next »

[sublinimal]

Spaces, space. Overlapping space. More space for me and my thoughts. They don't fit in this narrowly defined, narrowly defined space. Exclusive contingent narrowly defined, space. No space.

[sublinimal]

panic

[sublinimal]

Dawn of The Final Day -24 Hours Remain-

Well, I haven't posted all that much. It's difficult for me to confirm details of the game, and some of the stuff I've posted about is already obsolete. My way of making games is by no means a linear process where I'd just stack new things on top of each other. I think about mechanics far more than content, but seems like most gamedevs do it the other way round, and maybe they're more productive because of that. I'm critical of my own work, and it takes a lot of wrong turns before I come up with something I'm happy with.

I figured if the game design principles I've worked to incorporate won't be obvious in the end result, spelling out the things in some puzzle that are supposed to "prioritize algorithmic thinking over specialized solutions" would only sound like a pretentious excuse. I want to become someone who shows rather than tells.

Anyway, the game is pretty close to completion. Several puzzles are playable already, a few are pretty clever - the "final boss" involves implementing a kind of sorting algorithm using conveyor belts, different sensors, teleports, and pretty much all the hacks you've been taught before. However, there's still incomplete features, and ones I'm not happy with. There's not enough in the way of polish either, and I don't think I've managed to illustrate the more unintuitive mechanics well enough.

So I'm seriously considering whether it would be worth it to take a few more days to make more levels, do further playtesting, fix bugs, and add more polish & feedback. The event itself is pretty loose about deadlines. Which is how it should be, since #onegameamonth ain't a jam, it's just a resolution to finish some damn games this year.

Basically, I can't give up on this now. If I were to release it as an unfinished package, that would feel more like giving up than pushing the deadline would. I'm starting to feel overworked, which is likely to lead to a result nobody's happy with.

[sublinimal]

Okay, here's the plan. I had released a new version of Hunter to Hunted in early January, which would qualify as January's One Game a Month entry without me having to swallow my pride. This means that Control Room was February's entry all along! You were all fooled! Including myself!

I'll wrap this game up in the first week, then do some serious work on that other game for the rest of February, and likely get started on March's entry (which I don't expect to finish in one month anyway).

The evolution simulator I had planned for February will be postponed, but it's not the sort of game I would've wanted to work on right now. I hope this sort of scheduling will help with quality control and keep me sane.

What do you guys think? You've been awfully silent recently.

[koiwai]

Oh, this is an interesting project, I am really intrigued to see how it will shape up! I think, you have a nice approach to define functions, and going up in the call stack in case of failures.

Actually, why I am writing, I am planning to make a devlog of my game that involves similar robot programming concepts. Though my language is based on a little bit different computational model, strangely, visually our languages looks quite similar. Convergent evolution works..

Maybe, we can discuss or share some ideas. Probably, I should speed up and make my own devlog sooner. I hope, you don't think that I am stealing your ideas.. This is not the case, and moreover, my game is a roguelike with robots, it is very different. Btw, I had one screenshot posted in the screenshots thread.

Wish you good luck with the game. Very interesting work. Sorry that the  development is going slower than planned, but that's how it goes sometimes.

[sublinimal]

Actually, why I am writing, I am planning to make a devlog of my game that involves similar robot programming concepts. Though my language is based on a little bit different computational model, strangely, visually our languages looks quite similar. Convergent evolution works..

Ah, cool. I'd like to hear more about it, and exchanging ideas sounds good too. I'm actually hoping more people would make games inspired by programming concepts, because there's so much uncharted territory there.

[sublinimal]

I have somewhat of a "writer's block" with this project. I think this is partly because some of the rules are clumsy, so the gameplay isn't as fun as I'd hoped, and it's not convenient to make puzzles that make you think in states.

Firstly, even simple levels have pretty convoluted programs. That's not necessarily a bad thing in itself, but movement code is tedious. It feels like handling C char arrays when you really just want Python strings.

So I've been gradually taking the in-game language to a different direction, resembling finite state machines. Commands can place the bot in multiple states that have constant consequences, and each of them influences what states can be activated or deactivated. For a trivial example, you can't drop items when you're not in the carry state, and you can't carry when you're already in the carry state.

