725
I decided to stop working on fixing my BSP compiler in the way I was doing it and instead to implement CSG, maybe I said this in an earlier thread but I am making a lot of great progress now that I understand what it is all about, especially because I can use all of the helper functions in my BSP compiler to do CSG with! So I am no longer hitting a wall with my problems, and soon it will all be fixed and everything will be great. I really wish that I had just started working on CSG sooner, but I am glad to be making progress again, because I really want to finish this compiler so that I can optimize my rendering engine, which is pretty unoptimized in the way it is written now. It's written in a more OOP way instead of DOD, because of what I was trying to do last summer, when I wrote the fundamental systems for it.
>>14795211
Writing your own engine is fine, as long as you are enjoying yourself with the work you are doing, and making progress, you shouldn't listen to people who don't understand it. What kind of work have you done on your engine so far?
3068e2 No.14797769
Anyone know the Pro's / Cons of ARPG's, JRPG's, and Metroidvanias?
56f3ff No.14797770
>>14795211
>"this game looks amazing BECAUSE it's running in Unreal 4"
Disregard these kinds of children, they're idiots who don't know what they're talking about. 99% of what makes a game "look good" is the effort put in by the developer, not the physical limitations of the engine.
5d2e4c No.14797826
About 70% done with the Shinobu house. Need to furnish it and maybe do some more exterior designing. Blender Ivy gen is meh but the leaves are useful and can be used in combination with plant root cards. Using some leftover furniture props from the wedding project I never got around to using.
859f6e No.14798014
I'll take a bold leap of faith here and assume that trees cannot have two roots
859f6e No.14798058
>>14798014
Actually here's a good question. Let's say you wanted to add a link from 4->0 via a function like n4.Add(n0). This is an invalid operation, what would happen?
>Nothing, it just fails to do anything
>It raises an exception (circular dependency)
>It connects 4 to 0, creating a loop
>It connects 4 to 0, breaking 3 to 4
>It connects 4 to 0, breaking 0 to 1
bdecc2 No.14798100
>>14798058
There's another option
>It connects 4 to 0, summoning Satan
87fb0a No.14798229
>>14798014
What are you trying to do with this tree?
7cdbdb No.14798298
>>14798058
If cycles are not allowed, checking for them and signaling an error when they happen is the right thing, fuck everyone who thinks silent errors or quietly doing something subtly wrong is okay. To check whether a non-cyclic graph becomes cyclic after adding an edge A -> B, it suffices to check whether A is reachable from B, and since the graph is non-cyclic this is doable rather efficiently, too.
d9ae5d No.14798348
>>14796596
>i'm a programmer, not an artist, but this can't possibly be that hard, right?
1226c7 No.14798354
>>14796604
>I've heard that Java is not viable anymore
Correct. C# is better as far as managed languages go.
>c++ is not very efficent in Windows
Not so sure about this. Linux is imo better for programming since everything is laid bare. You can install libraries with your package manager and find them in /usr/lib, headers in /usr/include and you also have the coreutils to work with (grep, sed, awk, find, all that stuff), but I'd wager most c++ is probably written in a windows environment. The difference is you do all work in an IDE, so the details of how stuff is done is opaque to you. With linux, you'd write a makefile to build your code (or preferably use something that generates makefiles, like cmake). With windows, you click the build button and pray everything works. With linux, you can investigate problems and fix them. With windows, you're probably screwed. I assume mac and Xcode is similar to windows and visual studio.
1226c7 No.14798429
>>14798014
Correct, since then it would no longer be a tree. It would be a more general graph of some sort.
>>14798058
Decide on what behaviour you want, figure out what constraints are required, then enforce them. These "what would happen" questions applied to abstract descriptions of data structures are meaningless. It's like asking "what would happen if you divide by zero?". You can ask "what happens when I do this with this particular implementation?" and ask whether that is desired behaviour. Each of your options is a potentially valid choice for an implementation if that is the sort of behaviour you want.
35713e No.14798483
>>14796624
Successor actually, gone through multiple iteractions already, a top-down view with free camera is what I'd like to stick with.
Already got the atmospherics working (with liquids as well and any chemical element that is included), part of power lines as well and even the inventory, something that I seriously want to expand from the base game.
The issue with the limbs is precisely because I despise the way inventory works in SS13, with a certain hard-coded slots for your mob, instead of dynamic slots based on what limbs you have. It is especially limiting and one of the reasons alien races are always humanoid shapes or genetics can't really change that very much.
>>14798348
Oh look, un-modded Oblivion!
e9d297 No.14798519
>>14794392
Nice ubuntu wallpaper mate.
>>14798348
The good old "accidental gamedev lovecraftian horror".
5d2e4c No.14798528
>>14798348
I've never seen an animation telegraphed so well before.
8396e6 No.14798584
>>14798348
This is both excellent and hilarious.
Can someone recommend me a decent android tablet? I need to get my game on an actual device so I can see how hard it will shit the bed.
fc6892 No.14798609
>>14797547
>spoiler
>>14794392
>Previous thread: >>14771391
web.archive.org/web/*/https://8ch.net/v/res/ *insert thread number here* .html
which gives you:
https://web.archive.org/web/*/https://8ch.net/v/res/14771391.html
now update the archive
3f91fc No.14798622
Do games built by Unity contain any information about the user that made them? Like metadata in photos and so on.
35713e No.14798645
>>14798622
You can sign projects with your name and your company if you want, to display in the launcher (if you use it), so that probably sticks, but it's voluntary.
There's also an account you use to start Unity unless you're working offline, which might also be capable of connecting you to a project, assuming it's saved as part of it or that your projects are monitored by the Unity team (but both of these things likely don't happen).
3b7dd5 No.14798795
>>14797769
ARPG
>player skill matters
>seamless switching between exploration and combat (most of the time)
>easier to get into
>you can usually jump
<generally lower systems complexity so the core mechanics absolutely need to be great
<no parties or only minor party mechanics (there are exceptions)
JRPG
>menus allow more options in combat
>full party mechanics
>more character interactions
>lots of system variety
<battle scene transitions look cool, but generic battle backgrounds suck (there are exceptions)
<you can't usually jump
Metroidvania
>player skill matters
>powerup-based progression removes the grinding problem
>sequence breaking is pretty cool
>you can jump
<level design is incredibly important and easy to fuck up
<approaching diamond dozen status so don't expect huge sales
Also, regarding marching cubes versus dual contouring: Why don't you build an icosphere and use vectors to draw the terrain?
70cddd No.14798948
>>14797826
Looks like a great place to get kaka'd at in. Isn't she rather big compared to the environment though?
>>14798622
>>14798645
They most likely hide information about your machine and your Unity login, to track you in case you make enough money and don't pay them their one time fee.
0b4b57 No.14799372
Good morning, /agdg/! What are you up to today?
7c5d9d No.14799411
>>14799372
Brushing up coding skills by writing up a simple backup tool for my game files.
Gonna plan out some sort of schedule so that I would stop abandoning my other obligations to just make vidya.
d9ae5d No.14799414
>>14799372
i have done nothing for the past two months other than leaving a neural network on to train for the sake of eceleb porn
there is no god
seeing as we're on topic, i'm stuck because i need a writer for this shit >>14798348
3e0653 No.14799416
d3c1f9 No.14799421
>>14799372
Trying to find a font that covers my needed languages while still not killing the bank.
0b4b57 No.14799430
>>14799411
that's an excellent idea. A schedule is a bit cliche but nevertheless effective way to stay on track, although you gotta make sure you don't make it too strict, otherwise you're just gonna feel bad and tripped up when you inevitably slip and skip an item on it!
>>14799414
im very, very excited to see the results of your eceleb porn project. I may even make a request later down the line if you feel open to taking them :^)
>>14799421
filthy eurofag here, what does "killing the bank" mean? Spending lots of money?
d3c1f9 No.14799443
>>14799430
>Spending lots of money?
Yeah, exactly that. I was actually going to use Comic Sans but it costs fucking $1300 to license. But I'm thinking that I might only need a Desktop License after all, which means I'm down to like $30. I don't know, it's fucking retarded and jews everywhere.
19682a No.14799460
>>14799443
The fuck are you doing m8? There's tons of free fonts you can use.
c7a755 No.14799464
Day 4 of learning code. I have made a window, put a triangle and text inside using DirectX. Now for a basic camera.
I'm building everything from scratch using DX11 and C++. I'm too autistic about optimization and speed to feel okay with using UE4/Unity- not to mention there's MIT code for every major graphics effect floating on Github. It seems really doable until animations and art.
d3a5fd No.14799468
>>14799443
Desktop license means you can install the font and use it on your local computer. You'd need a license that lets you bundle the font file with you game. And some licenses will have terms that only allow for a certain number of installs before needing to pay more. You're better off looking for some sort of free/open font. Just be sure to check what license it uses.
d3c1f9 No.14799475
>>14799460
Sure friendo. Try finding a font that has
>Extended Unicode
>Cyrillic
>Greek
>free
>not shit
>>14799468
Yeah but apparently there's a licensing loophole if you render your font to a bitmap first. I'm not too far into it yet, need to keep reading more legalese, but it might help me.
>>14799464
>day 4 of learning code
>thinks he can out-optimize UE4 and Unity
Ok.
7c5d9d No.14799481
>>14799421
I made my own for a reason.
It took me about ten days, but that's because I was doing a bunch of special characters, the entire cyrilic alphabet, and there was quite a bit of family stuff going on as well. Could have finished it in four days if I was focusing on it more.
And I'm sure there's quite a few fonts that you can use for free. I think there was a site, 1000fonts.com or something of the sort, and you could sort by licenses.
c7a755 No.14799488
>>14799475
There's always DX12 if I don't make it.
19682a No.14799489
>>14799475
>Extended Unicode
I don't know what that involves but since you were talking about comic sans;
https://www.fontsquirrel.com/fonts/comic-relief
d3c1f9 No.14799494
>>14799481
What's it look like?
>>14799488
DX12 is not a magic bullet.
>>14799489
Extended unicode is primarily for nordic characters like ÅÄÖ. That font doesn't support greek sadly enough, thanks anyway.
19682a No.14799504
>>14799494
I don't know shit about greek so I won't argue, but it has greek in the supported languages.
d9ae5d No.14799505
>>14799430
not sure if ironic, but nah i can't take requests, doing this without a gpu so it's taking at least 3-4 times longer than it should
first day it went from 0.1 to 0.06 loss, second day to 0.05, third to 0.04, and now it's pretty much stuck at 0.031
it more or less manages to correctly copy the facial expression, but all the important details are still too blurry to actually recognize an eceleb
2b0b74 No.14799522
>>14799504
Odd, the test online doesn't work with Greek. I'll download it and check later.
