/agdg/ - Amateur Game Development General

Remember agdg

>doge go Raawrr
:3

postu proguressu

Hi agdg. Can you still make money from the tower defense genre?

How much do you want to make, over what timespan?
How clever is your game? Is it a straight genre game? Is it TD x another genre?
Are you willing or able to add some other element to increase marketability, such as waifus?

I managed to unfuck my animations but I still don't know what happened. For some reason, even in Blender 2 animations were playing at once, ie. the shooting animation which only uses the upper body bones, and the transform animation which uses the whole skeleton, were overlapping.
This never happened before and I don't know what changed.

Show us your game studio, agdg

youtube.com/watch?v=USPoN-GwzVk

Sure, if you have very interesting take on the genre, or time machine back to 2010

please use the term bishoujo
thank you

Ideally im willing to do whatever it takes (except waifus, sorry i cant draw for squat). Im more concerned with if the genre is even still marketable. Would it require a lot of wow factor to succeed, or does the basic formula still show results (like 1000 people playing for free = results)

Im not looking for a get rich quick with a TD game, just get a decent amount of people playing it

>dog in the preview image
>no dog in the video
what a fucking asshole

how does someone sell a game written in java? its not exactly hard for people to just download it from other people/hack it

Any recomndations for a 2d animation software? I want to make some spritesheet effects, but making it all in photoshop is unweildy as hell.

and whats stopping people from decompiling it and reverse engineering parts of it?

Same poster, I assume?

Dude, there is NO genre in 2017 where you can automatically succeed with an average competent game. You have to make something that is unique in some way, or hits some niche (porn or furbait games).

You should be more specific with your goals, too. What do you mean by "results". Do you have an actual target amount of money that you want? Do you just want to make something that people play, and don't care about the money? Assuming that it's on mobile, 1000 people total playing your game is simultaneously nothing, and far more players than most games receive. Ads WILL get you money, but the conversion rates are so low that 1000 players will give you a dollar or less.

Any singleplayer indie game can be pirated, and some portion of your audience will do so. As far as decompiling, some people will do it for fun. If you care about such things, use a different language. Java's not that popular for gamedev anyway

Nothing. Nobody cares about your shitty code though.

LIBGDX is a fucking mistake.

Different file types for different things, user:

CSV for game data where every instance has every or almost every value defined (examples: NPCs statistics, location names and coordinates, tileset definition)

JSON (with a suitably lenient parser) for other structured game data.

Simple property files for global constants and defaults.

Jokerscript files (or similar) for dialogue.

XML (with strict rules as to what is a valid value where) for game saves.

Nothing is stopping anyone who's not a bumbling moron from getting JD-Gui and a (de)obfuscator and decompiling your code, then having a laugh at how shitty it is. For an example of this in action and how well it worked out for the creator of the game, see Minecraft.

Are you saying minecraft is shit?

The code may be shitty, but the the (relative) ease of modding certainly helped it get big

alternativeto.net/software/opentoonz/

I care about my not-so shitty code, man. :(

What would you say makes it so shitty? As someone who has read quite a bit of the source, I'd have to say its better than 99.99% of the shit I see. But then, Ive only seen the modern source. Ive no idea what it looked like in the earlier days, before mainstream popularity.

99.99% is, btw, not a literal expression, but an exageration.

It's been a long time since I've touched it, and I don't use Java often. Just clarifying the intent of that post, which isn't even really about the quality of the code.

>As someone who has read quite a bit of the source
can you provide the source?

What're you gonna do? Give up?

>indie gamedevs thinking their shitty code is worth stealing
always makes me laugh

it's not corporations, it's other shitty indie gamedevs stealing your shitty code

Uhh, i could... but its easier if you just go decompile it yourself. Literally just google mcp and you should be golden.

so i have to google it myself? Typical aggydigy.

I can think of ONE case where an indie dev has code worth stealing, and that dev wisely hasn't put out a demo yet.

Well nigga, I aint even near a computer rn. Im 3 hours from home actually, and am not leaving to go back anytime soon. You can wait if you want, lol.

phoneposting needs to stop

----[ Recap ]----
Game: Space Dementia
Dev: rez
Tools: SFML/C++
Web: rezydev.tumblr.com
Progress:
+ Experimenting with new player sprite
+ Started new enemy animation and logic
+ Tweaks to current boss to make it less aggressive
+ Added cloth as serializable and interactable object
+ Fixed rare bug in collision handling
+ Visor light now rotates the same way as head, flashes when directly straight
+ Changed crosshair look and smoothing
+ Tighter boomerang trajectory
+ Fixed problem where platform movement would get stuck

wut gamu

Fleech's robot walking tech

Yeah? Whats wrong with phone posting? Even if i was home, i still dont know how i feel about uploading decompiled copyrighted shit, copyright reasons and such.

I have an incest problem!

