4 ms·

> While the game is my goal, i focus on the platform because i feel solo devs have a ton of overhead. I figure if i can optimize my the creative aspects of my w

> While the game is my goal, i focus on the platform because i feel solo devs have a ton of overhead. I figure if i can optimize my the creative aspects of my work i will have better long term success. I've been researching solo game dev for decades now, and that definitely mirrors my conclusions so far. Just keep the end goal in mind and don't reinvent the wheel, unless you desire a career in wheel engineering. While you're working on your stuff, remember that others are also working on their stuff, and a lot of useful things are shared freely and continually popping up new. Have the courage to recycle your ideas, and the ruthlessness to streamline your development process. Going solo is all about coming to terms with your limitations, and building solutions to work around them. Maximize the impact of work you like doing, while minimizing the necessity of work you don't. Sacrifice your creative vision until only the core remains. Resist the creeping scope of scaling it up, but also keep firm in your artistic requirements. Keep your "must haves" short, solve the hard problems first, manage your expectations, and above all be stubborn enough to tread through it.

boredtofears 4y ago
This is pretty much exactly why I (as a complete hobbyist) ended up with Godot. I spent two years writing things from scratch in various languages and lots of time Yak-shaving. The experience itself was worth something, but I don't really have a lot to show for it. I don't love godot - but it's far beyond anything I'll be able to accomplish on my own as a hobbyist. I find art to be the most limiting factor at this point. I'm not a great artist (what I can do takes a painstaking long time to finish) and it's hard to satisfy a creative vision with free art assets.
- verteu 4y ago
  Have you tried Upwork/Fiverr/Behance/Dribbble? In my experience you can get decent art for quite cheap.
  - thn-gap 4y ago
    Depending on what kind of art you need, Midjourney or similar models are becoming a more visible option for solo gamedevs to make decent art cheaply.
    - meheleventyone 4y ago
      The problem with ML art outside of concepting (and even then) is repeatability of scene elements. If you’re illustrating a children’s book you usually want each illustration to keep the characters and scenes consistent. The same is mostly true for game art and concepts.
      - steezeburger 4y ago
        It's been really cool to use for inspiration as well as generating seamless textures. I've generated some great grass and carpets for my ps1 style game.
        meheleventyone 4y ago
        I can definitely see it being great for that style of texture. Basically anything one-shot. How much post fiddling did making a seamless texture take?
        steezeburger 4y ago
        Sometimes almost none! I usually generate a normal map for it, but the last one I did was retro 70s carpet, and it looked awesome with zero cropping/editing. It was almost perfectly aligned. It helps that my game is quite lofi haha. I'm using Godot/Qodot/Trenchbroom.
        meheleventyone 4y ago
        Amazing. I think ML will also be pretty handy for generating incidental images. Like you want a picture of a log cabin on a spooky lake in your game and so on.
- lijogdfljk 4y ago
  For context, i don't suspect Godot has too much of what i'm focusing on built in. Versioning of my assets, build pipelines for asset source to optimize it or generate variations of it (2D images, renders, etc), generate configuration of available assets for use in game, etc. Though i am purposefully using Rust, so even if Godot did it's a bit more awkward. There are Rust libs for Godot, though i'm not entirely sold on the libs quite yet. My basic thought process is never do anything more than once for the creative work. If i can't find a way to keep overhead obscenely low then long term i will drown in one-offs. I think the same way with game code. A heavy emphasis on testing, integration testing, etc. I don't have the time to Q&A for all possible regressions, so i have to ensure i do everything i can to make the engine very testable. Minimizing repeat work is my goal with everything. Also insanely stable builds, i suppose. Ie i don't want a change in an asset to break historic builds or source. My source is tied to the assets at that time. etcetc edit: And to be clear, i'm not writing a game engine myself. I'm using Bevy atm.

boredtofears 4y ago

This is pretty much exactly why I (as a complete hobbyist) ended up with Godot. I spent two years writing things from scratch in various languages and lots of time Yak-shaving. The experience itself was worth something, but I don't really have a lot to show for it. I don't love godot - but it's far beyond anything I'll be able to accomplish on my own as a hobbyist. I find art to be the most limiting factor at this point. I'm not a great artist (what I can do takes a painstaking long time to finish) and it's hard to satisfy a creative vision with free art assets.

verteu 4y ago
Have you tried Upwork/Fiverr/Behance/Dribbble? In my experience you can get decent art for quite cheap.
- thn-gap 4y ago
  Depending on what kind of art you need, Midjourney or similar models are becoming a more visible option for solo gamedevs to make decent art cheaply.
  - meheleventyone 4y ago
    The problem with ML art outside of concepting (and even then) is repeatability of scene elements. If you’re illustrating a children’s book you usually want each illustration to keep the characters and scenes consistent. The same is mostly true for game art and concepts.
    - steezeburger 4y ago
      It's been really cool to use for inspiration as well as generating seamless textures. I've generated some great grass and carpets for my ps1 style game.
      - meheleventyone 4y ago
        I can definitely see it being great for that style of texture. Basically anything one-shot. How much post fiddling did making a seamless texture take?
        steezeburger 4y ago
        Sometimes almost none! I usually generate a normal map for it, but the last one I did was retro 70s carpet, and it looked awesome with zero cropping/editing. It was almost perfectly aligned. It helps that my game is quite lofi haha. I'm using Godot/Qodot/Trenchbroom.
        meheleventyone 4y ago
        Amazing. I think ML will also be pretty handy for generating incidental images. Like you want a picture of a log cabin on a spooky lake in your game and so on.
lijogdfljk 4y ago
For context, i don't suspect Godot has too much of what i'm focusing on built in. Versioning of my assets, build pipelines for asset source to optimize it or generate variations of it (2D images, renders, etc), generate configuration of available assets for use in game, etc. Though i am purposefully using Rust, so even if Godot did it's a bit more awkward. There are Rust libs for Godot, though i'm not entirely sold on the libs quite yet. My basic thought process is never do anything more than once for the creative work. If i can't find a way to keep overhead obscenely low then long term i will drown in one-offs. I think the same way with game code. A heavy emphasis on testing, integration testing, etc. I don't have the time to Q&A for all possible regressions, so i have to ensure i do everything i can to make the engine very testable. Minimizing repeat work is my goal with everything. Also insanely stable builds, i suppose. Ie i don't want a change in an asset to break historic builds or source. My source is tied to the assets at that time. etcetc edit: And to be clear, i'm not writing a game engine myself. I'm using Bevy atm.