19682a No.14799524
>>14799504
Also here's a bunch of fonts that supposedly support it. All the fonts here are free and most of them are basically libre.
https://www.fontsquirrel.com/fonts/list/language/greek
0b4b57 No.14799572
>>14799505
I'm not being ironic, I was genuinely wondering. It's okay, though - see if you iron out the kinks and then report back! Deepfakes are en-route to becoming illegal and so - I would love it if enthusiasts like you continued iterating upon different implementations.
Who knows, maybe you'll come up with a solution for the blur?
d9ae5d No.14799588
>>14799572
i'm not actually deving anything, i only said i'm using one
for the record i'm using myfakeapp, since the regular one only works with nvidia gpus
0b4b57 No.14799598
>>14799588
Ohh, I see! Okay, I'm sorry I misunderstood.
d3c1f9 No.14799771
>>14799489
>>14799504
It actually worked, you're a lifesaver anon. Must be something bugged with their site.
>>14799554
Looks pretty good, nice job.
859f6e No.14799795
>>14798229
>>14798298
>>14798429
For my engine, I want to have a lightweight UI system modelled after winforms (eg controls, forms, labels, buttons, etc). Since UIs are treelike/hierarchical in nature, I needed a good way to represent trees, so I made an INode interface.
I realized after posting last night, that both graphs and trees could be implemented with an adjacency/edge list, which would make implementing references (T parent, T[] children) on the nodes themselves redundant. That is, implementing extension methods like code related might not be a great idea, since I can't guarantee how the public implementation would work. https://hastebin.com/xiwoniweqo.cs
Anyways, to answer that question, I wanted to make a generic handler and query system for trees. Thus, if I only ever set values through Add(), and detect cycles at that time (checking the root for each node?), it should be robust enough if I did it that way, but like I said with the adjacency lists, I'm not sure how to proceed now
c0d2f4 No.14799815
3c17f8 No.14800187
>>14797826
I am not an expert in mexican architecture, but ain't these things supposed to let water flow out of the ceilings?
There should be a hole there
Very nice otherwise
3c17f8 No.14800195
>>14797826
also, she is a big girl
1226c7 No.14800256
>>14799795
One way would be to use append/detach. You create your nodes, then you append them to other nodes, or detach to remove their relationship. Note that detach doesn't delete anything, it simply removes the node from it's parent's child list. This way, I think it might be impossible to create cycles with this interface. Not sure though. To check for cycles, I think it should be sufficient to traverse the node's subtree (i.e. the node's children, their children and so on) and see if you ever reach yourself.
859f6e No.14800306
>>14800256
>To check for cycles, I think it should be sufficient to traverse the node's subtree (i.e. the node's children, their children and so on) and see if you ever reach yourself.
Couldn't you achieve the same thing with checking towards the parents or root? Also, in order to check for a cycle, wouldn't you need a list to store visited nodes, in the case that a cycle exists already? Eg, ABCA, but you attach node D somewhere - it will never find itself, but will still loop
3c17f8 No.14800439
In my atlas, normals become pink for some reason
Why there are no brown (the other texture is mostly brown) horizontal lines, only pink vertical ones?
The precision error should be happening equally on both, right?
At least I can fix by having the albedo+height atlas be separate from the normal+shin+metal one. but it still bothers be that the error is not happening on both directions
fcc654 No.14800442
3c17f8 No.14800480
>>14800439
forget I posted about it, it's just way harder to notice, but you can see if you change the textures to full black/white
d2b3a9 No.14800487
>>14797826
That looks amazing. Wish I could model like that.
915b9c No.14800556
>trying to figure out a proper class based ammo item i.e Ben G. will see a revolver boolit instead of a mag
>some of those ammo item will gibs 2 different ammo due to weapons needing different ones
>in order to prevent the ammo from being picked up when it is full I would need to figure out some hacky way with a_jumpifinventory
>waste several hours trying to use A_KillChildren instead to achieve it just later to figure out the ammo item needs the +ISMONSTER flag for it to work
Yes thank you based ZDoom developers for adding this (((limit)))
80ac97 No.14801502
>>14799414
You are a writer without peer, I fucking love this dialogue.
That being said, if you're serious, I've written for a couple games that never took off due to unmotivated piece of shit high school coders. Going to bed now, will check for replies later though.
d384ee No.14801552
>>14798948
>>14800195
The outer arches are too low. I have to fix them but raising them a bit. But the rest of the house has realistic measurements: 8+ feet walls, 6'8" doorways.
>>114800187
Yes there should be some hole there.
3afc24 No.14801582
Does anyone develop with older versions of Visual Studio? I've tried 2015 in the past and one of the required updates to Windows 7 causes weird OS-wide performance issues every time. 2010 seems to be the last unpozzed version but I'm not happy about the lack of basic C99 support.
adb70d No.14801688
>>14801582
I used 2003 to write mods for goldsrc at one point, I really liked the aesthetics of it but like you said I did not like the lack of C99 support.
46351a No.14801972
I know I said I'd stop a long time ago, but I can't stop jerking around with crts. I frankenstein'd a pair of old CRTs I had to run HDMI into em and I tried playin my game on it to see how it looked and to try and improve the optional shader a little bit. It's a lot of fun but I think I've basically hit the cap on how close I can get it with my understanding of this stuff because now I'm starting to go beyond just a simple shader into a bunch of separate post-process scripts. I might try to get that sort of luminous "blur" effect, but that's probably the end of it.
Anyways it was pretty fun playing it on the old tv's even though the game's resolution was too different to fit into the whole thing without a bunch of stretching or compressing. One of the TVs was a flat panel and the other was curved.
I'm almost done with the fifth stage at this point and I've already done most of the art for the sixth one so everything's coming along really nicely. My music guys say they're getting ready to start submitting some trial stuff too so I'm really excited for that.
adb70d No.14802171
>>14801972
If you're interested, this youtube channel has a pretty detailed explanation of how CRT's form an image, which might help.
https://www.youtube.com/channel/UCy0tKL1T7wFoYcxCe0xjN6Q
46351a No.14802360
>>14802171
This is great stuff, thanks dude. I mean, its kinda a waste of my dev time lmao but it's fun to play with.
5d2e4c No.14803427
In case you guys are curious how I make my own seamless textures. It's amazing how jewish Google Images has gotten these last few years. Stockphoto watermarks everywhere. Not anymore!
1.) Make a throwaway email here https://temp-mail.org/en/
2.) Get http://www.the-orange-box.com/
3.) Get http://www.the-orange-box.com/portfolio_page/free-seamless-texture/
4.) Follow my video instructions
fc6892 No.14803770
>>14801972
mmmm… this pleases me to no end.
Still, how did you manage to run HDMI into the CRTs? Are the pixels displayed on them accurate? NTSC? PAL? I'd imagine they are shadow mask based things, as that was the standard for TVs
What I'd do is get an older graphics card with S-video out and try to run the game on that. But I'm not 100% sure a Unity game would run well on such a set up.
fc6892 No.14804016
>>14803427
so it's automated, huh? I usually used a mix of effects -> offset, clone tool, mirroring, duplication and eraser myself. Will have to try this out, wonder how good the results will be
26783a No.14804410
8396e6 No.14804447
>>14804410
Just like.
Decided on my fonts, finally. I can hopefully have my entire game translated by tonight, then I'm just going to polish up until Nordic Game Conference, then I'll work on new content.
19682a No.14804731
Are there resources for getting gud at geometric math and matrices and shit? I want to be able to calculate rotations and transformations and find positions and intersections from them. Where should I start/look at?
8396e6 No.14804758
3c17f8 No.14805038
>>14803427
Gimp also have a 'make seamless' script thingy, but I don't think it have that "delete this texture feature but make it seem it had never been here before" magic used at the 35 sec mark
70cddd No.14805088
88fcf8 No.14805105
>>14803427
You can try substance designer for free for a month.
1226c7 No.14805117
>>14800306
>Couldn't you achieve the same thing with checking towards the parents or root?
Yes, but I assume it's more efficient to check below. The way I interpret "check towards root" is to basically traverse the entire tree except the appended node's subtree. Unless attaching a large subtree to a small tree is common, it should be more efficient to go below, since by ignoring the rest of the tree, you've cut down the node count to log(n).
>Also, in order to check for a cycle, wouldn't you need a list to store visited nodes
There are multiple ways of doing it.
One is a set of visited nodes. For each node you visit, check if it's contained within the set. There are many choices for the structure for storing visited nodes. A stack is one, which would require linear searching for membership test. Since you need to do a search for each node you visit, the complexity of lookup is O(n²). Another is a hashset, where everything is O(1), but beware reallocations and cache thrashing.
Another option is to let each note have a visited flag that you set when traversing and clear when unwinding. This way it's simple: if you visit a node with the flag set, cycle! You could even make the flag set external and make clearing a simple memset. This is easy to do if the nodes are stored in a vector instead of being plain heap-allocated. Then the bit index just corresponds to an index in the node array.
>in the case that a cycle exists already
Do you expect this to occur? The answer to this question is the answer to "do you intent to load the tree from an untrusted source, such as a serialization format, a config file, over the network etc?". If the structure is entirely built in-memory, as long as your mutating functions cannot create cycles, your tree can never reach an invalid state and you don't have to worry about existing cycles, only about new ones. Ofc, you can still add a checkForCycles() in a bunch of places during development to help yourself find bugs. In fact, I'd strongly recommend adding ways of detecthing things you don't expect to happen during development, like nodes becoming orphaned, cycles being created and any other kinds of illegal events. Really helps debugging.
Anyway, I gave the append, detach model some more thought and I came up with an edge case. Consider the following:
A=Node()
B=Node()
B.attach(A) // B becomes child of A
A.attach(B) // vice versa
This is a cycle. It occurs because a parent gets attached to one of it's children (unf). However, let's say this is a subtree we're building for attaching to the real root R.
A.attach(R)
Now, A is no longer a child of R and the cycle is broken. In general, the single parent relationship ensures no cycles can occur, I'm pretty sure. Except for the root node. If you ever attach the root node to anything, a cycle occurs. You could prevent this from happening by having a tree class which has a node reference it calls root that exposes the methods attach(child, parent) and detach(child). These simply check if child is root and if so, do not permit the operation. Or just never attach root to anything.
However, if you decide to allow multiple parents (by changing to T parent[] or removing the explicit parent reference entirely), things change. You may actually want to do this, since it's a very powerful feature to have. What you want is essentially a scene graph and if the single parent constraint is removed, you can re-use subtrees in multiple places. Then you can have stuff like the health bar apperaing on the regular hud as well as in the inventory, with no need to build two identical nodes referencing the same internal game state. Your can build a library of composite UI elements that you can re-use anywhere at no extra cost. If you do this, you have to check for cycles on every attach, by traversing the to-be-attached node's subtree and checking if it ever reaches the would-be parent. It can't happen that there already is a cycle anywhere in either tree, because that cycle would have been created by performing an attach, which checks for cycles.
