Notices by Jacek Wesołowski (jzillw@mastodon.gamedev.place)

My prototype uses tilemap levels. Each tile has an entry for each of the four walls. In the current implementation, each entry is a struct. When you build a wall and then delete it, the struct changes wall type to none, but doesn't clear other params, such as the tint.

The bug was that east walls were being tinted with the colour of the north wall from the same tile. Even if the wall wasn't there, there was leftover colour from before it was deleted. Result: walls in seemingly random colours.

tired: red, green, blue, alpha
wired: lavender, geranium, basil, transparency

A custom tool like this can speed up a designer's work by an order of magnitude. Most teams I've worked with were reluctant to invest in tools, but I'm giving them priority, because I'm broke. Also wanted to learn new Unreal tricks. Can't wait to build an actual level now (hopefully soon).

This project has gone from zero to this in about 3 months, but a lot of that was me learning all kinds of Unreal weirdness. If I were to reapply knowledge to a new project, it would probably take one month.

I just added a debug message "check filenames on disk" and realised my PC doesn't even have a "disk", because SDDs have replaced them.

It sometimes amazes me how transient modern tech is. When I was a kid, we still used tape. Uncle had a floppy disk drive and it seemed so high-tech.

Big success today. My pet project uses an old-school presentation style that emulates 3D graphics with sprites. Problem is, the further away you let the player see, the more sprites you need. Old dungeon crawlers usually had a view depths of only 3 tiles.

In fact, old dungeon crawlers faked perspective and scaled their sprites uniformly! I took inspiration from this trick. This view uses correct perspective for depth <= 7 tiles, and then switches to uniform scaling. Total view depth: 12 tiles.

We live in the stupid timeline, part over 9000: some random store that seems legit but I don't even recall buying from them, sent me best wishes (and a sales offer) on the occasion of my name-day.

My name-day is in February.

Solo dev means doing a lot of work outside of your area of expertise. In this case I've spent last three days improving my texture packing algorithm. This is a heuristic that takes the nature of the set into account, complexity is technically O(N^2) but close to O(N) in practice.

There are two edge cases I'm not handling, but I'm only going to add that if it becomes relevant in practice.

That feeling when reading a bunch of wrong advice on Unreal forums helped you realise that some other advice you took from there was also wrong, while repairing the damage from said advice also fixes the issue you were trying to solve in the first place as a side effect.

*sighs in C++*

Today's success: managed to display the level (which is a custom asset in my case) in a custom editor's 3D viewport. Made easier by 3D viewports in Unreal having a UWorld of their own, and me having a separate Actor class whose sole responsibility is to display the level.

The main advantage of Unity over Unreal is that Unreal essentially forces you to read its engine code and in doing so exposes all the parts where it's held together by spit and duct tape, whereas with Unity you live in a state of glorious ignorance.

Unreal prototype is now officially more advanced than the old Unity version. I've added a texture importer that speeds up asset creation. Materials support a detail layer and tinting. Next thing to do is cutouts so that I can have doors and windows without adding variants for each texture.

The wall material doesn't tile correctly, but that's an issue with the data and not the code.

The range of view is limited and it shows, but that was the case in old crawlers as well and it can be mitigated.

Weirdly enough, Gamescom is giving me FOMO this year. I never had this with E3, PAX or GDC. Could be just a simple case of Gamescom being easily reachable, either by plane or by train, i.e. it feels like a place I could actually be at.

Plus I feel like I don't know enough people in EU gamedev.

Finally managed to switch the personal project from manual to automated texture management. The success here is that everything looks the same as before, hence no screenshot.

Took me two and a half weeks instead of a week, but I feel excused: I was taking it easy after stressful period at work.

@thejessiekirk @stilgherrian I think 1989 CEE revolutions are more relevant to today, and they all featured:

- pop-culture, including what we would call memes today, as a recruitment tool

- protests that de-prioritised (but crucially did not rule out) violence, such as general strikes

- building their own propaganda with movies, songs, "second circulation" etc.

- negotiated compromises that involved going to the polls

The Romanians did execute their dictator, but that was an outlier.

Stuff like this isn't attractive, it doesn't scream creativity, but it's actually crucial if you want the project to get someplace nice. Being able to generate this atlas automatically will shave a few hours of work off every single surface material in the game.

This atlas was generated automatically by a naive algorithm that leverages a heuristic about the texture set that's being processed. Efficient development means I'm not going to complicate the algorithm for the sake of optimisation, unless I actually need a more efficient solution.

@goshki Mmmm, old-school joystick controls... exquisite. :-)

The goal of this pet project is to replicate the look of dungeon crawlers ca. 1992, i.e. make it sprite-based. Most modern dungeon crawlers use true 3D environments with heavy filters on top instead, but that's just not the aesthetic I'm aiming for.

To be clear, I'm under no illusion that any publisher ever is going to pick it up. I wanted to do something just for myself, have a testbed for gameplay experiments, and learn some more Unreal in the process.

That feeling when any layperson from outside the project would tell you to gtfo because this doesn't even look like a game, let alone play like one, but you're just so happy to have crossed a milestone (i.e. levels in my dungeon crawler can now be multi-storey).

Decent wall materials are next. ^_^

(I started this project in Unity a few years ago, but then Unity went nuts and then I was busy; now I'm recreating it in Unreal)

@cstross On some level I appreciate the craft that went into implying this is about football in such a way that if you assume it's actually about money the exact same wording also applies.

(i.e. "having listened to all the views it has become clear that even if we manage to push this through you won't let us make any money")