Potentially Acceptable Rendering

[@maliciouscompliance6489]

RIP Matthewmatosis. He fell into the engine dev void. Still rooting for you, man!

[@alexandrechoin]

Probably just echoing here and a late comment: I really would just make a bunch of crummy game jam games in gamemaker and have people play them. Game design really doesn't come out of the theoretical or having an airtight engine. It comes from throwing something together based on a few assumptions and realizing how people typically react and basing your decisions on that experience. You can have the jankiest thing ever but if it hits some basic design goals, it can still be fun and more importantly faster to learn/iterate from. The mental trap isn't just diving down to an almost assembler level just to have a basic thing going but probably having an extensive youtube career that talked about games and the nature of how they're good and suddenly having this pressure to back that up somehow with perfectionism. It's a dangerous cycle. As someone who shipped a game on steam and several on mobile. It's SO HARD to make a game, even with gamemaker or unity, even with existing templates or cloning a previous game. It's hard to ever know how much time is actually needed because so many tasks have unknown time variables and it's a miracle a game comes out at all. If you're just interested in the computer science aspect only and only that then sure that makes sense. But if the goal is to make a game in your limited lifespan, i'd really suggest the make things "good enough" fast approach. Looking back i wish i didn't get burnt out on 3 year long projects so much.

[@sneakman9873]

Somehow I knew this was going to end up with you making your own engine. It just seems like the most logical endpoint for some game/games made by Matthewmatosis.

[@RoboGuy2K]

The final title of the game is just gonna be "Potentially Acceptable Mushido"

[@serpentingestor3]

more like Potentially Passable Pagodamania

[@matthewmatosisextra]

The compression artifacts are particularly bad this time because the camera is much more zoomed out but that's just the video, the fleas look the same as they did before.

[@augustdahlkvist3998]

I fear that you are falling into the perfectionism trap where you will spend forever dithering on small details and never make big strides because the engine/foundation isn't perfect. I've seen many gamedevs do this, a classic(and extreme) example is the Australian KZbinr "randy" who rewrote his game more times than I can count, and then eventually scrapping the whole thing. In the end all he had to show for it was a character walking around a test environment. Making your own engine in c++ almost always ends in disaster as your project gets plagued with undefined behaviour, memory leaks, and stale pointer issues. Especially since you are going as low level as openGL and SDL. I've worked on projects that did this and it happens every time. Having said that, I do hope that you don't do another rewrite and instead get on to making some content. You don't want to spend months on the engine only to realise that the game isn't as good as you had thought. Plus, design will be the brunt of your total work, so the quicker you get to that, the better. I don't want to discourage you or anything, just some friendly advice. I am looking forwards to the release of this and wish you all the best ☺️

[@ebnflo]

The line between feature delivery and tech debt is a painful one

[@augustdahlkvist3998]

​@@ebnflothis is true, but the act of deciding to make your own engine creates metric tonnes of tech debt, when you consider the amount of work and bugs it will generate, compared to using an established engine.

[@matthewmatosisextra]

Understandable concerns and I'm definitely the type of person to fall into the perfectionism trap. One thing I'd like to say in my defense though: The prototype is still being made in GameMaker and when I was putting it together I hammered in features as quickly as possible, incurring a bunch of technical debt just to see if it was a viable idea. My current plan is to keep adding features without much concern about the structure until we have something that looks like a real game. Then, if we're actually going to release it, I'll do one significant re-write (still in GameMaker) to tidy everything up once I know what the requirements of the new game are. That seems like a reasonable middleground to me and I've stuck to it so far.

[@MenacingSnail]

I agree, but Matthew is at least pretty cautious about what he tries to do. He's not some 22 year old hot shot bragging about how he's making an engine. It's just the solution he genuinely believes will be best. This in my opinion, bodes well.

[@augustdahlkvist3998]