3c17f8 No.14805122
for shaders that support heightmap (like Standard Shader in unity). If there is no height texture, which value is set to it? 0.5? 0?
1226c7 No.14805142
>>14801582
>subjecting oneself to this kind of shit
Just run loonix in a VM and do coding there if you're so attached to windows. Better yet, install loonix and put windows in the VM. PCI passthrough is not difficult to do and you won't notice the maybe 0.1% CPU overhead.
46351a No.14805154
>>14803770
Just converting from.HDMi into analog signals, thats all. I'm sure they're not 100% accurate, but they're decent I guess. The tvs I'm using dont have a wealth of ports so my options are kinda limited(one only has composite input and the small bubble one only has a working RF port).
1226c7 No.14805159
70cddd No.14805164
>>14805142
Or use something like Cygwin if you're attached to Windows and don't want to run a full linux environment in a VM.
And don't forget to select apt-cyg (a cygwin package manager) for installation when you're using the installer, so you don't have to use the installer every time you want to add a program. Why this isn't in the standard installed files yet I will never understand.
d9ae5d No.14805222
>>14801502
i am serious
but i am also an unmotivated piece of shit university soon to be dropout coder
it's your funeral
1226c7 No.14805226
>>14805164
Last time I used cygwin (maybe 8 years ago?) it was a really uncomfortable environment to work in, nothing compared to native linux. Has it improved?
915b9c No.14805317
>>14800556
Welp I finally figured out this shit, I had to resort to some ACS scripting in order to get the item respawn to work as some items such as the shell will spawn 2 items now (shell_1) and (shell_2), as well a check if sv_itemrespawn is off or on, so basically those 2 newly spawned items will behave as it was placed by map. All I need to do now is adding support for the few other classes then its done.
859f6e No.14805508
>>14799372
>Fell for the Factorio being fun meme
>It's actually fun and I blew 8 hours in the blink of an eye, my automation is disgustingly bad, and it made me realize I can't ever compete
I had fun though
c3c718 No.14805644
>>14799372
I'm close to finishing this educational game for display signage tablets that will run at a local aquarium.
7205c7 No.14806187
>>14805508
Factorio is objectively not fun.
80ac97 No.14806305
>>14805222
Whatever, dude. Just drop me a line or some shit, I guess.
d9ae5d No.14806313
>>14806305
sure, is steam alright? id is docktard
3afc24 No.14806469
>>14805142
I hate Linux, and I want an IDE with a proper debugger but none of Qt Creator, KDevelop, CodeLite etc. ever worked acceptably for me.
adb70d No.14806595
I'm working on CSG, I believe that I have sucessfully created all of the clipped polygons, but I think I have also failed to discard the polygons that are supposed to be discarded by creating those. Hence the Z-fighting where the clipped polygon and the original polygon are in the same place in this image.
>>14802360
I dunno, doesn't it add something to how your game feels if you can get it right? Sounds like a useful thing to me.
91efcc No.14806682
>>14806469
>IDE
>Linux
No wonder you hate Linux, you're going into it expecting a Windows clone. Pick a decent source code editor (if you really want something close to an IDE, try Emacs) and familiarize yourself with Valgrind and the infamous GNU Debugger. LLDB also exists but I know fuckall about it.
0b4b57 No.14806787
>>14805644
That's awesome, anon! Are you getting paid or are you doing it for fun?
3c17f8 No.14806852
I don't get Unity's normal maps, it's nothing like it's explained
Here is what I get in the shaders reading from a non-normal mapped texture whose RGBA was taken via getPixels() on a normal-mapped texture.
Here some info:
First pixel of the normal map when not marked as normal map from GetPixel(): RGBA(0.322, 0.698, 0.937, 0.063)
First pixel of the normal map when marked as normal map from GetPixel(): RGBA(1.000, 0.667, 0.678, 0.322)
So clearly, the packing already happened, R- > A and G - > G while discarding R and B. If my texture is made by reading A to R and G to G it should work, right?
Not really
As seen on the image, my results and whatever comes out of tex2D(_BUMP, IN.uv_BUMP) are completely different, but I can't figure out how and why.
My normal is made like this: fixed4 normal = fixed4(1, nrsm.r, nrsm.r, nrsm.g);
This is exactly what I get from GetPixel, which can be proven by unmarking the normalmap as a normalmap. The results from tex2D becomes exactly what I have except for the R value.
What exactly happens during tex2D if the image is marked as a normal map? I can't find this documented ANYWHERE.
UnpackNormal is unrelated, everything I have shown or discussed here was about the steps prior to executing o.Normal = UnpackNormal(normal)
my shader:
Shader "Custom/AtlasShader" {
Properties{
_ALBH("Albedo + Height (Array)", 2DArray) = "" {}
_NRSM("Normal + Smooth + Metalic (Array)", 2DArray) = "" {}
_MainTexture("Main Texture", Int) = 0
_BUMP("Main Texture (Atlas)", 2D) = "bump" {}
}
SubShader{
Tags { "RenderType" = "Opaque" }
LOD 200
CGPROGRAM
#pragma surface surf Standard fullforwardshadows// vertex:vert
#pragma require 2darray
#pragma target 3.5
struct Input {
float2 uv_ALBH;
float2 uv_BUMP;
};
sampler2D _BUMP;
int _MainTexture;
UNITY_DECLARE_TEX2DARRAY(_ALBH);
UNITY_DECLARE_TEX2DARRAY(_NRSM);
void surf(Input IN, inout SurfaceOutputStandard o) {
fixed4 albh = UNITY_SAMPLE_TEX2DARRAY(_ALBH, float3(IN.uv_ALBH, _MainTexture));
fixed4 nrsm = UNITY_SAMPLE_TEX2DARRAY(_NRSM, float3(IN.uv_ALBH, _MainTexture));
fixed4 normal = fixed4(1, nrsm.r, nrsm.r, nrsm.g);
if (floor(IN.uv_ALBH.x * 10) % 2 == 0) {
normal = tex2D(_BUMP, IN.uv_BUMP);
}
if (floor(IN.uv_ALBH.y * 10) % 2 == 0) {
o.Albedo = normal.ggg;
}
else {
o.Albedo = normal.aaa;
}
o.Metallic = nrsm.b;
o.Smoothness = nrsm.a;
o.Alpha = 1;
}
ENDCG
}
FallBack "Diffuse"
}
0b4b57 No.14807006
>>14806946
Anon, please use Handbrake to web-optimize your recordings.
453b30 No.14807045
>>14799554
it looks like the spongebob font
c3c718 No.14807383
>>14806787
I'm getting paid, but I'll be lucky if I make much money out of it. The project includes the tablets and metal casings to mount them on the wall. The casings were more expensive than expected, and the tablet customs and tax expenses were completely ignored beforehand.
fcc520 No.14807523
How do I improve my face? It looks somewhat weird but i have no idea why.
adb70d No.14807545
>>14807523
>Tim willits is posting in /agdg/
0b4b57 No.14807959
>>14807383
That's okay, anon! It's practice nevertheless, even if you don't get much out of it.
c3c718 No.14808081
>>14807523
Looks anatomically correct, just not very attractive.
>>14807959
Thank you!
a7cc38 No.14808203
>>14807523
Face doesn't have enough depth; she looks like a blowup doll.
Try giving her eyelids and do something to the chin and nose area to make them stand out from the rest of her face because it just looks flat. And I don't know if the appearance of her mouth slightly parted to show tongue is intentional but she would look less realdoll with her mouth closed.
3c17f8 No.14808255
Tree District 60% done (trees, platforms and bridges)
Aqua District 20% done, making floating houses next
Market District 0% done
Central Area 50% done (bar under the tree roots and central platform)
Temple Lake 40% done (need the temple itself)
Beach Area 40% done (need to make the lighthouse into an actual building and add a second pier)
Mountain Area 50% done (need to add a tower and improve the bridges)
fcc520 No.14808261
>>14808203
Thanks, i'll give sculpting another go.
203235 No.14808276
>>14807523
>tfw I'm one of the first to see a sneak peek of Life Is Strange 3
be5567 No.14808330
>>14808276
>Posting beanspeak
Remove yourself.
be5567 No.14808347
>>14808340
No beanspeak, Paco. Out.
3015b6 No.14808442
>>14808347
Donde crees que estamos?
3a6385 No.14808451
Does anyone know of an algorithm to jam as many rectangles into a given area as possible without any overlapping or going outside of the area?
07bb70 No.14808590
I've been learning to map for the game F.E.A.R
Right now I'm learning how to add in custom textures by porting Source engine dev textures to the game. Since they look sexy in every game I put them in. And this one is no exception. It helps that the game's engine allows for a ton of flexibility with shaders.
It's also a dead game to mod for so documentation is really scarce. I noticed the game's particle editing utility's documentation is incomplete. It has a heading for how to make a particle from scratch but that section is completely blank. Almost everyone I asked who modded this game in the past admitted they never figured out how it works. It's like I'm trying to learn Sumerian.
There's also a lot of flexibility with the game's lighting systems since it allowed for dynamic and volumetric lighting. There's so much flexibility with it that it's actually somewhat challenging to get lighting to look "right" since it's really easy to make a light look like a thick fog or to make it almost non-existent.
ec22b2 No.14808650
>>14808451
Why not do it backwards?
You have an area and you split it up, like you were making a puzzle out of an image.
Then use those rectangles in the game.
3a6385 No.14808665
>>14808650
Different areas require different kinds of rectangles. This is for placing trees based on whether or not the biome requires trees, and what kind of trees depend on the biome.
0567ec No.14808699
>>14808451
Look into packing problems, that's pretty much what they are about. It's a field of mathematics though, so what you find might be a bit hard to understand and possibly an optimal solution isn't even known yet.
ec22b2 No.14808702
>>14808665
There's an algorithm called Huang/Korf, but I'm not sure how up to date it is.
ec22b2 No.14808726
>>14808702
>>14808699
Huang/Korf was what I used a couple of years ago in a non-gaming project. There are some samples on GitHub. I haven't really been following the exciting optimal rectangle packing scene, so there could be better solutions out there.
5ad6b2 No.14808746
>>14808590
>I've been learning to map for the game F.E.A.R
F.E.A.R. is one of my favorite games ever, godspeed.
07bb70 No.14808785
>>14808746
Same here
The biggest flaw I have with F.E.A.R is it feels too much like a one trick pony (it only has 3-4 good enemy types and the level designs are really similar) and I think with just a few varied levels it would be a lot better. Max Payne for instance has distinctly different levels spread throughout and it's much better for it.
dc466a No.14808805
>>14808590
>I've been learning to map for the game F.E.A.R
For what purpose? However I commend you.
adb70d No.14808816
>>14808276
Its a model of nyx from quake champions. Close guess though.
