12-26-2023, 06:46 PM
Leo: Well, on the code, it depends. I wrote the code that handles the resources moving in this game once. Then I rewrote it a second time. Then I rewrote it again a third time.
The third rewrite is the way I should have written it the first time, and if I had, then it would have saved me time. I couldn't leave it the way it was because it was needlessly complex and confusing. As a result, the game had bugs that I couldn't find in the code, because the code was too fucked up to follow. I felt a real sense of dread having to go back to that part of the project. I didn't write it the best/clearest/least confusing way I could because I wanted to go fast.
And to be sure, it's not necessarily wrong with some code to go fast. Not everything involves actual data structures and algorithms types of issues. Sometimes there are five ways to change a variable on user input and the method you choose doesn't matter. But when you run into a higher level of problem, it's absolutely better to plan things out first, as long as it takes. That's when 'slow is smooth, smooth is fast' kicks in.
I suspect you haven't yet had your ass kicked by programming in this particular way, but you could be a much better programmer than me. I'm not sure how it is for other people. When I was in the coding bootcamp, there were people who seemed to fly through the code, and people crawling through it. I felt like the only person somewhere in the middle.
The enemy, yeah, he's not meant to look too imposing. He's the weakest enemy and, at least initially, will die in a couple shots. I probably will buff the amount of veins/tendrils on him, a bit. I'm worried about creating too much detail/polys for Godot, but we'll see.
I like the paintover. I think your idea would be an improvement on the image, taking advantage of what turned out to be a little bit awkward of a pose and expression.
That particular image has a funny story to it now. The initial reference image is a girl in her room with a totally different expression and hair over her face. The only thing I definitely wanted to get from it was the color vibe and pose vibe.
![[Image: 7pvkd8.jpg]](https://files.catbox.moe/7pvkd8.jpg)
I started drawing/painting it, then ran my partially finished image through AI for a paintover. There's a pretty great classical painting lora out there. This is what it spit back out at me.
![[Image: l3cxkk.png]](https://files.catbox.moe/l3cxkk.png)
I showed this to Muzz and he pointed out that the AI created a better flow with the pose and eyes than both my painting and the reference. In the AI image, the torso, neck, and head all bend at an increasing angle.
![[Image: ueo74d.png]](https://files.catbox.moe/ueo74d.png)
The eye pose is also better in the AI image than my painting or the ref. Out of the three, it feels the most natural. The eyes lead the head, which leads the neck, which leads the torso. I was too focused on fiddling with the texture brush to notice this. But this is the kind of stuff I'm talking about when saying we ought to be using AI as a learning tool. You can bet I'm going to be thinking more about this kind of flow in the positioning of the torso, neck, and head from now on.
The third rewrite is the way I should have written it the first time, and if I had, then it would have saved me time. I couldn't leave it the way it was because it was needlessly complex and confusing. As a result, the game had bugs that I couldn't find in the code, because the code was too fucked up to follow. I felt a real sense of dread having to go back to that part of the project. I didn't write it the best/clearest/least confusing way I could because I wanted to go fast.
And to be sure, it's not necessarily wrong with some code to go fast. Not everything involves actual data structures and algorithms types of issues. Sometimes there are five ways to change a variable on user input and the method you choose doesn't matter. But when you run into a higher level of problem, it's absolutely better to plan things out first, as long as it takes. That's when 'slow is smooth, smooth is fast' kicks in.
I suspect you haven't yet had your ass kicked by programming in this particular way, but you could be a much better programmer than me. I'm not sure how it is for other people. When I was in the coding bootcamp, there were people who seemed to fly through the code, and people crawling through it. I felt like the only person somewhere in the middle.
The enemy, yeah, he's not meant to look too imposing. He's the weakest enemy and, at least initially, will die in a couple shots. I probably will buff the amount of veins/tendrils on him, a bit. I'm worried about creating too much detail/polys for Godot, but we'll see.
I like the paintover. I think your idea would be an improvement on the image, taking advantage of what turned out to be a little bit awkward of a pose and expression.
That particular image has a funny story to it now. The initial reference image is a girl in her room with a totally different expression and hair over her face. The only thing I definitely wanted to get from it was the color vibe and pose vibe.
I started drawing/painting it, then ran my partially finished image through AI for a paintover. There's a pretty great classical painting lora out there. This is what it spit back out at me.
I showed this to Muzz and he pointed out that the AI created a better flow with the pose and eyes than both my painting and the reference. In the AI image, the torso, neck, and head all bend at an increasing angle.
The eye pose is also better in the AI image than my painting or the ref. Out of the three, it feels the most natural. The eyes lead the head, which leads the neck, which leads the torso. I was too focused on fiddling with the texture brush to notice this. But this is the kind of stuff I'm talking about when saying we ought to be using AI as a learning tool. You can bet I'm going to be thinking more about this kind of flow in the positioning of the torso, neck, and head from now on.