@@matthewmatosisextra Ah fair enough. The only other thing I would say is to really make sure you are using profiling when you do these optimisations. You spoke about a couple of things optimisations which I am not sure are necessary for your kind of game. Optimisation without using a profiler is like trying to solve a maze while blind. Looking at the profiles are your eyes and ears when optimising a game. There are two reasons for this: 1) You should make sure that you are actually optimising the bottlenecks of your game. You could spend a couple days working on a really clever system that bundles adjacent tiles of the same type into two triangles, but is geometry really the bottleneck of your game here. You don't know if you are saving milliseconds or only nanoseconds. Also, there are often things which seem fairly innocent but take a surprisingly long amount of time, and these are easy to miss without a profile. 2) Even though conceptually some algorithm seems faster, you will often find that in reality it isn't. I can't count the number of days I have spend working on optimisations which ended up being slower than a more simple/naive approach. There are so many factors to juggle, it really is impossible to say if something will be faster without actually looking at a profile. Going back to the tile bundling example, if you tried this on tiles that moved, you would need to reconsider the adjacencies of tiles every frame. I reckon the cost of evaluating that + building the geometry would far out-weigh the nanosecond savings on the GPU. But like I said, don't take my word for it, trust the profiler! I can highly recommend this profiler: github.com/wolfpld/tracy If you setup a couple of macros you can instrument your functions quite easily.

[@DanF2006]

Is this Matthew's ulillillia arc?

[@MellyMellouange]

having functions executing in a linear order is called "Procedural programming"

[@gloomywheel8378]

its actually just torture watching this video when halfway through you realize no fleas are going to make it on top of the 7x7

[@gloomywheel8378]

​@@Senpaipls253 please lord just timestamp it for me

[@gloomywheel8378]

@@Senpaipls253 i just rewatched the video on 2x speed and i must say you are probably the most evil person i have ever met in my life

[@matthewmatosisextra]

It's interesting how there are certain paths they never end up taking. After a few minutes they stop falling down the right side. If they had their glide mixup they would probably spready out more evenly.

[@gloomywheel8378]

@@matthewmatosisextra thinkin about this situation the culprit for a flea never reaching the 7x7 block might be the sloped block to the right of the step block. it seems very visually feasible that a flea would be able to jump from the edge of the step block to the 7x7 block. However its impossible for a flea to reach this (very close the the end of the third step position). A flea can land at the end of a "step" but, because of their jump distance, to do so it would have to start it's jump at the start of the same step. The problem is that the fleas always enter the step block very close to the beginning of the first step (because the slope block is lower than the step block the fleas have a lower horizontal range than if the were jumping from a higher platform to a lower platform). Because of this the fleas have only have one possible path, land on the very beginning of the first step -> jump to the end of the first step -> jump about 3/5 of the way on the second step -> jump to around the middle of the third step -> fall down the hole. I wonder if because of all this it might be a general rule of systems like these that the higher up you go, the more constricted the fleas become since there are a constricted number of paths that can get that high. It's really weird how this system actually can evoke irritation because it really looks feasible that a flea can jump from the end of the third step to the top of the 7x7 but what's hiddenly not feasible is that a flea will never make it to the end of the third step.

[@lucyfrye6723]

That change to a 4-vertex quad should be trivial, replace GL_TRIANGLES to GL_TRIANGLE_STRIP in the draw call. It really is that simple.

[@thomchiovoloni3877]

i came here to say this too lol

[@yaccobarata6574]

I'm just here to say no fleas make it to the top of the top-left square, and no amount of wishful thinking will help them get there. You may now go back and actually pay attention to what's being said.

[@shade0636]

It's a good learning process to make your own rendering without a game engine, nice job.

[@username449]

I'm a little perplexed at what the purpose of crafting your own engine is meant to be? What wasn't GameMaker doing that this is meant to cover? What is the benefit meant to outweigh the cost? I was honestly waiting the entire video for you to elaborate on why you went and did this to begin with. I can really respect wanting to understand something "from first principles". I'm sure it's something a lot of your viewers can relate to, myself included. But from experience I also know it can be used to justify digressions that just aren't worth the time - especially as an after-the-fact justification. There;s aphoristic phrase I'm fond of that says something like "you never really learn to understand mathematics - you just get used to it". I think that's true for a lot else. Philosophizing aside, I'm just hoping that eventually you find the right way and start carrying the momentum you've been building up over the past ~2 years. PS. Watching the fleas was nothing short of hypnotizing. Sometimes certain pairs of fleas looked like they were bouncing off of each other like "Burt"s in Yoshi's Island. Might be a new enemy concept in there somewhere.

[@chrispeng5502]

"The game may be full of bugs when it releases" lol Matthew already talks like a real game developer

[@adammalm4733]

Big fan of your stuff Matt, really excited to see what comes of this engine! also thank you for the rain world and nex machina recommendations, especially nex, I'm currently trying to play a lot more twin stick shooters at the moment. Anyways, love your stuff, keep doing what your doing.

