It's been longer than a week. But that's okay—things were slow anyways.

Perhaps the biggest change—at least the biggest noticeable change—is that Jeremy made the water look wetter. See for yourself.

It's all still a work in progress, but it's much better, right?

Other than that, I created several animations for cannonfire:

and for cannonballs hitting things:

We've also spent some time generating and refining game mechanics, I've started prototyping ships sinking, and we've gotten in front of our to-do list, organizing it quite a bit using Github Projects. I've even started an intranet site for us to start documenting things.

Oh! I almost forgot. We started narrowing down a music/sfx style. Still some work do to on this, though (like everything else).

That's it!

Over the past couple weeks, I finished several ship animations. There's still a lot to do, but this goes a long way to making them feel like they're actually in the world.

That's my big completion for the week.

I also did the following:

• added to our game-files a couple tower assets
• at Jeremy's suggestion, darkened the waters a shade
• added ship shadows; they are partially transparent, so they will darken and lighten based on what shade of water you're on

One of the things we're noticing is how dead everything feels when you're zoomed out. All of the embellishments are really for being zoomed in, but the action is easier to follow when you're zoomed out. I'm hoping adding a bunch of eye interests, like water glints and rocks and things, will help.

Next comes a bunch more animations, like ports being built, ships sinking, and cannonballs hitting things.

I reworked the ports, and it wasn't too bad. They're working great now.

And the paper texture is better, but it's not quite there yet. The issue is that the map grid is too uniform. We're looking into adding some overlays that would break up the uniformity, like a fancy compass, little dingbats and asymmetric, map-spanning lines. I haven't started on that yet, but I'm looking forward to it.

I also did a little more prototyping of the night variant of the map. It's an idea we like a lot but haven't worked on much, given the complexity. But it has a lot of promise.

And finally, I've started animating the ships. This might change as I gain experience (and speed) and start seeing the results, but it's been one of the more tedious aspects of asset creation. So far I've created wakes and sail/flag movement for a few directions of the main ship. There's a lot more work to do.

Jeremy has always suggested that ports have walls around them. It never appealed to me, though. Maybe it just seems unlikely Caribbean ports were surrounded by walls. Or perhaps I had big medieval stone walls or mind. Or maybe I just failed to conceptualize it right.

But as a result of his gentle prodding, and since I wasn't happy with the two variants I had already created, I did to some walled-port testing.

So, we're moving to walled ports.

And guess what. This led to more learning. The repeated textures I had started making for our tiles didn't take into account that our tiles overlap. That is, the tile images are 48x48, but Godot fits them into 32x32 tiles. (The overlapping edges are how diagonals are possible). So since I was already making tile images at 48, I made the repeating pattern on them 48x48 as well. The result is that when the engine puts them together, their seams are all wonky. What I should have been doing was making them 32x32 and then just expanding them to fit a 48x48 image. That's what I'm working on now.

This is a 48x48 texture in a 48x48 square. If you place this in a 32x32 grid, its edges overlap other tiles and mess up the seams.

This is a 32x32 texture with 8 pixels all around it. In those 8 pixels, I've tiled the opposite edge of the 32x32 texture. If you place this into a 32x32 grid, the edges overlap, but because they're the same image where they do, you can't tell.

So not only do I have to rework the ports (again), I also have to rework the paper texture I just did. One component of the gameplay will include a paper variant of the actual game map. We're pretty excited about it, actually. But, I won't say too much about that yet.

I'll add some images to this post later.

Arrrchipelago​

Here's a couple clips of the current state of the game.

The main elements to note are

1. the diagonal tilesets--you're seeing five tilesets of 256 tiles apiece (and change) all being dynamically generated and interacting with each other
2. port creation and the pathfinding the merchant ships do between ports
3. the newly-animated shoreline (which is basically another three tilesets...)

The ship and other items are in a rough-draft state. Everything's rough-draft until it's not.

And you'll notice a lag bump whenever I place the ports. Jeremy was able to resolve that earlier today.

More as we develop.

Arrrchipelago​

The trees disappointed us--think sand-boxes lined with monkey grass. But we may have found another way.

Some context:

1. We're using the Godot game engine.
2. It can dynamically generate maps using tile templates that we create.
3. Each tile has to connect to other appropriate tiles in a way that visually gels and that Godot can understand it.

We were using a 48-tile set. It's got some nice curves, and it provides a lot of the major shapes you'd want. But it only allows tiles to connect to each other side-to-side and top-to-bottom--no diagonals. Diagonal routes are zig-zagged, and zooming out reveals a map of squares.

We were also using 128px x 128px for the texture size. You get more control over micro-details but less control over macro-ones. That is, you can make lots of little things on a single tile, but the tile, itself, is 3x bigger than the player's ship. Straight edges go on and on between abrupt corners.