Ultimately, it has to do with spaces. Functions are spaces that have their interrupt values and return values. The language can be seen as a space containing a variety of states. In a larger sense, even a puzzle is a space, where you're supposed to get into a state where it's considered solved. It's something that I feel like I've been trying to grasp, and makes me think why didn't I approach the in-game programming like that in the first place.

In the meantime, I've already started building the foundations of March's project, because I just want to get some work done. I've reached the point where I can't view this project objectively anymore; there are times when it stops being a fun project and instead just some task I want to "get over with". Luckily I'm too much of a perfectionist to release it before I'm completely satisfied with it.

It's time to look at Control Room from a different perspective, to re-evaluate what's supposed to make it fun.

[koiwai]

You can try to turn this game into programmer's Super Meat Boy. It's a different perspective, but you already have a functioning game, it is just not fun enough.

Try to make the robot repeatedly evaluate your comamnds, even if they are incomplete. Make him move very quickly. When the program crashes or ends, it start over from the beginning. The robot is doing this, until the program is correct. Also, make a button to restart evaluation in case of long programs.

I mean, frustratingly hard games may look fun ..

Easy tutorial levels can help a lot too. don't give up, man, don't give up!

Wondering, is it possible to let the robot do something on his own? Say, if you control Bender, he would not just follow your commands, but rather often deviate and do random stuff on the way..

[sublinimal]

After a creative pause, I think I'll actually try to get this damn thing done now, by looking at it from the "finite state machine" perspective.

Besides that other game, I've actually been working more on March's project in February (which correlates with the fact that it's frankly more fun to work on). I'll make a devlog for that soon.

[sublinimal]

Hm, I wonder. What if this game was in a top-down, "turtle graphics" perspective after all? It'd eliminate some of the more tedious parts of movement, like falling and jumping, which are hard to represent as states. Also, it'd give you more options for using sensors.

[sublinimal]

Perhaps the main "problem" making this tedious is that I'm trying to base the puzzles around one actor (the robot). It encourages specific solutions like most puzzle games do, only with the difference that you don't have real-time control. So I'm playing around with ideas like crates that live their own life. They could be programmed with simple rulesets that are supposed to create order from a chaotic startpoint after enough time. So you'd handle sets of objects rather than just one.

Whatever. My brain just locks up when I think of this nowadays. And if I ever finish it, it still won't be as good as Manufactoria.

[koiwai]

In top-down view you can try using cellular automata for crates. If you have time for that.

[sublinimal]

That's it you have 24 HOURS

TO BECOME A FUN GAME

OR I WILL USE THE RIP ICON

(AND I'M NOT AFRAID TO USE IT)

AND GO WORK ON MY FUN PROJECTS INSTEAD.

[sublinimal]

(But if I do get it right, I'll just disregard the whole One Game a Month thing and finish this anyway)

(And if I don't, I will eventually resurrect it because I can't stand losing a battle of wits to a friggin Flash game)

[sublinimal]

Alright. I guess I'll have to admit failure for now, since this is taking time and headspace from other projects. What I can do, though, is try to take it constructively and sort out my thoughts in a post-mortem.

What went right?

• It had potential. A game about algorithmic solutions is something I've wanted to make for a while.
• A working development process. I worked pretty hard, it seemed to be on schedule for One Game a Month, and even after that, I didn't immediately call it quits. It wasn't because of distractions or procrastination that I failed, and I did learn a thing or two about parsers and game design. Pages of pen-and-paper design too that forced me to think about the fundamentals.
• It had some neat production values. I like to be a jack of all trades when it comes to game development, and I pulled it off adequately again. The graphics were good enough, it would've had some catchy music, and the neurotic humor gave me a few giggles.

What went wrong?

• It just wasn't fun! The in-game programming wasn't abstract enough not to be tedious. And save for a few that kept me going, the puzzles felt either forced or incomplete. Towards the end of the project, I just couldn't grasp the soul of the game anymore, since the mechanics were as if pieced together from different games because of all the redesigns and duct tape. There was a point where I had to go "this doesn't work". Every time I think about it now, my mind just starts to wander as if oversaturated.

Playing the aforementioned Manufactoria was like the final blow - it puts me to shame with the way it manages to do something really clever with simple rules. If I ever revisit the concept of Control Room again, it's going to be something abstract like that, a completely different game with the same name and some recycled assets.

Until then, it's time for me to move on.