[@ebnflo]

its so frustrating that they are on this set path and will never get to the top block

[@swisspunker94]

always enjoy your videos man! Im sure you learned a lot from doing this. As far as using it for your real game, what would worry me is a lack of support/advice. You're gonna run into some issues that you cant just google the solution to. If you have no one around you who has gone through that journey it will be tough to sort out all the bugs/optimization issues of the engine. If you really enjoy that type of work and have the time for it then do it! If it really motivates you to use your own engine then do it! But if your primary concern is getting a game made then I would say continue with game maker. To me the benefits out-weight the negatives.

[@lllavoie]

I've done small projects with thousands of unoptimized 3d drawcalls and AABB collision checks per frame on a crappy 8 years old laptop with an integrated GPU and it hasn't chugged once at 60hz. I wouldn't overthink it too much until it becomes a problem, the biggest optimizations usually come from good partitioning.

[@47ejecting2]

Your manner of explaining to the layman without condescension is greatly appreciated! As one such layman, I wonder if you have a well-sorted list of priorities for your engine/"thing" - i.e., which additions would be luxuries vs. necessities. Regardless of whether you go back to Game Maker or not, learning from first principles is always valuable. Apologies I don't have anything more substantial to say, but I look forward to what you end up releasing!

[@wushiba]

Thanks for covering games like Rain World! If not for your twitter and some random person's question and your recommendation then I would have never found that game. Good luck on this project of yours! (Btw Returnal is a pretty cool game too, I'm having little bit of Rain World vibes at certain moments)

[@LearntheLore]

Platform compatibility is definitely *the* thing that's kept me away from rolling my own engine after years of using Game Maker, since I also really only use a very small handful of built-ins in my own projects. I've also been keeping an eye on Godot. Maybe someday I'll make the jump. In my own experience, it hasn't been thaaaaat big of an issue to save optimization for a kind of "clean-up" phase at the end of development while working at above the engine level, since most of the real "big performance gains" stuff tends to live in an obvious 10% of the codebase, even if you're not thinking about that too much when you're writing it. This is assuming you're already broadly in good programming "habits", which it seems like you are. With the smaller (assuming "small-ish") project you're tackling now, you might find it valuable to just try to bang out something playable as quickly as possible and then apply the insight you gain from doing that to either refactoring the more "performance-critical" areas of the code, or else to the planning/early implementation phase of whatever the next project happens to be. This is something that you can get away with reasonably well at small-scope that's much more difficult at medium or large scope. Odds are good you've heard similar advice before, though, so I'll append to it that your preferred way of working is the best way to work. Best of luck.

[@GregBadabinski]

This is my favorite box of fleas. EDIT: very impressive! I think making this leap will really help advance your understanding if nothing else.

[@siphillis]

The Matosis Engine

[@Majikn]

The FleaBox Engine

[@joebailey8294]

Pizza Tower was made in game maker and the creator of that game seems to be a follower of yours

[@matthewmatosisextra]

That's cool, I haven't played it yet but I'm looking forward to trying it. I think UFO50 is being made in GameMaker too, probably many others as well.

[@TheJols]

@@matthewmatosisextra Pizza Tower is probably the best game of the year for me. Actually thought of you when I started playing it.

[@joebailey8294]

Matthew “brainiac” matosis

[@SaadTheGlad]

Hey Matthew, I'm just happy you're uploading on KZbin still, please continue to do so, on this channel or the main one, it doesn't really matter. I can't count how many times I've slept to your videos, very soothing voice.

[@wushiba]

I agree. Matthew should make game manual reading ASMR videos.

[@veselekov]

Excited to see your growth. Even more excited for your debut

[@bugwrld8891]

cant even hear what youre saying because im so focused on the fleas being unable to reach the top block. love the little guys tho food work

[@xThuby]

For a map editor I'd suggest Ogmo or Tiled. You only need to write the code for reading the map file data.

[@matthewmatosisextra]

I honestly forgot there are dedicate map editors like that, thanks for the reminder.

[@numidium3]

I get the desire to do things from "first principles" but keep in mind that taking that to its extreme would mean writing everything in machine code. Don't be afraid to leverage the work that's already been done for you. An existing engine is probably the way to go.

[@EnvyTankos]

I've always associated your voice with Niall's on the SleepyCabin Podcast and if I come to find you look nothing like him I will be disappointed

[@sbmoonbeam]

Nice, interesting you've decided to move away from game maker. I've been working on my own thing from scratch in Vulkan, finally got things like polygon collisions and blending working with a mish mash of glfw and sdl for audio . There's obviously more learning to do but I think it's worth it to understand the underlying graphics systems and get into the GLSL business of what shaders are actually doing.