So, I set out to experiment with a tileset designed for diagonals and with smaller textures. We liked the results visually. But not understanding the engine yet, I designed something that Godot couldn't use.

Luckily, Jeremy found a tool that takes a 25-tile template and creates from it a 256-tile set that allows for diagonals. I spent a couple days working out the details, and it's working. Not only has it cut down on the iteration time--I only have to design 25 tiles and at 32px x 32px--we have our diagonals!

Now we just have to make sure it works with everything else. It's a step back, and it'll take some tinkering, but as Jeremy told me, "When it's right, it's right."

I will add pictures and links shortly.

Arrrchipelago​

The topic of the weekend was palm trees. And guess what? They reminded me that I am still learning. How hard can it be to draw a bunch of palm trees, right?

I made a prototype a couple weeks ago and was happy with it. It started with a single tree, which I just copied over and over.

That was okay, so I made some variations--

--and then copied and pasted those randomly. I was less pleased with this one.

I dabbled with drawing shadows onto a base, lighter color green so that the base color had its own branches poking out into the shadows. From there, I just highlighted some fronds and ended up with this.

We were pretty happy with that, so we moved on. Then, after figuring out that I needed a tree solution before a port solution (I'm also designing ports), I came back and attempted a cleaner version.

The weekend just wasn't having it. I spent a couple days attempting to develop a reproducable process, and I wasn't able to even reproduce the original effect. One workaround I floated was to do individual trees again.

But after realizing I was just digging a big, angry hole, I peeled myself away and waited until the next day.

This was the product after that night's sleep. And though I suspect I'll be able to find a lot wrong with it after a couple years' experience, I'm pretty happy with it.

And let me say that Adam C. Younis is super helpful. Here's his vid on environment: https://www.youtube.com/watch?v=yBmaz45eHFA&pp=ygUZYWRhbSBjIHlvdW5pcyBlbnZpcm9ubWVudA%3D%3D

Other Thoughts​

I should start recording every time I remember that I am still learning. Other than the various hobbies I've had, I do not have a background in visual arts, and my last training was before high school. But I get so carried away with wanting to create that as soon as I produce anything keep-able, I forget about education and, more importantly, practice (and failure).

I need to develop a visual-arts education-regimen, and I need to remember its importance. This includes The Arts, pixel art, aesthetics--things I'm already interested in and am always working near but have never formally studied.

Anyone want to buy me an MFA? Or maybe a practice-based theological aesthetics PhD?

poindextergameco.com:​

• Created a basic presskit page. (Not that we have any press).

Arrrchipelago:​

• Started working on an Arrrchipelago logo. It's not done yet, and we're toying with changing the name, so I'll post it later. But we're liking it.
• Finished first draft of the sand -> water and water -> deep-water tilesets. Also developed a bunch of buildings and things out of which we'll create the ports.
• Started thinking about how we're going to animate these. Not much work on that yet, though.

Other Thoughts​

It's important to note that we're using AI as much as possible. Some of it is learning to future-proof ourselves, and some of it is to bolster our limited manpower.

For example, my normal process for any new visual object is to assault GPT-4 for a bunch of iterations of that object. Then I just pick and choose the things I like or the things that come to mind in response to GPT-4's results and mash them all together in my own style. I jump between Aseprite and Illustrator a lot. Also a bit of Photoshop.

I suspect that in a few years, at a minimum, creators will mainly just have to have taste. AI is to cognitive faculty what electricity is to manual labor. Like going to the gym, thinking will be the exercise of the healthy.

I'll just summarize some of the things I've been working on for the past couple weeks. Hopefully I can make this a more regular thing and add some screenshots and whatnot to document a bit better.

poindextergameco.com:​

• I've deployed a basic site.
• I've also configured initial email addresses on the new domain and the set up social accounts at the various providers. Instagram didn't like the profile at first and disabled it, but I got through that.

Arrrchipelago:​

• I realized I need to still think of this as learning. I had already fallen into performance mode, and the tiling that I was working on was starting to disappoint me. I'm toying with 128px x 128px for the tileset size now--up from 16px x 16px--and will be seeing some test results soon. One issue I ran into was not understanding how the tiles in Godot connect to each other. I realized after creating 100-some-odd tiles that I wasn't going to get the more-natural results I was looking for and moved to the bigger size.
• I need to put into my schedule different learning avenues. Again, this is new. I tend to just want to make stuff with learning just helping me when I get stuck. Being more strategic with such would certainly help.

Poindexter Game Co.​

• Created a logo, and we agreed on a company name. Creating our identity now will help with things later on.

More later.