07bb70 No.14808830
>>14808805
>For what purpose? However I commend you.
I really like Lithtech for some reason, it's a somewhat fascinating engine to me. Mostly because it had a ton of solid games on it that were abandoned by the developer later. Games like No One Lives Forever, Shogo Mobile Armored Division and so forth. Games I could imagine sequels for but will never get them. That sort of thing.
Also since it's a solid game that I feel could be improved upon massively with a superior presentation (mostly things like the weapon animations/models) and aesthetic (more varied maps). I have a lot of ideas for modifications of this game and it's unfortunate the only ones people made were for its multiplayer that nobody plays anymore. I see it as a superior game to something like Half-Life 2 and people still map for that game to this day which is a shame to me for some reason.
adb70d No.14809135
I used my visual debugger to find a bug in my CSG routine, so now I know how to fix the code. This is a video of me flying around in it. You can see the Z-fighting which shows that even though a polygon split has occured, the original polygon wasn't deleted, on the second polygon split. Not very good… I'll fix this and post a new video of what it's supposed to look like.
adb70d No.14809262
>>14809135
And here is the code where CSG works correctly, you can see that there is no conflict and the algorithim correctly clips the polygons together. This means that my invalid geometry problem is fixed, I will no longer have bugs in my BSP compiler. It also means that almost 1000 lines of code have been made irrelevant by this, and I can delete them ALL which I am going to be very happy to be able to do. I dont like the look of that old code that I wrote all last month, so im happy that its gone.
8396e6 No.14809305
>>14806469
Have you tried CLion?
3b7dd5 No.14809391
>>14809262
Looks great, Anon. Now redo the whole engine with meshless geometry and distance functions.
fc6892 No.14809412
>>14807006
that's what I get for trusting OBS. Oddly enough, hiccups only happen in some browsers. Let's see if webm will work better.
>>14808830
from what I read, the F.E.A.R. SDK contains source code for the whole GOAP AI statemachine so I'd take a look at that if I were you.
Kinda sad that the fantastic engine that it is never really got much traction and was only used in Monolith's titles and CI Games' shovelware.
07bb70 No.14809512
>>14809412
>from what I read, the F.E.A.R. SDK contains source code for the whole GOAP AI statemachine so I'd take a look at that if I were you.
They released the source code to all of the game logic, this was actually something they usually did with their games. It doesn't contain the engine code which I honestly don't need. The game logic code is mostly useful if you want to do things like alter the UI or add functions to the game. (I'd imagine adding dialogue options like NOLF or MCAs like Shogo would require altering game logic).
Adding things like weapons can actually be accomplished fairly easily by just altering the game's internal database. It's odd it was designed to be so modular.
One huge downside to F.E.A.R's sdk not a lot of people know about is Monolith put DRM into it. The sdk refuses to install unless the game is installed. And the Steam/GOG version's don't count you need to use the disk version. These Russians hacked it a few years ago so now it works although it's for F.E.A.R Combat (Sierra released F.E.A.R's MP as free to play for a few years) not SP. I can still make singleplayer maps but I'm suspecting there's things missing from FEAR Combat's sdk. So I'm doing some investigating. It also requires the game's cd to be put in presumably so people don't just extract the game's archives off the cd and run it via the editor but it's easily fixed with a nocd patch.
adb70d No.14809997
I can now compile the map that I was trying to compile, and the BSP collision detections work great! The only problem is that the textures are not working on all of the polygons, so I will work on this bug next.
>>14809391
maybe in the next engine, when I am better at math…
1226c7 No.14810261
>>14806852
There are various ways of encoding normal vectors. Unity may use an optimized one.
The naive way is to simply say that R->X, G->Y, B->Z, but this has a pretty massive problem. Each component has a range of -1.0->1.0, encoding 256 distinct values (assuming 8 bits per pixel). With 3 such components, you should have 16777216 distinct values at your disposal, plenty of precision right? Except that you'll only use a very small subset of that. Normals are always unit length, so the values that get encoded are those that lie on the surface of a unit sphere. So if you imagine a cube spanning -1³->+1³ quantized into 256 steps along each axis, thus consisting of 16777216 points in total, the only points you actually use are those close to the surface of a sphere inside said cube. Actually, only half of a sphere, since z>0. That's far fewer values. There are many ways of improving the data utilization, ranging from constructing a lookup table to various space-stretching functions. It's all a tradeoff between better utilizing the data space available versus the computational complexity of decoding back to cartesian space.
The point is that unity might be using some sort of optimized representation under the hood. So when you pass it a normal map, it might be converting it to some other representation before uploading, then wrapping texture access in the shader with an unpack function. To make this work, you probably need to declare your texture access as normal access, so the required code gets generated. Alternatively, this could be some sort of sRGB hijinx. The non-normal access might go though gamma correction, while the normal one doesn't (or vice versa). It's tough to say when you're working with an engine and don't know everything that happens under the hood.
One thing you could do is construct a texture containing a representative subset of the values a real map would contain, like pic related. Then construct the corresponding two textures sampled textures like you did for your picture, then stare at the result until the solution occurs to you. Experiment with the data, like convert to heightmap, then use said heightmap as a displacement map in blender and see if you recover something that looks like a sphere. Compare color histograms. Anything you can think of.
1226c7 No.14810281
>>14808451
http://blackpawn.com/texts/lightmaps/
I've implemented this algorithm before, but I don't have any of the test images I generated and cba to revive the code.
1226c7 No.14810319
>>14808590
Scan the credits list. Look for people that held positions related to environment lighting. Start specific then go more generic. So if you can't for instance find one with the title "head of environmental particle effects", try "environmental engineer", "effects engineer" etc. Substitute "engineer" for "programmer" if the former term is absent. Also try tools programmers. Only go for "artist" if you can't find coders relating to the subject. When you have a list of names, start digging. Look for linkedin profiles, twitter, tumblr, facebook etc. Try to find an email address. Then mail them explaining your situation, asking if they remember anything or have any documentation still laying around. I think you'll find many of them happy to help. Just make sure to be excessively polite and humble. Hide your powerlevel and contain your spaghetti, even if it turns out they're pink-haired nu-males.
1226c7 No.14810337
>>14809262
I think I read somewhere that a good programmer is one that finds as much pleasure in removing code as he does in adding it.
a8ca51 No.14810595
Erni now can wear a toque. I'm not sure why I added this.
8396e6 No.14810603
>>14810595
Cute. I recall being able to fish out a boot? Make him able to wear that as a hat as well? Might be fun.
a8ca51 No.14810618
>>14810603
That's hilarious, I'll have to do that
ff43d7 No.14810623
>>14810595
>>14810603
What about a lootcrate that you could fish up, which the player can open for a meager purchase of 2.49 USD, allowing the player to wear the lootcrate as a hat.
a8ca51 No.14810648
>>14810623
Heh, I made a treasure chest just as a joke type of fishable "garbage", but it might be fun to add some actual functionality to it.
74a59f No.14810651
>>14810595
Did you think about the proposal of allowing the player to re-assign the controls from demo day?
Playing it on QWERTZ keyboard was a pain in the arse.
c0c1c0 No.14810704
I've kinda hit a roadblock on how I can implement an idea.
I'm wanting to do an interactive point and click game in the style of a vn.
I'm wanting all the rooms/environments to be hand painted 360 panoramics.
**https://www.artstation.com/artwork/KQmKW
https://www.artstation.com/artwork/EKo92** <- like these
Would it be possible to render 2D character sprites separately on the screen while retaining the ability to turn the screen without having to create a distorted sprite for each spot they'd be in?
a8ca51 No.14810734
>>14810651
Yeah, still on the todo. I'll have to make platform specific input code as right now I'm using a generic solution since I work on my game a lot on my (((smartphone)))
74a59f No.14810741
>>14810734
>on my (((smartphone)))
Looking forward to it anyway :^)
c3c718 No.14810774
>>14810704
If the sprite is some distance away it might look ok.
70cddd No.14810897
>>14810704
If you have stylized spites and just use billboards for them, plus having multiple layers on your panoramic environments so you can obscure them partially, I think it might actually look quite nice and be helpful for UX as it would be very obvious what you can interact with. Somewhat like Professor Layton's sprites, but with 360 views.
1226c7 No.14811291
>>14810704
Yes.
A 360 panorama is just an image that can be sampled to construct a "3D-looking" view of the scene. Basically, you have a rectangle which is your viewport and a pair of angles describing which direction in the panorama you want to "render". You loop over the pixels of your viewport and transform them using a transform built from a set of parameters, including the viewport direction, aspect ratio, fov and maybe some more. You sample from the panorama at the transformed coords and paint the sampled pixel to your view et voi là: something that gives sort of a 3D effect when the camera pans. To draw something into this virtual world, you just have to perform the inverse transform on what you wanna draw. From what I remember researching this stuff, there are a ton of different panoramic projections to choose from, but your pic looks like a simple cylindrical projection. Research how cube maps work, they're sometimes used as a starting position to transform a 3D scene into something panoramic.
Alternatively, just draw them like you would in a regular 3D engine, i.e. decide on a camera matrix (camera orientation, then perspective projection) then toss sprite billboards at it. Though depending on how the panorama is drawn, things might not look right in motion. It's difficult since you're mixing an explicit 3D space with an implicit one and hand-draw is never 100% correct perspective-wise.
c0c1c0 No.14811464
>>14811291
Yeah I was thinking something kinda like this but couldn't formulate it in my head properly.
Currently using grids as a base in photoshop's 3D spherical panorama mode. Think the hardest part will just be having a consistent fov and perspective between all environments so the sprites can be projected relatively accurately across the board.
91efcc No.14811783
Partially inspired by this recent Steam censorship thing ( >>14810393 ), a friend and I are looking into setting up an open source, federated vidya marketplace. The idea is that developers upload games to a federated server running the marketplace software (usually run by a publisher or someone taking a small cut from sales to pay for server upkeep, but nothing stops you from running your own) and buyers are given a cryptographically signed receipt which is stored in a client-side wallet. This receipt is used to authenticate game downloads and updates through said client, which likely doubles as a launcher too.
Thing is, to reduce federation bandwidth costs we'd likely implement a P2P sharing system. I'm unsure whether we'll go with a locked down system with peers checking each other's receipts (which could complicate refunds) or just go with a protocol like BitTorrent or IPFS (meaning buyers could peer with pirates, which can be a negative or a federation bandwidth-saving feature depending on your outlook).
If we implemented this, would any of you fags be interested? It would be much easier to set up and host than a regular web store, and fairly censorship-resistant as long as there's people willing to host federated servers. There's also nothing stopping someone from hosting H game-friendly federations.