[@MiguelMarques1402]

Hey Mathew, you're a big inspiration of mine!! I wish you the best of luck with this project! I have a barely related question though, what are your favourite Side-Scroll action games if you have any? Seems to be what you're going for here and it's my favourite genre when it comes to gameplay. I would love to try out the favourites of the CEO of gaming himself! Also to see where some inspirations might come from in this project later.

[@matthewmatosisextra]

Ghosts 'n Goblins Resurrection is phenomenal. I'm also a big fan of Rondo of Blood too. The other GnG and Castlevania games are usually worth playing too, some are better than others naturally. For anything faster than those, I tend to prefer straight up platformers like Dustforce or run-and-gun games like Cuphead, Kero Blaster or Alien Soldier. For faster action, Viewtiful Joe is one of the few truly 2D beat-em-ups and has a lot of great intricacies to learn. Katana Zero also comes to mind, great presentation with satisfying movement. I also enjoyed the movement of Luminous Avenger Ix but the camera zoom level is a serious, crippling flaw with that game which can really sour it, still it's worth a look. Not linear but Super Metroid is also a must play. I think that'll do for now, hopefully there's something there for you.

[@MiguelMarques1402]

@@matthewmatosisextra Oh wow! Thank you so much for the reply! I have a lot to look forward to trying! Rondo of blood is the only one i havent played, i really like Portrait of Ruin and it's easily my favourite cause of how the story and gameplay work together, also the music. Also omg i literally forced my friend to play Luminous Avenger IX yesterday and he forced me to play Super Metroid, i just think that's kind of funny. Currently playing Metroid Fusion so i can force him to play Luminous Avenger IX 2. My favourite games of this genre are the Megaman Zero games, they awoke my love for the high difficulty combat heavy side scrollers, from there i consumed all of Inti.Creates games (they made Luminous Avenger IX which scratches that exact ich). That company are the kings of this type of game and i highly recommend you try the Megaman Zero collection, it will hook you in the first 30 minutes. Also despite it's flaws there is a lot to love in Luminous Avenger IX2, and in the end i feel like i ended up enjoying it as much or slightly more than IX1. I actually don't have a problem with the camera thought, or at least i can't really imagine what your problem is outside of not having time to react to enemies when you're going fast which for me isnt a problem since you'll probably tag them and not take damage, am kind of curious to hear your thoughts! This is a bigger issue in the Zero games though but also doesn't bother me much cause i like everything else so much. Anyways once again thank you so much for replying it means a lot to me!

[@chrispeng5502]

Matthew's enthusiasm about GnG Resurrection does concern me a little bit because the game is really niche. It is frustrating and extremely devious. And I played a few NES/SNES platforms, I even speedran a fan-made Megaman back then. But I wouldn't have beaten the game if Matthew was not recommending this game. It was excruciating. One thing I think Matthew was spot on was that he knew if game developers made games he liked, no one would have bought them. And I think he was right.

[@neqkk]

I don't know if anyone mentioned it already: the established terminology for "texture pages" I believe is "texture atlas"

[@superplayer009]

As a Technical artist student who has built games in engine and built a 3d engine form scratch in C++ (using DX11) i must ask why not just make it easy for yourself and use an engine? I mean its cool if you want to get an idea how engines are made and what shaders are but its not a very effective way to make a game, picking a engine and learning its tools is far more efficient than building your own from scratch.

[@KillAllCommies2]

I heard that the Godot engine is astounding for 2D games.

[@tenv]

From the way you describe your code it sounds like procedural or data-oriented might be a better descriptor than functional, at least where I draw the line on whatever "functional" means.

[@steeznson]

I think using objects might help with the memory management since data associated with an actor can be stored in its class.

[@mirrorblade6268]

Behold, the future rulers of humanity on their infancy.

[@tunczyk110]

regarding that thing about using 4 vertices instead of 6 to draw a rectangle, you're looking for Element Buffer Objects

[@Coreagrus]

When you release a game, I will play it, judge it respectfully however it turns out, and overall - honor you. Best of luck, Matthewmatosis and co.

[@soyunbonus]

Making an engine is great for lerning, but not for making games for the most part. It takes a lot of time that you could be using in improving the game itself. But, to understand what other game engines do under the hood, it's a great learning process.

