Hello!
Just released my most recent homebrew, an NES demake of Thekla's recent indie game "The Witness". You can get it totally free at
https://dustmop.itch.io/the-witnesFor this project, I used cc65 and
shiru's neslib, which made development move a lot faster than raw assembly, and the performance of the game wasn't too affected since a puzzle / exploration game isn't that CPU intensive. Early on I also made some python scripts that added source level debugging to FCEUX (letting me see C code in the FCEUX debugger as comments), which made fixing problems *way* easier. I plan on releasing them very soon on github, in case anyone is interested.
I also made a 4-way scrolling engine for the first time, which was *much* harder than I anticipated. Having C code do this made it easier, but even still, there's so many edge cases and quite a bit of tricky math. It also necessitated writing some parts in asm, for performance's sake.
I'm interested to hear what everyone thinks! This was a really fun project to work on, and so far the people I've shown it to have been fairly positive.
Screenshots:
Attachment:
thewit-garden.png [ 2.36 KiB | Viewed 7062 times ]
Attachment:
thewit-lake.png [ 1.87 KiB | Viewed 7062 times ]
Attachment:
thewit-puzzle.png [ 1.2 KiB | Viewed 7062 times ]
Heh, interesting.
I think it would be really helpful if you accepted diagonal input. Trying to get around corners is a bit of a pain the way it is; it seems like there is a priority to the 4 cardinal directions and it just takes the first 1 on the list. What I'd suggest is that if one direction is blocked, try the other. At intersections favour a turning direction over one that continues along the same path (e.g. holding up+right on a vertical passage means travel up until you hit a possible right turn). It would move a lot more smoothly with some diagonal input possibilities.
Totally true! Someone I was showing it to yesterday said something similar about hugging corners, but I was already committed to releasing today. If I ever do an update that'll definitely be one of the first things to fix.
The gate is open! I'm free!
What's the solution?
Telling you the answer would be to entirely miss the point of both this game and the original one
To be fair, that particular puzzle is introducing a new rule about how those types of symbols combine that no previous puzzle elucidates.
Edit: ...or, never mind, maybe the rule
is discernible from a previous one.
If you've already played through The Witness it's hard to look at these in a fresh way.
pubby wrote:
What's the solution?
Redo the puzzle before that one. Think about what its answer means.
So, brief thoughts: I found the final puzzle much harder than all the ones leading up to it.
I predicted the rule for the grey dots correctly from the get-go.
I found the purple/green dots to be easy but I did have to revise my belief of what the rule was several times.
The yellow blocks took me a while to finally believe I understood what they meant. Actually, I'm still not entirely certain I do, given the penultimate puzzle in that series.
It was a fun way to spend an hour or two. A "backup" button on the line-drawer would be nice, without having to manually backtrack or restart from scratch.
dustmop wrote:
pubby wrote:
What's the solution?
Redo the puzzle before that one. Think about what its answer means.
I found the solution on youtube.
The previous puzzle's rule introduction was ambiguous to me. My first thought was that the grid was on torus, my second thought was that flipping was allowed, and my third thought had to do with binary logic.
Call me stupid, but I really don't enjoy puzzles where the difficulty lies in interpreting unexplained rules. It's as frustrating as that awful riddle I used to hear a lot as a kid:
Quote:
You're in a jail cell with no windows or doors and you have to escape. All you have is a piece of paper. How do you do it?
The answer is "Tear the paper in two to create two halves. Combine the two halves to create a whole. Climb out of the hole."
Sheesh!
Anyway, I noticed that you already got press from gaming sites like Kotaku. Pretty cool!
Did you contact them or did they just find out on their own?
I don't think you're stupid. On the contrary, I agree with you, Rainwarrior, and lidnariq that the last tetris puzzle doesn't proceed as well as it could. Problem is, I sorta coded myself into a corner by only having 5-bits to represent a puzzle on the map, and not having a good place to cut puzzles in order to add some to better introduce the full ruleset. Fixing this would have involved rewriting a ton of the engine, which I decided against, instead releasing a flawed demo. I also developed the tetris section last, didn't playtest it as much as the other parts of the game, and underestimated how hard it was to teach. Learned a lot of lessons for sure!
That being said, I hope you still enjoyed it, despite the moments of frustration. Honestly though, having some frustration actually helps in mimicking the full experience of the original "The Witness" (for better or for worse).
I started to reach out to a hand-full of press people, but it didn't even matter as the ROM went viral pretty fast. I also know that Jonathan Blow saw it and RT'd it very quickly, which was basically the tipping point.
Thanks for the kind words, and also for the totally valid and fair criticisms!
Well, honestly I was impressed with the way you managed to make a small/compact version of it like this.
The original game usually had 20 or more puzzles to teach the rules gradually. You were trying to do it with such a small number of puzzles, and I can see a lot of care went into the planning of their progression. Obviously you kept the overall complexity down as a result.
I'm a little disappointed that I couldn't find any "perspective" puzzles, though.
Lol. Part of the nature of a demake is losing some aspects of the original game, and I made the choice here of completely dropping perspective / environmental puzzles. A commenter on Reddit suggested doing a side-scroller with parallax to simulate them. While that could be interesting, I think a top-down view better preserves the exploration feeling, and also fits in more with how NES games would present this type of world.
Non-spoiler: you can B-dash in this game.
Redrawing bug: bumping into the start circle and withdrawing will take a bite out.
I had trouble with the last one, but it was more an execution than a figuring-out-the-rule. (Figuring out the rule is still annoying, but is so earlier, and…(ROT13)gur nqinaprq lryybj ehyr vf abg arrqrq sbe gur ynfg chmmyr. )
Finally got around to trying this. It's a pretty cool demo. Quite short and easy having beaten the original game, but that's understandable.
The main critique I would have is that moving in the world felt more like a chore rather than having that nice feeling of exploration as in the original game. I'm not sure how much it would be possible to improve on this in a 2D game, though. (I'd expect improved audiovisuals would help somewhat, but who knows how much.) Might be interesting to try the same idea in the context of a 2D platformer.
I also noted the same minor issues with controls as rainwarrior.
That is not a valid solution to that puzzle.