Have you tried Sil? http://www.amirrorclear.net/flowers/game/sil/ The manual is a bit of a requisite to understand the systems, but it plays a lot more like a table top RPG than most roguelikes. I'm digging the setting and focused gameplay.
So, there is a bit of drama in the RL community about a failed kickstarter. The sole dev originally asked for $5k but ended up getting over $30k. He showed detailed land generation and some basic history lore gen, but was touting this as a DF successor. Anyways, turns out he got bored of the project less than a six months into the game and gave up. He did, though, finally open source his code, which is a complete mindfuck.
The guys who safeguard the DF backups have confirmed that the code is essentially an impenetrable black box, which makes sense, as Toady is self-taught, and the project is the result of a straight-up decade of nonstop coding.
To a degree it's also how rogue-likes are made. Roguelikes are pretty much the most outstanding example of a single-coder project you can imagine. Everything about them encourages sloppy code and cut corners. Halting, pseduo-turn-based structure offers little incentive to optimize from a gameplay or computing standpoint. It all happens behind the scenes, so as long as numbers and a mad-libbed description paragraph comes out the other end, you can do whatever you want. Lack of representation means a sense that there are no restrictions; the programmer never has to work with anyone else and never has to worry about something looking right in action.
In Bloodcrusher II, which does a lot of rogue-like generation, my trained attack coder and I have to very carefully plan the sequences by which weapons, levels, layouts and enemies are generated so we can so smoothly, in as few frames as possible while also avoiding a performance hit. This means we have super-clean code that does exactly what we need.
If we were making a traditional roguelike, we could just pack all the generation shit into a single frame and call it a day. And shit, everything could path on every single frame; we'd never have to do the boring shitty work of optimizing that, we could just add more guns!
On top of that, think about how rogue-likes are actually played. There never has to be any worry about how things work "in practice" because the structure means numbers in, numbers out. There also doesn't need to be any concept of balance, because rogue-likes are expected by the audience to have wild imbalances, and most of the gameplay is just discovering them. At no point does the coder have to sit back and see how their game is doing; nothing but actual fail to compile stops them from just plowing forward with more features, and when something does break there is usually just a patch fix.
The internal logic of rogue-likes is also very simple from a computing standpoint. Things that would require complex structures to represent in other games can be done with very primitive code, so people working on rogue-likes have little cause to learn more. I'm probably the least programmer-y person on the face of the planet, but with what I've learned in the last six months just working with a programmer and studying so I can be helpful, I'm fairly certain I could make a simple rogue-like. From there, I could endlessly add bullshit items and systems to it without ever getting any more complex than that in terms of actual infrastructure.
Finally, rogue-likes are 99% UI and it is a simple and eternal fact that all UI code for all video games ever is maximum janky at all times.
TL;DR: To make a rogue-like, you don't have to be a game designer, you don't have to work on art or with an artist, you never have to worry about representation or optimization, you don't ever have to learn anything complicated or challenge yourself in the process in favour of stacking on more primitive structures, and the audience of rogue-likes systematically fail to call this behavior out.
Gahhh! I just started exploring that site and now I wasted an embarrassing amount of time playing this one-dimensional roguelike.
It didn't show up in my browser, only a single link saying edit in JSbin. So I played it with all the source code right there. And I beat it (after many attempts).
Not only is the game genius, it kinda illustrates Axel's previous points about grinding and skipping the grind perfectly. Or, maybe, the opposite points. Probably the opposite points.
I played a game and had the opportunity to change how much health I had at the start, or how much money I had at the start. Seriously, right there in the same browser window!
So after getting used to the game and not beating it, I decided to give myself loads of health, something like 10,000 hp. Weirdly when I spent money to fill my health up again it reduced my health to 100, but then upped my money to just under 10,000. Anyway, I beat the game, and got the message that I'd killed the doomking. Of course I could have just scrolled down through the source and find the same message, but that wouldn't have been as fun.
And then, after "cheating" to beat the game, I put the start health back to 100 and tried beating the game again, and kept playing until I did beat the game. The second time I beat it was waaaaaay more satisfying than the first. I got the same story with the same ending, but this time I had to grind for it! Back and forth from shop to monster to shop, then onto the next level, jumping and fighting.
I guess it illustrates both sides. I could skip the grind, and did, and it wasn't so fun. And yet I also played it through on the harder difficulty, just to prove to myself that I could, even though there was a more optimal path to winning right on my screen.
Comments
Brogue is amazing.
Seriously, I cannot believe that someone gave this guy $30k to work on code of this quality. Full repo here
https://bitbucket.org/dmhagar/empyrea-public/
In Bloodcrusher II, which does a lot of rogue-like generation, my trained attack coder and I have to very carefully plan the sequences by which weapons, levels, layouts and enemies are generated so we can so smoothly, in as few frames as possible while also avoiding a performance hit. This means we have super-clean code that does exactly what we need.
If we were making a traditional roguelike, we could just pack all the generation shit into a single frame and call it a day. And shit, everything could path on every single frame; we'd never have to do the boring shitty work of optimizing that, we could just add more guns!
On top of that, think about how rogue-likes are actually played. There never has to be any worry about how things work "in practice" because the structure means numbers in, numbers out. There also doesn't need to be any concept of balance, because rogue-likes are expected by the audience to have wild imbalances, and most of the gameplay is just discovering them. At no point does the coder have to sit back and see how their game is doing; nothing but actual fail to compile stops them from just plowing forward with more features, and when something does break there is usually just a patch fix.
The internal logic of rogue-likes is also very simple from a computing standpoint. Things that would require complex structures to represent in other games can be done with very primitive code, so people working on rogue-likes have little cause to learn more. I'm probably the least programmer-y person on the face of the planet, but with what I've learned in the last six months just working with a programmer and studying so I can be helpful, I'm fairly certain I could make a simple rogue-like. From there, I could endlessly add bullshit items and systems to it without ever getting any more complex than that in terms of actual infrastructure.
Finally, rogue-likes are 99% UI and it is a simple and eternal fact that all UI code for all video games ever is maximum janky at all times.
TL;DR: To make a rogue-like, you don't have to be a game designer, you don't have to work on art or with an artist, you never have to worry about representation or optimization, you don't ever have to learn anything complicated or challenge yourself in the process in favour of stacking on more primitive structures, and the audience of rogue-likes systematically fail to call this behavior out.
Not only is the game genius, it kinda illustrates Axel's previous points about grinding and skipping the grind perfectly. Or, maybe, the opposite points. Probably the opposite points.
I played a game and had the opportunity to change how much health I had at the start, or how much money I had at the start. Seriously, right there in the same browser window!
So after getting used to the game and not beating it, I decided to give myself loads of health, something like 10,000 hp. Weirdly when I spent money to fill my health up again it reduced my health to 100, but then upped my money to just under 10,000. Anyway, I beat the game, and got the message that I'd killed the doomking. Of course I could have just scrolled down through the source and find the same message, but that wouldn't have been as fun.
And then, after "cheating" to beat the game, I put the start health back to 100 and tried beating the game again, and kept playing until I did beat the game. The second time I beat it was waaaaaay more satisfying than the first. I got the same story with the same ending, but this time I had to grind for it! Back and forth from shop to monster to shop, then onto the next level, jumping and fighting.
I guess it illustrates both sides. I could skip the grind, and did, and it wasn't so fun. And yet I also played it through on the harder difficulty, just to prove to myself that I could, even though there was a more optimal path to winning right on my screen.