[@elizawada6583]

If you you're trying to make a makeshift level editor for your OpenGL build, rather than typing the level's information into a text file, which is quite abstract and subject to a lot of fine tuning, would it be possible to map everything to a bitmap file so that an actor/tile maps to a colour and the pixel position of the bitmap could correspond to the tile/actor's position on the grid of your level? This would have the benefit of at least giving a visual representation of how the level would be laid out come time to render, and make things maybe more expedient (hopefully) to throw together a level layout. But then again this feels like it'd have the pitfall of requiring a colour legend to navigate the different actors and tiles, and would get more comprehensive/complicated as you'd have to add more support for things like particles, all of which have their own individual behaviors that need to be defined, in which case you'd just end up requiring ANOTHER text file just to accommodate these more complex elements.

[@zomologorg]

I'm at minute 6 now, and my pattern recognition machine brain seems to have concluded that it's actually impossible for any of the fleas to reach the 7x7. They could easily make the jump from the edge of the platform before it, but due to the layout they'll never actually be able to jump from that point. What a scam.

[@RA-eg8tw]

Could you share the resources you used to learn? Are there any books that you found really helpful in your learning progress. I've heard many say 'Game Programming Patterns' being useful

[@blblblblblbl7505]

Sorry if it's too personal a question, but what's your background outside of the KZbin content? I'd love to know how you learned all this and got to this level of expertise. I ask as a self taught programmer, who would find this kind of project a pretty daunting task. Feel free to not answer. I love your videos, and can't wait to play the game.

[@ankurama42]

I often wonder if this is right path to get the most done with your limited time. Shouldn't one first gather enough capital to have people work for you or assist you in embarking on such an enormous project like yours? Wouldn't that be more effective with what you're trying to achieve? Tbh, all this seems like you want to make things more difficult for yourself rather than actually finish making a game.

[@saysolmao]

I need this game finished and on my desk by the end of the week

[@samueldeandrade8535]

Is he still uploading videos?

[@thesimpsoan7946]

Are you gonna recommend ghost trick again when the remaster comes out lol

[@okami7847]

Hey Matthew how long would it take for someone with zero experience to get to this point? I know the obvious answer is going to be like hundreds of hours but is it as difficult or daunting as it seems? EDIT: I'm actually referring to coding I suppose, not actually creating an Engine

[@matthewmatosisextra]

I'm probably the wrong person to ask as I can be quite stubborn and insular which slows me down some. I was the same way with the channel which I think ultimately paid off after some growing pains so I'm just hoping the same is true here. This project was easier than I thought and took about a month of work but obviously I have a lot of programming experience already. The collisions/animation/state machine stuff all took months initially but porting them to my own engine only took a couple of days each. I'm not sure what level of understanding you're at but with each passing year I'm more sure that playing the Zachtronics programming games had a positive impact on how I understand and write programs myself. They demystify assembly and assembly is what your processor actually runs, no matter what language you're using. Coming to terms with assembly is probably the single biggest leap I ever made as a programmer because it gave me a much better idea what things I should and shouldn't care about. If you can complete those Zachtronics games without any help you can probably do everything I've done in these videos, it just takes time.

[@okami7847]

@@matthewmatosisextra Okay I see, thanks, I appreciate the insight!

[@SpooderW]

Not sure if anyone else mentioned, but even with Unity/GameMaker/Unreal, you can NOT release on console without an SDK. There is platform specific code required by all platforms for supporting their particular certification requirements (achievements, dashboard interactions, etc), and of course you will run into performance issues and shader incompatibility with specific platforms that will need fixing. Even if you could just export and cross your fingers, that would be a crazy thing to release without testing lol. Even disregarding all that, you cannot install or use the modules that allow you to export to propreietary platforms in Unity/GM/etc unless you have a license from the platform holder, in which case you would also have access to the SDK. Not gonna weigh in on the technical aspects of whether or not its "worth" rolling your own engine, just saying that if one of your big concerns is "needing an SDK", you'll need one no matter what path you choose.

[@Kriss_ch.]

You've dreamt of getting an engine that feels as good as the Team Ladybug engine, right? I think it's worthwhile to explore what you can do to get to that point.

[@EdgardR.]

This is interesting

[@deleteTF]

Is this engine just for the new prototype you are doing, or if you come back to Mushido, will you use it there aswell and ditch GameMaker?

[@matthewmatosisextra]

Right now, neither. The prototype is still in GameMaker and this new engine is mostly a curiosity. It's not currently viable to swap over but it might be one day.