I'm making a breeding sim thing and so far it's all working smooth enough, with the 'basics' already in place and bug free ('breeder' objects mate and produce offspring, passing their 'genes' down in the process). I want an easy way to display to the player the 'incestyness' of any given paring between individuals with various levels of shared ancestry. I have two thoughts in this regard:

1) Do a branching search up to n branches up in the family tree and generate an incest-level based on -potentially- shared genes. IE immediate relatives all get the same incest rating, then each successive branch 'away' halves this rating.

2) Do a direct comparison of the 'dna' (really just a 1000 character long string of arbitrary characters with half inherited from each parent), the incest-level is a direct percentage of shared 'genes' between the breeding pair.

1 is how incest is typically assessed (fucking your third cousin isn't as bad as fucking your sister) but 2 gets down to the point much quicker (what is the likelihood the offspring will inherit a specific phenotype?). As a micromanager myself I'd much prefer to have 2, but maybe 2 gives too much information and people might prefer rolling the dice and not knowing the exact odds.

No, the code is shitty (for a game), which is one half why it runs like a one-legged dog. Have you seen the kind of GC activity that game creates?

Are you weston edwards?

>2017
>Not being a lazy NEET that has never completed a project since getting into programming 4 years ago.

In most cases, choose 1. It's more ambiguous and more intuitive. Potentially don't even show a rating, but use something more descriptive, like their relation (cousin, etc), or display the tree for the players to parse manually.

It depends on the rest of your design goals, of course. Option 2 can be unlocked or a special ability. Maybe only at the level of specific genes.

post game
or are you just ideaguying?

Honestly no, i have not used analytical tool on minecraft. From experience, i have not noticed anything too out of the ordinary. The chunks probably arent too gc friendly, but honestly ive never noticed any serious lag just because of the gc

I guess I should start doing the recap again. it'd be good motivation if i could get into the habit but i rarely feel like doing it.

----[ Recap ]----
Game: Idol
Dev: Marson
Tools: GMS
Web: twitter.com/MarsonNG
Progress:
+ Did a few new animations
+ Added a dodge roll
+ Applied slope physics to rolling state
+ Optimised old hook shot targeting code

>What would you say makes it so shitty?
Java is a "don't think about objects or the memory they take, creation is nearly free and cleanup is done by the garbage collector behind the scenes" kind of language. Great for writing, but a game has the choice between producing millions of small, immutable-for-thread-safety objects (positions, vectors, ...) or thousands of small, mutable and expensive to multithread with objects of the same time and the GC runs hurt your performance. A lot. And lead to "FPS stutter".

No problem with a JavaFX business app or server-sided things where you can just throw in another 64GB or more RAM or a bunch more machines/blades at the "problem" and be done with it, not that big of a deal if you have a turn-based strategy game either, but a big problem for an interactive game.

Also, not knowing where in RAM your objects are means you're basically losing 10x the performance you could have via thrashing the CPU data cache every so often.

What you need to do (and what libraries like LibGDX or Artemis help you to do, by the way) is use direct memory buffers as pools and create objects from those pools. Re-use object space after you don't need the object any more. But the code to do so looks very little like normal Java object creation and usage code - you have to basically fight against the "Jave best practices" advice you'll find on the internet in many places.

Minecraft was coded without these insides, and it makes for a very fun game with a shitty (for a game, that is) code.

----[ Recap ]----
Game: Bullet Hell Lane Shooter
Dev: ADMAN
Tools: Unity, Blender
Web: sgtadman.itch.io/untitled-bullet-hell-game
Progress:
+ Most of the UI for the content creator is done and working
- Need to do saving and loading stuff

I'd use version 1 (it's called consanguinity, by the way). Leave the actual values for the debug/cheat mode.

fishing with a pole works, visuals still need work though
this isn't quite what I was planning to do with the fishing pole but I think it's funny and more engaging than what I was going to do, which was making "reel" throw the fish out of the water towards the player unit
now I think I'll just make reel gradually shorten the line if you click/tap quickly or something so it's easier to drag the fish out of the water

I've been thinking of switching to C++, and just might do it now (for the client, any server stuff will always be java-based) as LWJGL is nice and all, but there are so many times where i find myself wishing i had actual pointers. What is this cache thrashing you speak of? Not once have i ever heard such a term.

cute

----[ Recap ]----
Game:Devil Engine
Dev:shmoopdev
Tools:GMS,photoshop,aesprite
Web:shmoopdev.tumblr.com/
Progress:
+Stage 4 miniboss mostly done
+Prototyping the 2nd half of stage 4
+Bosses now have alternate hitsounds when low on health
+Fixed some performance issues related to bullets not deleting themselves
-somewhat slow week

thanks

right now the 'game' is just the code and some ui dangly bits, I generate a population of breeder objects and they go at it for a couple hundred cycles. Then I can make some neat family trees and click through names to see parents/siblings and such.