>inb4 payment processing
To keep things FOSS and simple-ish, https://taler.net/en/
d9ae5d No.14811788
>>14811783
can't you just use itch? is it censored too?
91efcc No.14811806
>>14811788
Itch.io has fairly strict filesize limits and it's a traditional web store run by SJWs, meaning they're free to kick you off the platform as they see fit. This is intended as a final solution to the vidya store problem, so there isn't a single point of failure and federations/publishers are free to host whatever the fuck they want without the overhead and complexity of hosting a regular web store.
19682a No.14811828
>>14811783
I don't have high confidence in these P2P/federated/distributed services, they tend to cater to a very niche audience and generally be shit, but if you think you can do it and make sales secure (see: people can't exploit it to download all the games) then sounds good. Call again when you have more than an idea.
>>14811806
All of itch.io's limits can be lifted by contacting them. As far as I remember they're in place just so some random faggot won't suddenly upload tons of shit.
91efcc No.14811891
>>14811828
>see: people can't exploit it to download all the games
Preventing people from downloading games from a federation or receiving update notifications is pretty easy if you check their receipt before sending them anything. Theoretically you could set all clients to check each other's receipts for the federation's signature, but adding refunds into the mix would complicate things. Alternatively, you could just say "fuck it, buyers can peer with pirates too since it simplifies shit and saves us money" and use Bittorrent or IPFS.
People buy games in the first place for a mixture of convenience and wanting to support the devs. With either P2P method pirates lose the convenience of automatic updates.
a8ca51 No.14811896
>>14811828
itch.io are a bunch of SJWs; they're the ones who hosted that "resist jam" bullshit
19682a No.14811901
>>14811896
I know, they also constantly promote progressive walking and talking simulators in the official blog.
I haven't heard of them removing any politically incorrect games or anything like that though.
7c5d9d No.14811904
>>14811896
>>14811901
Is Gamejolt any good?
1226c7 No.14812214
>>14811783
Not gonna work.
>cryptographically signed receipt which is stored in a client-side wallet
So if you lose that wallet you lose all your games. normalfags can't handle the responsibility of maintaining valuable data, they need an avenue of recovery, I.e. a password, which can be recovered through email, said email being recoverable through phone number, which is recoverable by going to a store and showing their ugly face.
>P2P sharing system
Fuck off. Why should I share my bandwidth, that I paid for, so that you can avoid paying for hosting? I alreay paid for the game and you're making me pay a second time. Either I get everything for free and share bandwith or I pay for a service that works through and through.
>taler
Requires cooperation from a bank or a government. Or do you suppose coins will just pop out of thin air?
Tech-savvy communities can maintain these sort of things amongst themselves, but it will never fly with the general public. Decentralization is antithetical to ease of use, which is a hard requirement for mass appeal.
1589ba No.14812341
>>14812255
that's one lanky-ass sonic
1226c7 No.14812416
>>14811891
>check their receipt before sending them anything
>>guiz, i cracked the nusteam client so now you can download any games without paying
>crack simply removes the outbound check, mostly making you a seeder
>crackers set up clients with large collection of games as seed boxes to make sure content is available to early adopters
>people adopt the cracked client and download loadsa games, adding to the swarm volume exponentially
>client reaches critical mass, every game now readily available from hundreds of peers at any time
>publishers abandons platform
The only variable is whether people will block outbound data as they did with the regular client, but then the crackers will just keep running their seed boxes, like they do already for the bittorrent network. Fun fact: 99% of the seeding is done by 1% of the seeders. The networks runs off a bunch of dedicated autists with secure seedboxes and fuckhuge libraries. Almost nobody seeds back what they download. So if you think that "surely nobody is willling to pay for all that bandwidth out of their own pocket", think again.
1226c7 No.14812576
>>14811783
I should probably add something constructive and not just criticize. If you wanna launch a non-pozzed steam competitor, launch a non-pozzed steam competitor. Do it the way this has been done forever. Offer a superior service at a superior price point. Make a storefront that is well curated. Undercut the competition. Hosting isn't nearly as expensive as you think it is, even at these volumes. The big cost is development. Coding. System administration. Reachout to publisher. Deal-making. Basically, all the human resource stuff. Humans are expensive, work limited hours and have flaky availability and quality. Computers work 24/7 for almost nothing and deliver with 100% consistency, given a consistent environment. That's why valve (and everyone else) want to automate everything.
I'd do something like this: start out with the itch.io model of letting people sell games with no entrance fee. Free games should cost nothing to distribute, ever. Then, only start taking a cut if they reach a threshold of sales, like say $1000. Take a very small cut and be open to negotiating deals. Big publishers might not let you take 1% of their shekels, so try tiers. 10% of the first $10 000, then 1% until $100 000, then 0.1%. Consider the value of having a big title on the platform vs getting greedy with cuts. You don't need that much money. Valve only think they do because gaben uses it as toilet paper. With indies, larger cuts might be acceptable, but with a higher free bar. Heck, many would probably be okay with something like "Give me everything until $2500, then you take the rest, even if that's $1 000 000". Be fair and communicate. In the current world of excessive automation, human interraction is an immense competitive advantage.
Also, you're gonna need maybe 5 years of development and a ton of investment capital. It's far from impossible to disrupt entrenched industries, as itunes, netflix, tesla and spacex all have proven. But it takes time, hard work and exceptional people.
323da2 No.14812657
>>14806187
This anon is objectively wrong.
91efcc No.14812759
>>14812416
>Fuck off. Why should I share my bandwidth, that I paid for, so that you can avoid paying for hosting?
You don't have to seed, faggot.
>>14812416
He and I discussed that scenario internally, which is why I personally lean towards removing a peer receipt check entirely or making it optional so legitimate buyers benefit from pirate peers.
>>14812576
I'll be honest: I'm probably not up for hosting my own Steam competitor and my country would probably force some content restrictions. The federated marketplace concept was an idea this friend and I had on how to simplify selling vidya without making an assload of separate stores with their own accounts and high bandwidth costs. The main obstacles to this idea are people losing wallets, devs/federations being uncomfortable with pirate peers assuming P2P isn't dropped entirely, and payment processing.
When I sell vidya myself, I'll just go with a much simpler concept for fags who don't want complicated storefront bullshit: selling email subscriptions of game downloads and updates using regular download links and magnet links/IPFS hashes. Pirates peering with buyers is alright with me and gives better service for both of them.
Thanks for the feedback. I'd rather not waste time or money on setting up a weird federated marketplace if I can't get people to use it.
ff43d7 No.14812982
>>14811783
>>14812759
Since people are giving criticism I guess I'll just throw in my 2 cents too.
I can't seem to find the article about why you shouldn't start your own platform, but the tl;dr is that the market is pretty much covered already.
The main problem with starting a new platform is that in order to get customers, you need developers willing to sell games on your platform. But those developers wont waste their time on you if you have no customers. The good old chicken'n'egg-a-roo.
The only way to get anywhere at that point is to try and fill a niche that nobody else does. But there-in lies the problem; there is no big enough niche.
Sure, an open source platform sounds great, but the harsh reality is that almost all games are closed source; if you want to play anything you already have to accept that you have no fucking clue what code you're running. And when you have that going on, you really would like someone to pin the blame on if the game makes your computer explode. Making the marketplace distributed here removes that, which is a clear step backwards for pretty much everyone.
I understand the concerns with Steam's recent censorship of lewd weebshit, but the reality is that most developers would prefer not to share their platform with porn.
That all said, I think a platform like that can have its share as success, but only as a cool tech hobby project. If it ever reaches mainstream appeal, then people will inadvertently ruin the entire thing, most likely by flooding it with absolute garbage or leeching torrents off of shared receipts.
For a good example of that happening, just look at bitcoin. Decentralized money sure is a cool tech project that has legit use cases, but suddenly we got a gold rush of retards driving up GPU prices trying to make a profit. The cryptocurrency became straight-up unusable as transaction costs soared.
1226c7 No.14813077
>>14812759
I'd definitely try it, assuming it actually had games on it. I just doubt it's feasibility in the form you presented and think your assesment of what is easy and what is not is incorrect. You need to think hard of various adversarial scenarios and how your scemes could be broken. Also, consider this: given the failure of bitcoin, youtube content id and others, shouldn't we kinda have realised by now that societal problems can't be solved by throwing tech at them? If you wanna fix the problem of biased videogame retail, I'm convinced it's gonna take more that coding to solve it.
>>14812982
>most developers would prefer not to share their platform with porn
Curious: do we have proof this is the case? I'd imagine most publishers wouldn't give a shit as long as they sell cop- uh, lootboxes. My assumption was that this was a sign of valve getting more pozzed. In my experience, while big companies are willing to throw money at leftists for exposure, they generally don't embrace the ideology. It's just another marketing expense. However, since virtue signaling and inserting dykes doesn't sell games, it's a failing scheme. Then again, marketing in itself is a huge pyramid scheme that provides little real-world value. Some, ofc, but the billions being spent don't equate to billions in sales. Point is, either publishers are rational capitalistic entities and will follow the money or they really have drunk the kool-aid, in which case they would never cooperate with this upstart to begin with. A competing store isn't gonna fix the later case. Only good games from good developers is gonna fix that. Which this very thread sort of is trying to foster.
91efcc No.14813148
>>14812982
The idea was that the wallet and server software would be FOSS, not necessarily the games being sold. Also, the system is more federated than decentralized: games are uploaded to federated servers which, like Mastodon or Matrix servers, are free to be as lax or curated as they want. At one point we also considered tying player identities to federated servers instead of a wallet client, but if one of said federations died it would take loads of player accounts with it.
>most developers would prefer not to share their platform with porn
They could always use a federation which doesn't host h-games.
>>14813077
>You need to think hard of various adversarial scenarios and how your scemes could be broken
Which is why I'm bringing it up with you guys instead of diving in right away. The concept is still new and I'm figuring out if I can make it work with some tweaking or not.
ff43d7 No.14813190
>>14813077
The reason why Valve has to avoid porn games on Steam like the plague ties into regulation.
The video game industry has very little actual regulation. Most of it is self-regulated. For example; the ESRB, which is funded by the ESA, which is comprised of industry leaders like EA, Ubisoft, Activsion, etc., rates games, and based on those ratings stores decide whether or not they accept these games. It is because of this self-regulation that there is almost no legal regulation.
Unfortunately this doubles as a bit of a racket; you can't really sell your game unless you have an ESRB rating, but those are expensive and cost time and other resources. The reason why digital distribution was suddenly a success is because there FINALLY was a fucking store for games that don't have ESRB rating. This lowered the bar significantly, which was great.
However, the previously mentioned self-regulation is not used by Steam, which quickly grew to become the biggest store on the internet. If Steam would start selling porn games, then that would mean they're also selling porn games to minors. A couple of concerned parents later and you have a shitstorm bigger than the lootbox regulation nonsense, with legal regulation following very quickly.