[@likessox92]

So when I play Matt's game in 5 years i twill run good? Nice

[@TOPGWAFDude01]

I'm a person online trying to find a channel... Who made Nintendo comment the same name as the oneni chose for my channel. I made it to find RandomGuyDude01, but i have no information about them anymore. Would you know them in any way?

[@Sharo93]

Even if you use Unity and Unreal for consoles it won't be easy as you will have to optimize. I will make the argument that if you have a nice abstraction on your renderer porting to a console is much much easier and it's code you know so you can optimize it. Getting the DevKits is the biggest problem, in that regard I have no experience. To add to this, the APIs on consoles are very similar to each other, if you are okay with putting up with/porting to vulkan you will have an easier time on consoles too as they are very similar.

[@GizouGitai]

Game Maker is the clear answer to me, simply because it makes the next 5-10 years of this game even viable. I don't want to see Matt's game never get made because he just kept optimising the game down to ASM instead of adding features, like some kind of reverse Yanderedev. At some point the game has to even start being functionally playable. Secondly, let's pretend that the game will be immensely successful. Matt will be able to subcontract the porting work out, even if GameMaker dies. And he can go on making his next blockbuster genre revival, and go off the deep end like he's the next Kojima. Thirdly, The RenderWare engine. Game Maker might not go bankrupt, but they can be bought and killed gently, like EA did with RenderWare. Persona 3, 4 use RenderWare, and they had to be ported. Persona 4 Golden and P3Portable apparently didn't use RenderWare (that's the rumour), which resulted in some broad changes, and the PSP/V version were the only viable ports in this dystopic EA-ruled future. But, they WERE ported. And ported just fine, at that. Kind of. The Grand Theft Auto PS2 trilogy were also RenderWare.

[@matthewmatosisextra]

Yeah I think some devs have even used an existing engine to release and then updated it to their own engine once it was successful and they had more time on their hands.

[@Thisisthegreatestatofalltime]

We all know what happened with the port of that trilogy. Though tbh that was probably more a combination of incompetence and a bs deadline.

[@gooberusmaximus5362]

Mega micro engine

[@evillogberry]

Any new updates on the game?

[@FeatherDory]

Not to be a backseat driver here, but if you're not using OOP in C++, you're going to have a hard time cleaning up garbage and you'll likely face issues with memory leaks. Your code won't be DRY and you're going to have a helluva time organizing, updating, etc.

[@beanpole4701]

the only problem with using opengl directly is if you ever want to port it to consoles or macos youll have to do a LOT more work than if you just abstracted it out a layer. i appreciate doing everything yourself, but its kind of bad for future proofing

[@beanpole4701]

i didnt finish the video lol

[@m8subject996]

33:23 uh oh

[@wyrdplae8586]

Matthewmatosis has a whole second channel!!!!!!!!!!!!!!!!!!!!!!!!! OMG

[@CornishCreamtea07]

I had assumed it was your own engine giving how much low level programming you seemed to be doing.

[@UneededStudios]

It's not precisely true that everything in a 2d game using OpenGL would be represented by a rectangle. You may reach a point where you need to optimize against wasted transparent pixels and have to tailor the underlying meshes of your sprites as a result. Very unlikely someone would reach that point with a shader heavy game though.

[@matthewmatosisextra]

Interesting, I hadn't thought about cutting out that empty space. I suppose it's more important the larger the sprite is. Probably not worth worrying about for the kinds of games I want to make right now but a good thing to be aware of nonetheless, thanks.

[@supersonic159]

You're not a fraud game dev Matt. You may not have shipped a product yet, but that doesn't mean you can't be the real deal before then. What makes a dev isn't shipping a game but making one.

[@megamozgs9959]

Imho, you should avoid using geometry shaders. First of all, they are really slow, but more importantly, they are not supported on many devices, apple gpus are an example of that.

[@vj7248]

hopefully matthewmatosis did NOT switch to Unity for whatever reason

[@part_time_commie]

I am pretty impressed by your journey to be frank. Your thinking should give a good foundation for a steady pace forward. I have seen people in the industry not go that far. As for which tools to use, it's all up to you. If you want to have fun and have maximum flexibility use your own engine. If you want a faster turn around and compatibility use Game Maker. Personally I would choose Game Maker even as someone who worked on custom engines. Use the right tool for the job and here it seems to fit the needs of your game very well. As I said before, if you want some mentoring, help or just throw ideas at someone you can message me. I really like the way you approach problems, don't give up and you'll most certainly get somewhere if not far.