Yeah I suppose 'both' could work. In the end if this shit ever makes it to demo-day I'm sure I'll have some feedback to work off of anyway. Right now my design goal is 'make a sandbox that vaguely simulates genetic inheritance', because I always had a lot of fun doing things like that in games that allowed it. There's player controlled breeding (pick two compatible units and make an offspring) and AI controlled breeding (right now, randomly pick something and try and make offspring with it). Not much of a game at the moment but that's life.

Gc cant occur if there is no garbage to collect tho, right? Like, during runtime, as long as everything just stays the same (example only, reality is unpredictable), there theoretically shouldnt be much if any gc going on

when you access memory a line of data is copied into cache on the CPU where it can b more rapidly accessed. If you access memory in lots of random spaced out locations you get cache thrashing because the data you're accessing cant be cached in a line. it's best to process data by iterating over the array elements in order if you can. google it for more

----[ Recap ]----
Game: Ghost Knight Victis
Dev: Anomalous Underdog
Tools: Unity, Blender, Substance Painter, Photoshop CC
Web: hypertensiongamedev.tumblr.com
Progress:
+ Finished making code that prevents hitting characters on the other side walls/barriers
+ Cleaned up some code and moved common settings among characters into an asset file so they all share those values

When neither you, nor any mobs, nor anyone else on the server move, and no physics updates or landscape generation happen, Minecraft doesn't indeed generate much in the way of data for the GC to collect.

>I have an incest problem!
2 actually isn't testing for incest, it's testing for genetic similarity. it's entirely possible for a couple to be genetically similar but not family.

Doing interger/fp math doesnt create garbage, does it?

In Java, it actually can, unless you're careful (thanks to automatic boxing/unboxing of primitive data types). Generally though it doesn't, and in particular Minecraft (at least reasonably current versions thereof) avoids much of it by saving chunk data in big int[] and short[] arrays.

----[ Recap ]----
Game:UW
Dev:user
Tools:Unity
Web:twitter.com/DuqueHayashi
Progress:
+ Broke AI down into smaller pieces to add more behavior options.
+ AI optimized
+ Grenade Added
+ UI Script cleanup
+ Armor/Cover bonus started
+ Skill/Ability set revised

Alright man, thanks for the tips. Ill be sure to put them to use soon. :)

just you wait

>it's details

It is details, indeed.

How come only native English speakers do this type of spelling error?

What are you talking about? I am not making a western erotic game.

Yeah, that's kinda the point. I want a pokemon experience on my phone like pokemon red\blue. Although there will be many differences: The story will be different, different monsters, it's set in a fantasy medieval time, the combat will be more like duelyst and essentially will be smaller in scope for the first version.

----[ Recap ]----
Game: STD
Dev: Toasty
Tools: C++\SFML\AseSprite\Visual Studio
Web: null
Progress:
+ Data driven projectiles
+ Waves loaded from file
+ Levels added
+ First refactor done
+ Towers now correctly identify closest target
+ T1 towers can now only have one target at a time
+ Art update (Eyes, Orcs, Tiles)
- Need level selection
- Need tower upgrading

>CPU: main 25 ms average

What?

I am doing a procedural 3D game with thousands of objects visible on the screen at the same time and never go above 5 ms.

This is the power of Unity.

He probably hasn't refactored the mono-behaviours away yet. Your objects look really simple from that screenshot as well.

----[ Recap ]----
Game: Jappu -Heart of Karina-
Dev: takkongguy
Tools: rpgm
Web: takkongguy-blog.tumblr.com/ takkongguy.itch.io
Progress:
- Having hard time drawing grass tile

recording in fraps makes it jump up a little
it goes down to 6.5 ish when I disable all my fish... going to need to do some optimizing I guess

for a person's native language, they learn it by ear because that's what they do growing up. so if it sounds close enough, they end up using that without thinking about grammar

Fuck off fag. I like sitting on my back deck and phoneposting. All a pc does differently in regards to the board is that it allows a faster typing speed (for me anyway) thus more frequent shitposting. Fucking basement dwellers need to get out and see the sun more often.

His are even simpler. Look at his Tris and Verts.

Do a profiler.

why is this thread so fucking fast holy shit

While more tris\verts is pushing more stuff to the GPU it will impact the CPU very little compared to actual behaviour which is what I am referring to.

That sort of behaviour does not behoove you.

I can only guess at some physics checks outside FixedUpdate or some while-loops going astray, or maybe the use of OnTriggerStay instead of OnTriggerEnter and OnTriggerExit.

Fuck thats painful to look at. Gives me phantom pains of ripping off a fingernail. Jesus. Did that dog die?

Fellas is it that hard to make multiplayer with game maker?? should I just start with unity?
Im making a 2D Dungeon game

>dog
i think it's an horse man, and he was dead when they did it

it's a mushroom.

I can look at wrekt threads all day, but this is just painful to look at everytime i see it

please delet that picture

...

I want to dev gory games

Do it then feggit