If you look at the games featuring pornographic content that Steam ignores, you will notice that they're all games with an ESRB rating that deems them acceptable for stores.
The games Valve is removing are games that most likely get an AO-rating that would disallow them to be sold in stores. The reason why these games would get AO-rating, despite containing roughly the same levels of nudity, is simply because of cultural reasons. People don't consider GTA V porn, but they do consider Huniepop porn, ergo Huniepop deserves an AO-rating and GTA V gets off easily.
It's unfair, but considering that the point of the rating system is to prevent regulation of the gaming industry, its what you'll have to deal with. I dislike the double standard too, even though I don't give a shit about porn games.
ff43d7 No.14813205
>>14813077
Anyway, the reason why selling non-porn games alongside porn is a bad idea is because your customers need a way to sift through products to find what they want. If your game is being sold next to porn, then they will automatically assume that its very likely that your game will be porn too. If they don't want a porn game, they'll fuck off to a store that doesn't have porn. That way they're sure they wont end up accidentally buying porn for their shitty kid.
c0c1c0 No.14813212
>>14813205
That's as easy as putting porn games into their own category that you can disable from view.
6db0cc No.14813336
>>14813212
Kids can easily get around that, and the only real way to eliminate that; in addition to the potential bad PR, is to eliminate it from their storefront entirely.
They really only think of profit anon, and H-games are smalltime in terms of what they could lose; i.e. all those parents not wanting their kids on a "disgusting game service that sells my kid porn games".
Speaking from a realist's pov (devil's advocate in this case), there's a time and place, and steam is not a place to sell porn games; it's simply too mainstream (meme on how mainstreaming a hobby ruins it.png).
70cddd No.14813358
>>14813336
Anon, they sell GTA5, a game where you can fuck prostitutes and get lapdances.
c0c1c0 No.14813397
>>14813336
If the parents are just handing their kids the credit card and say "ya sure get whatever" without looking what they're buying, they've no room to complain.
If parents were that stingy they wouldn't let kids watch cartoon network because adult swim comes on at night. They wouldn't let their kids on the internet in the first place because there are porn sites.
6db0cc No.14813495
>>14812416
>Fun fact: 99% of the seeding is done by 1% of the seeders.
Maybe it would work if it was system like what private trackers do, and have a seed/leech ratio tracked per user; you put back in what you take.
Although there would definitely have to be an incentive todo so, and a persistent record of who seeds/leeches.
>>14813358
true, but case and point
>They really only think of profit
I agree with what you're implying, and that it's double standards… but u know, private company and all.
>>14813397
>If the parents are just handing their kids the credit card and say "ya sure get whatever"
Generally in the case of games, parents don't play games, nor do they care about games.
They do not know what GTA is, or what it's about. A kid can easily show their parent select clips of the game that demonstrate, "it's just like CoD but open wurld mum!", and get their parent to buy it.
>they've no room to complain.
you would think that, but that's not the case in the real world; they have to have someone to blame.
>the rest
>all parents know of adult swim
>all parents know kids are deviants, and not little angels
Also it's not always that they "care care", it's just they want someone to blame for X, Y, or Z issue.
Humans are one thing, and that's consistently irrational.
>>14813366 (checked)
nice
1226c7 No.14813500
>>14813205
>>14813190
Doesn't that mean that 91efcc's idea is stillborn? If any mainstream marketplace is gonna have to keep kids away from porn games, won't a federated model just get coerced to censor at some point? Alternatively, get slandered until it's a liability for publishers to do buisness with them?
>>14813397
If there is one constant of humanity, it's irrational behaviour. Parents let their kids play in the street, then complain to get the speed limit lowered because it's "unsafe". Tons of laws exist because too many are too lazy to behave responsibly. They have no room to complain, but they will anyway and since said parents have the right to vote, politicians will be more than happy to oblige to score brownie points.
c0d2f4 No.14813503
>>14813500
>and since said parents have the right to vote, politicians will be more than happy to oblige to score brownie points.
just another example why democracy a big fat SHIT
91efcc No.14813545
>>14813500
>won't a federated model just get coerced to censor at some point
That wouldn't work unless you either coerced every federation hosting adult vidya to stop (unlikely) or set up a client-side lewd federation blacklist on top of whatever age restriction system is used.
80ac97 No.14813781
>>14813495
>Generally in the case of games, parents don't play games, nor do they care about games.
>A kid can easily show their parent select clips of the game that demonstrate, "it's just like CoD but open wurld mum!", and get their parent to buy it.
Evidently they don't know what an ESRB rating is, either. Which doesn't surprise me, given the populace's general intellect.
>Consisently irrational
Well there we go.
3b7dd5 No.14814219
>>14813503
Democracy is the worst form of government, except for all the other ones.
0b4b57 No.14814640
>Finally got linear storage octree to work
>Got some weird garbage collection type of deal working to maintain order of indices for existing nodes (keeping it all in a linear array is just handier for serialization and iterating over existing nodes)
>suddenly contouring algorithm freaks out
>check octree with debug graphics - seems fine
>check algorithm, seems fine
>check if I'm actually marking redundant nodes as pruned - its all fine.
>tried setting half of each chunk to be filled with voxels, the other half is air - shit's totally fine and looks good
>the issue only occurs when using noise to generate blobby terrain
call 911 im having a fucking stroke lads
adb70d No.14814939
>>14814640
If nothing else works, try writing a debugger that slowly steps through the algorithm, if you can't figure anything else out. I wrote one for my project and it's really helping me a lot.
98a3e3 No.14815056
rapey monster shaping up (blendshapes.)
8f803a No.14815570
I published a post about my scripting language in my engine, if anyone is into that: https://kircode.com/post/designing-a-custom-scripting-language-for-my-game-engine
>>14810595
Is that a life counter in the top left (5 lives)? Maybe the hat could serve as an additional life when equipped. In Kero Blaster you can put on a winter jacket that absorbs a bullet and disappears when destroyed.
>>14806946
>>14809412
Damn, this looks really neat. Good job.
ff43d7 No.14815588
>>14813500
>Doesn't that mean that 91efcc's idea is stillborn?
Yep.
The only way a marketplace can succeed whilst selling porn games is if it is exclusively selling porn games to adults. If the site is very clearly marked as for adults only, it can avoid regulation just fine. Unfortunately that severely shrinks down your potential customer base, to the point where it may not even be profitable.
>>14813545
>That wouldn't work unless you either coerced every federation hosting adult vidya to stop (unlikely) or set up a client-side lewd federation blacklist on top of whatever age restriction system is used.
I think you're misunderstanding what makes distributed systems work.
When it comes to torrenting, its hard to suppress since the only one liable for claiming damages is copyright owners. They don't know who the fuck you are, and they don't want to ban torrents altogether either since many of them cram them into their own software to cut down on update bandwidth costs.
Then you have a federation hosting adult vidya. This federation doesn't infringe the copyright of some bozo's that have a hard time tracking you down. Instead, it breaks the law of the government, which by the way, you're required to completely fill in on your operation in order to be compliant with tax law. With that information, they'll drag you to court, tell you to never do this shit again and slap on a fine and potentially a couple years in jail.
Sure, you can simply avoid giving the government your information. But then you're dodging taxes. And the punishment for that is way worse.
Taking down every federation isn't necessary. Throw the owners of a couple big ones in jail, and the rest tends to follow suit. As long as it appears to be a high risk low profit venture, it wont gain enough traction to be widely successful.
d9ae5d No.14815952
only now realized that i'm also supposed to serialize buffs
went and added buffs to items while i was at it - only example right now is that boots increase your speed by 2 when worn
0b4b57 No.14816203
>>14814640
>>14814939
Wow I figured it out. Apparently for whatever reason with the way my system works, It only works correctly if I instigate a "collapse wave" (my way of maintaining octree sparsity) it only works if I'm instigating it from "solid" voxels as opposed to air. I have no idea why that is, and it's not making me very comfortable, but it's progress.
0b4b57 No.14816376
>>14816203
turns out I was missing a check for whether or not a node was a leaf or a branch during the recursive collapse so it was collapsing branches that had the wrong material index assuming they were leaves.
3c17f8 No.14816419
>>14810261
>>14806852
nrsm.rg = normalize(float3(0.295, nrsm.rg)).yz;
seems to be the answer, there nrsm.rg is the green and alpha values of a normal map texture.
Is there any significance to the number 0.295 or is it close to any significant number? like log(10) or cubic root of some number or something?
This is the third time I get a "solution" by messing around randomly with numbers, the last two times it didn't work on all situations and I fear this one will not either, even if it have passed all previous tests
final shader: https://pastebin.com/seYkALUw
It's only this shiny to showcase the normals
How do I fix the shitty mip mapping of Unity's Texture Arrays?
Using UNITY_SAMPLE_TEX2DARRAY_LOD twice and interpolating the results seems overkill, but the only way at the same time
3c17f8 No.14816751
I've managed to enable parallax height mapping thingy you see in the standard shader. But this is so fucking terrible that I am thinking of using the last color component for something else instead
7c5d9d No.14816773
Well, I think the front of the head is pretty good for now. I don't really know what to do in terms of the back, though. Any suggestions? Modifying the hairstyle seems like the only real idea I can come up with.
3c17f8 No.14816786
>>14816751
disregard this, the heightmap is wrong
This new WEB is Unity's parallax vs Some proper (and expensive) parallax/height stuff whatever it's called
0b4b57 No.14816799
>>14816786
The second one looks really good! Do you mean computationally expensive?
7a4d82 No.14816815
>>14816419
>How do I fix the shitty mip mapping of Unity's Texture Arrays?
You can always build the mips yourself cpu side (fixed the issue for me when i was messing with texture arrays ages ago).
You could also use a gradient sampler, as explained in this post:
https://forum.unity.com/threads/texture2d-array-mipmap-troubles.416799/#post-2730309
>>14816751
The default parallax method is just… bad imo.
It has distorted UV lookup coordinates for the pixel height offset, and the normal map lookup coordinates are distorted too.
I much prefer raymarched parallax, and having some raymarched AO thrown in there can help the look a lot too.
>>14816786
nice, much better.
I'd recommend to implement some type of lod for the expensive parallax (single branch).
If it's done in a similar manner to mips, in the sense of, after rate of change is above X (ddx, ddy); then it will have massive performance benefits (basically, it will be able to group threads better, even with a branch, it offers massive performance benefits).
016fc2 No.14816998
>>14815952
thanks mr skeltal
016fc2 No.14817030
>>14815570
>Maybe the hat could serve as an additional life when equipped.
Yeah, that's the life counter. I was thinking the hat could be something like that. Each hat could be its own powerup, I guess, since I wanted to make a number of silly ones.
3c17f8 No.14817133
>>14816815
>If it's done in a similar manner to mips, in the sense of, after rate of change is above X (ddx, ddy); then it will have massive performance benefits (basically, it will be able to group threads better, even with a branch, it offers massive performance benefits)
I only marginally understand what this all means, I took the 'actually good parallax heighty thingy shader' from the internet and I don't exactly know how it works.
I will just leave Unity's and focus on finishing the map system rather than keep spending time on the shader, it does what it HAVE to do, even if it could do it better
The only thing I am worried about the current shader is the arbitrary fix to the normals, I fear it didn't actually fix and I will have to come back to it yet again later
3c17f8 No.14818078
Holy shit this thread is dead, I feel bad being 50% of the posts in this thread in the last 8 hours.
Now I have Triplanar texturing with heightmap blending.
What do I do about the angles where the axes are at 50-50? Just hope no mesh will have faces at these angles?
Currently It's a boolean blend, either one axis or the other, can't have 30% one, 20% the other and 50% the last, but it seems it needs to be that way?
0b4b57 No.14818171
>>14818078
It's not dead, anon! Myself, and I'm sure everyone else are hard at work on our games. I come here to read other people's posts every so often, though.
Sorry though, I'm not at all familiar with texture logic and such - I'd help you if I was.
7a4d82 No.14818253
>>14818078
void surf (Input IN, inout SurfaceOutput o)
{
// Find our UVs for each axis based on world position of the fragment.
half2 yUV = IN.worldPos.xz / _TextureScale;
half2 xUV = IN.worldPos.zy / _TextureScale;
half2 zUV = IN.worldPos.xy / _TextureScale;
// Now do texture samples from our diffuse map with each of the 3 UV set's we've just made.
half3 yDiff = tex2D (_DiffuseMapX, yUV);
half3 xDiff = tex2D (_DiffuseMapY, xUV);
half3 zDiff = tex2D (_DiffuseMapZ, zUV);
// Get the absolute value of the world normal.
// Put the blend weights to the power of BlendSharpness, the higher the value,
// the sharper the transition between the planar maps will be.
half3 blendWeights = pow (abs(IN.worldNormal), _TriplanarBlendSharpness);
// Divide our blend mask by the sum of it's components, this will make x+y+z=1
blendWeights = blendWeights / (blendWeights.x + blendWeights.y + blendWeights.z);
// Finally, blend together all three samples based on the blend mask.
o.Albedo = xDiff * blendWeights.x + yDiff * blendWeights.y + zDiff * blendWeights.z;
}
3c17f8 No.14818260
>>14818171
>I'm not at all familiar with texture logic and such
there is also the lack of progress posting.
I guess some days are more active than others
7a4d82 No.14818286
>>14818253
better version, that i have in an include file, forgot about this one. other one has some inconsistencies
half4 Triplanar (float3 wNorm, float3 wPosi)
{
// Find our UVs for each axis based on world position of the fragment.
half2 xUV = wPosi.zy / _TextureScale;
half2 yUV = wPosi.xz / _TextureScale;
half2 zUV = wPosi.xy / _TextureScale;
// Now do texture samples from our diffuse map with each of the 3 UV set's we've just made.
half3 xDiff = tex2D (_Tex0, xUV);
half3 yDiff = tex2D (_Tex1, yUV);
half3 zDiff = tex2D (_Tex2, zUV);
// Get the absolute value of the world normal.
// Put the blend weights to the power of BlendSharpness, the higher the value,
// the sharper the transition between the planar maps will be.
half3 blendWeights = pow (abs(wNorm), _BlendSharpness);
// Divide our blend mask by the sum of it's components, this will make x+y+z=1
blendWeights = blendWeights / (blendWeights.x + blendWeights.y + blendWeights.z);
// Finally, blend together all three samples based on the blend mask.
return half4(xDiff * blendWeights.x + yDiff * blendWeights.y + zDiff * blendWeights.z, 1);
}
0b4b57 No.14818287
>>14818260
Some days just have a lot of hidden progress, anon. Optimizing your game a little, refactoring code to make more sense, hell - even just going through your old code and adding comments is still progress, but it's progress that you can't really share with a picture. I'm sure we'll see a burst of posts soon.
476974 No.14818334
Should I look into Lua at all? I saw some guy made a 3D extension to Love2D. Looked good, but I don't think it'll replace Godot for me. I still want to delve into more barebone engines/frameworks. I just don't want to be some bitch while I let the engine do all the work for me. Any good barebone frameworks for 3D games?
2523f8 No.14818352
>>14818260
I prefer not posting that much progress until I've gotten to a certain point. Hopefully I'll start posting what I have in about two months. Enginedeving is a hell of a pain sometimes.
4c7809 No.14818383
>Drill boy is actually happening
Congratulations /agdg/
19682a No.14818414
>>14818393
It's like a real videy gomo
0b4b57 No.14818415
>>14818383
>>14818393
Are you the anon who's working on Drillboy? If so, congratulations on this important step!
e17511 No.14818430
>>14818393
>>14818383
Pretty neat. Where was this?
3c17f8 No.14818486
>>14818286
>>14818253
Making the shader actually become triplanar I got already working, the two things I am having problems are "how do I blend the 3 planes" and ParallaxOffset being inverted in some axies, which I will solve via trial and error using the axisSign variable
The "how do I blend the 3 planes" is just a question of figuring out some way to get a mix of taking into account the height map texture and the default blending technique
The shader: https://pastebin.com/g7W41nn5
>>14818334
>I just don't want to be some bitch while I let the engine do all the work for me
Well, that depends on whenever you want to be come a game developer or an engine developer. I believe I am already wasting too much time with terrain/grass/trees and too little making a cute farming game
>>14818393
>a fucking wrench
0b4b57 No.14818491
globby, cloudy, generates way too much unnecessary geometry, but guess what? It's generating a fucking terrain. I'm such a happy guy right now
3c17f8 No.14818531
>>14818486
>which I will solve via trial and error using the axisSign variable
Took less than expected, I know full well that there is bound to have a super small and neat equivalent of these two shitty IFS, but it's working. So it's good for now
if (axisSign.y > 0 && axisSign.x < 0 && axisSign.z < 0) {
offset *= -1;
}
if (axisSign.x < 0 && axisSign.y > 0 && axisSign.z > 0) {
offset *= -1;
}
4c7809 No.14818640
>>14818415
Nah, I just saw it on social media and figured I'd share.
7c5d9d No.14818685
Do any anons here have experience with Gamejolt? I've been trying to look into alternatives after the Steam debacle and it seems solid. Their terms of service, from what I've seen, don't seem to prevent any content from being posted other than the usual legal shit, and asking for things to be tagged properly.
4cea06 No.14818921
I started reading C# Yellow Book and I think it might work out, I already know a handfull of basic shit, dunno if they'd translate to Game Engines though
4cea06 No.14818923
>>14818685
On gayjolt you're gonna have to confirm your identity to earn shekels
d2b3a9 No.14818943
>>14818260
busy debugging a random gfx driver crash that only appears every 3 hours or so of continuous execution. CodeXL is helpful, and I've fixed many opengl issues, but still crashes. Tested on Intel gfx drivers and no issues. No access to AMD hardware. Just the NVidia driver dll crashing occasionally.
d2b3a9 No.14818953
>>14818334
"Engine development" is really not a thing unless you are working on a large team.
If you mean you would like to write your own 3D renderer, then go ahead. Download GLFW and build off of one of the included examples.
859f6e No.14819069
>>14806187
>>14812657
I bought it approximately 68 hours ago.
c3c718 No.14819629
>>14818923
Is it even possible to make money from a store without?
0b4b57 No.14819669
okay anons don't laugh but how the fuck do I optimize shit properly? how do I track down the function causing the most computational cost? i'm using Visual Studio, so honestly if you could just nudge me in the right direction that'd be great.
My shit's currently like a bag of bricks on my CPU and I'd like to fix that.
adb70d No.14819819
I was looking in my saves of some old builds of Sigma II. The build on the right is the engine on 5/21/17, while the build on the left is the current build of 5/20/18, a year later.
>>14819669
Doesnt visual studio come with a profiler that lets you tell which functions take the most time?
0b4b57 No.14819846
>>14819819
>performance profiler
>directly under the debugger option
>i didn't see it the twenty times I tried to fix perf using a normal debugger with breakpoints
>mfw
thank you very much, anon - I'm going to go contemplate the extent of my retardation.
adb70d No.14819930
>>14819846
Everyone who has used Visual Studio has had their time wasted every step of the way trying to navigate that shitty UI. Don't worry about it anon, one day you will remember where all of the stupid little buttons and drop downs are to make it do what you wanted it too (sometimes)
c3c718 No.14819953
>>14819819
You have much to be proud of, anon.
1226c7 No.14820056
>>14819930
Or drop it for emacs, cmake, gdb and valgrind.
5d2e4c No.14820639
Making more props. Probably just going to finish the living room for now and continue working on her logic.
>>14819669
It's hardly about what functions use the most cycles and more of the nature of your program. Which functions get called the most? You should be familiar with Unity's documentation because sometimes it tells you if a certain function might incur a performance hit. Do some random tests first.
d9ae5d No.14820689
minor ease of life features
right clicking items in your inventory will attempt to equip them based on their primary item type
if you already have something equipped, it will swap them
right clicking equipped items unequips it if you have space in your inventory
consumable items are still handled the same, they get their effects executed on right clicking
b5fccb No.14821093
>>14820056
Emacs is just a gigantic time sink. You spend more time fixing the finickinesses of it than you do actually coding.
>>14820199
Vim is for pussies.
3c17f8 No.14821351
>>14818921
>dunno if they'd translate to Game Engines though
They won't.
The upside of engines is that there is a lot of magic already done and ready to use
The downside is that they sometimesmostly implemented in some arbitrary way that you will have to learn and it's the type of knowledge that can't be reused anywhere else.
0b4b57 No.14821486
>>14821361
Anon, use hooktube.
>>14819930
Thanks! I keep hearing that VS is babby tier and getting embarrassed when im confused about the way certain things are done.
>>14820639
The reason I was thinking about it function-wise is because I currently have a process that generates terrain, and then a mesh to represent that terrain, and shit takes an unholy amount of time to do. It makes use of maybe 20-25 functions and uses each one tens of thousands of times. I wanna see what functions I should be prioritizing and rewriting because at the moment I have a clue, but no definitive answer as to which part of the algorithm is causing this heft.
>Unity docs
UE4 here, for better or worse.
3c17f8 No.14821526
>>14821486
>UE4 here, for better or worse
doesn't UE4 have a profiler window?
0b4b57 No.14821586
>>14821526
i'm gonna look, I'm not actually sure because I've never needed to use it before. Pretty much all game code I've done before was just you know, gameplay framework.
dd3f6d No.14822102
>>14821486
>UE4 for better or worse
It better not be fuckin blueprints. I mean they can help for quick prototyping and all, but don't use them for performance heavy things.
>>14819819
Great progress! Keep up the good work, anon.
a51a1d No.14822108
>>14818640
where can I follow him then?
19682a No.14822118
>~1000 lines of code just to get decent string functionality in C
I'll have a game in no time
0b4b57 No.14822142
>>14822102
I tried blueprints for prototyping once, and I found them to be a perfect tool to completely and utterly obfuscate your "code" to the point where refactoring shit in C++ later becomes nigh impossible.
I only ever use BP for high-level, slight adjustments - every step of my worldgen hierarchy has a BP class derived from it and used instead of the pure C++ class, specifically because of that, it's just too handy to pass up for this purpose. It's like a workflow lubricant.
Anything else though? Any kinda logic? Fuck that, C++ is faster, easier to write and better to read. It's also more hands-on. I find trying to implement functionality in BP feels like putting legos together with chopsticks.
>>14822118
It gets faster, friend! Just keep in mind to abstract, encapsulate, and comment your code. You're gonna be pumping out functionality at sanic speeds.
b5fccb No.14822217
>>14822118
FYI If you use malloc it wont be any faster than std::string. You need to create a giant buffer of memory, then have your string library create strings from that memory. This was something I only recently learned and it fucked me over hard.
c3c718 No.14822218
>>14822142
Of course blueprints are worse. They're for non-coders. They're even slower to make.
5ad6b2 No.14822278
225130 No.14822283
>>14822118
Just use a fucking high-level language already Jesus fucking Christ. You are writing a game, not yet another badly designed string library.
70cddd No.14822307
>>14822217
"Ideally" you'd allocate all the memory you'll need for your game in one go and "allocate" everything in there manually. Except only AAA games need this, and it doesn't really work well on PC where your amount of available RAM tends to be variable.
19682a No.14822348
>>14822283
There's no such thing as a high level language that satisfies my autism. Although that doesn't mean I particularly like C either.
225130 No.14822388
>>14822348
This autism is called procrastination. Stop making excuses and make a game instead.
0b4b57 No.14822392
19682a No.14822400
>>14822388
Just because there's more work doesn't mean it's not work.
0b4b57 No.14822416
>>14822400
effective work vs. unnecessary work. You could be writing high-level gameplay systems instead of fucking with string processing.
b5fccb No.14822419
>>14822307
>Except only AAA games need this
Why? Are indie games magically immune to performance issues? If I boot up Dwarf Fortress can I have a dwarven city at +5 fps?
>and it doesn't really work well on PC where your amount of available RAM tends to be variable.
Then resize your buffer. Better doing it once every couple of frames than +2000x in a single frame.
19682a No.14822427
>>14822416
Posting in /agdg/ is procrastination, why don't you get back to work fag
0b4b57 No.14822428
>>14822427
Because I'm running a performance profile right now.
70cddd No.14822435
>>14822419
Except DF doesn't have any issues with allocations, it just has too many things being calculated at one time on a single thread. What I meant is that you only need the extra performance gained that way when you're squeezing the maximum out of a system, which your indie game isn't going to.
19682a No.14822437
>>14822428
Premature optimization is unnecessary work.
0b4b57 No.14822445
>>14822437
Who said it was premature? I'm in the final stages of implementing the functionality I'm working on right now.
b5fccb No.14822447
>>14822435
>Except DF doesn't have any issues with allocations, it just has too many things being calculated at one time on a single thread.
Yeah. And I'm sure slow as shit system calls (memory allocations) and cpu cache have nothing to do with it.
>which your indie game isn't going to.
This mentality is why so many indie games have shit performance, hog resources and have frame drops, despite looking like shit or being 2d.
225130 No.14822484
>>14822447
Indie games have shit performance because they are badly written. Even Minecraft wasn't slow because hurrrr java is el slo x-dddd but because Notch couldn't fucking program his way out of a paper bag. The retarded micro "optimizations" (i'd bet money that what you write will perform worse once it has the features of any sane HLL) you are doing are a complete waste of time. Write your game in something that makes it easy to write the game, and then optimize once you have performance issues. And when you do that, profile before doing anything.
Unless of course you want your game to be eternal vaporware. What's the matter, afraid it'll be bad?
d2b3a9 No.14822527
>>14822118
hey, check out STB on github. Many very useful pieces of very portable C code -
github.com/nothings/stb
I think you might be interested in stb.h and stb_textedit.h
b5fccb No.14822542
>>14822484
>Don't prematurely optimize XD
Only works at the micro level, which is the only level you java/C# coders seem to think at. I didn't read too much of your post though, because it's the same regurgitated garbage every webshite uses and it's in every thread now. It's kinda sad because it shows the state this threads coming to. Just a few threads ago we had a fucking sysadmin trying to wave his dick around because he could configure servers.
d2b3a9 No.14822548
>>14822217
This leads to a critical piece of information that many newcomers are unaware of: Never allocate memory per-frame. Never ever. Memory pools are the easiest way to avoid this. Other, more specialized routines may benefit from a ring buffer or something more exotic and specialized - for example, particle effects, where you don't really care that much about individual particles.
allocate all of your memory up-front, or on level-load, or some other convenient point. Dynamically allocating and resizing buffers every few frames is never required, never a good idea.
0b4b57 No.14822554
>>14822542
The core population seems to be doing a good job at weeding out and bullying the cancer into submission, though. Which is good.
3c17f8 No.14822857
>>14822348
Assembly is waiting for you with open arms
d2b3a9 No.14823721
>>14822484
Picking good data structures for your game objects and then implementing a sane memory handling scheme isn't micro-optimizations. It's what is required to build a game that has more than a handful of dynamic objects.
If you are building a platformer with a few onscreen objects then sure, maybe it's over-optimization for the task. But then why are you building it by hand and not just using GameMaker?
What kills one-man indie projects (literally everything in this thread) is A) overly-ambitious scope, B) scope creep and feature creep, C) spending time coding things you don't need because it's fun and not just picking the simplest solution that supports your end-goal, no matter how disgusting and filthy a design it is.
0b4b57 No.14823823
Alright, so shit seems to be pretty cash at the moment, I fixed the issue with the algorithm that created excessive geometry and I'm left with this beauty that I just need to stitch up and it'll be a nice, continuous mesh.
Multithreading is up next! oh boy why do i feel like it's going to take me half a month to get to work properly
>>14823721
A) The overly-ambitious scope thing fuckin hurts me to read not because it's not true but because I fear it applying to me. I basically aim to make a not-shit /v/approved cursed-runes game that is meant to be actually fun for a change without getting bogged down by the trash fluff mechanics.
That sounds like overly-ambitious scope and it scares me but it feels to me like I got a handle on the concept almost entirely from the programming standpoint (only saying almost entirely because I'm anticipating a Great Filter of sorts to suckerpunch me).
B) I don't understand how the fuck this keeps happening to seemingly competent people. Just like, make a GDD and stick to it. And when you inevitably modify the GDD because you're burning with an idea, let it sit for a month or two until you cool off. It's basic project management and planning, not rocket science you guys.
C) People who don't use libraries and code shit on their own out of principle amaze me.
>>14822484
Modded Minecraft runs faster than the vanilla game even when running 150~ mods at once thanks to a community patch that makes Minecraft actually optimized. I'm amazed at how poorly the core game is coded.
4cea06 No.14825307
>>14821351
thats good because i didn't understand a bunch of other shit if i wont have to know 100% of c#
bb95ae No.14825380
>>14822542
>I didn't read too much of your post
No shit.
>>14823721
That's what I said. But implementing your own strings doesn't fall into that category, not even remotely.
489939 No.14825964
Any opinions on Gamemaker? I'm basically a rank amateur, I have some previous programming experience though nothing on the level of making a game, a lot of great games have been made with it and it seems to move a lot of fluff out of the way to help with just like making game.
Idea is for a Roguelite turn-based RPG kind of in the vein of Into the Breach but with more emphasis on action than problem solving. I'm not really a fan of ITB so it wouldn't be anything like that if it ever comes to fruition. Ideally I want to capture the fun and class variety of Dofus/Wakfu but in a singleplayer format and with shorter combat times.
d9ae5d No.14826003
>>14825964
it's shit
just use godot
327823 No.14826145
>>14825964
It's ok as long as you're not trying to make something revolutionary.
If you have no experience with any engine then consider going with Godot instead, since it's a free open source non-jewish engine unlike Gamemaker. There's less tutorials and documentation for Godot though so I'm not sure if it's as beginner friendly as Gamemaker.
d9ae5d No.14826512
ripping off skyrim one step at a time
adb70d No.14826997
>>14822118
Why do you need the extra string functionality? Will you actually use all of it in your project?
Maybe the best approach is to start on your game, and then only write extra utility functions for strings that the standard library doesn't provide as you actually need them.
19682a No.14827303
>>14826997
Because I'd lose my mind if I wanted to have anything more than hardcoded words with C's functionality, plus I hate having to think about what length every string needs to be and I don't even always know.
Like if you want to put something into the beginning of a string, you'd have to jump through hoops to make that happen and make sure it won't fuck up. Now I can just do arr_injectC(&array, 0, "here we go") and I don't have to give a shit how long either of the strings are because it's dynamic length.
>make a game and THEN add features that should be in the base language
I'm not sure how you think this way. I made this feature because it's going to make everything involving strings less painful from now on.
adb70d No.14827343
>>14827303
If you want to put something in the beginning of the string use strncat. If you want this stuff to be in the base language (which it already is, but whatever…) you could use C++.
adb70d No.14827346
>>14827343
maybe my post is a little dumb, telling you to use C++, because there is a middle ground, I just wouldn't bother writing an "arr_injectC" function until I ran into a situation in my code where I would use it. That's all I mean.
19682a No.14827375
>>14827343
>If you want to put something in the beginning of the string use strncat
That won't work unless you create a second string and move them around, because strncat adds to the end of the string. And then it'll break if it won't fit, and you have to go back and rethink how big the array should be.
adb70d No.14827600
>>14827375
strncat is bounds-checked, you're thinking of "strcat". But I see what you mean. You could write a function like so:
char *strnappend(char *dst, char *app, size_t count){
size_t applen = strnlen_s(app,count);
memmove_s(dst,count,dst + applen,count - applen);
strncpy(dst,app,applen);
return dst;
}
Which doesn't require you to create another string to dump the characters into, if this pattern comes up in your code a lot. It doesn't increase the size of the buffer, which you want. But I try and avoid allocating memory outside of initialization functions, so I would just try solving it by making the buffers bigger.
9533eb No.14828509