Show HN: A simple Wordle clone in 60 lines, using Hyperscript

jaidhyani 4 years ago |

This is awesome, but there's a slight bug in this implementation: Wordle won't count additional instances of a letter as wrong-position, it'll just flag them as wrong. For example, if the word is TRACE, and I guess TRACT, the first T will be green, but the second T should be black, not yellow. In this implementation, at the moment, TRACT would return GGGGY, but in Wordle it would be GGGGB.

I actually made the same mistake in a Wordle solver I was writing recently - it's easy to miss, since Wordle doesn't make this explicit and you have to infer it yourself.

ajusa 4 years ago | |

Oh wow, I never noticed that myself. I'll fix it once I'm at my computer, thanks for pointing it out! The fix will probably be something like

    if letter not in (.bg-warning in first .guess).innerText then add .bg-warning

Haven't tested as I'm on my phone, but seems like a simple enough fix.

anonymoushn 4 years ago | | |

I don't think it's that simple. For HOUSE, guessing BEERS should get you WYWYW

Aethylia 4 years ago | | |

I have been trying to figure this out myself. The algorithm I'm using is to colour yellow only if the count of that letter before that occurance in the guess, is less than the count of that letter in the target word.

joeframbach 4 years ago | | |

Looks good, here are some test cases off the top of my head.

guess: "SASSY", target: "BRASS", expect either "YYBGB" or "BYYGB". I believe your algorithm would return "YYYGB".

guess: "BRASS", target: "SASSY", expect "BBYGY"

guess: "ASSET", target: "BRASS", expect "YYYBB"

guess: "BRASS", target: "ASSET", expect "BBYYY"

xmonkee 4 years ago | |

Haha, I made the exact same mistake in my wordle solver. Well, not a mistake since that's what the worlde docs say.

aldebran 4 years ago |

I found a bug. When there is a repeated character in the guess but not in the word, you’re given an incorrect color key.

E.g. Word was GROUP I guessed SPOON Output was green O at location 3 and yellow at 4. Happened for a couple more guesses making me think there is another O eventually to realize there are no more spots and it’s a bug.

ajusa 4 years ago | |

Sorry for that (and thanks to all the folks that pointed out the word matching code was wrong). I've pushed an updated version, it's 63 lines now.

Turns out, HN finds and reports bugs much faster than family and friends :)

joeframbach 4 years ago | |

wordle-play.com fails the test. wordlegame.org passes. Implementations disagree, but I agree with how you are defining it should work.

huhtenberg 4 years ago |

Not take away from the brevity of the code, but these 60 lines translate into 168KB of JavaScript.

dane-pgp 4 years ago | |

Wait until you find out how big the browser that runs that code is!

ivanbakel 4 years ago | | |

This is a blithe statement - the browser isn't downloaded whenever a user navigates to the page. 1/6th of a MB for a page this simple is pretty substantial.

antihero 4 years ago | | |

Do you download the browser on every request?

Of course 168kB is still nothing but your comparison is flawed.

teaearlgraycold 4 years ago | | |

Don’t give web developers any ideas.

recursivedoubts 4 years ago | | |

hyperscript is ~22k gziped and cached after the first request

the size is not an issue

now, the runtime on the other hand...

skavi 4 years ago | | |

wait is this not being compiled to js?

irq-1 4 years ago | |

_hyperscript_web.min.js is 84.1 KiB (86,093 bytes) and when zipped it's 22.1 KiB (22,641 bytes)

ajusa 4 years ago | | |

Looks like they doubled the size by accident, in the network debugger the first is a 302 that resolves to the actual file. It shows 84kb twice because of that.

d--b 4 years ago |

Wow, first time I see hyperscript. It is so remote from what I am used to, I have no idea how to read the code in details, but what it does is completely self-explanatory.

aloknnikhil 4 years ago |

On Safari on iOS, at least, typing fast will accidentally zoom into that part of the page instead of registering it as a character press.

But pretty neat implementation otherwise. Awesome work!

ajusa 4 years ago | |

Yep, I have the wrong meta tag I believe:

        <meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no" />

That should fix it, I'll push a new version after fixing the word matching issue others have pointed out. I underestimated Wordle's complexity :)

recursivedoubts 4 years ago | |

i see it slowing down on mobile safari as well

will be a great chance to optimize the hyperscript runtime for mobile

martini333 4 years ago |

Does not check for real words.

e67f70028a46fba 4 years ago | |

“The goal here wasn't really to create a good version of Wordle, it was to build 80% of Wordle in a different language to see what it looks like.”

ajusa 4 years ago | | |

Yup, exactly. I've considered adding a bloom filter check for words but I've accomplished what I wanted to try out. Maybe if I'm ever interested in the future!

It'd be very easy to check against just a list of words, but that'd require more storage and be slower than a bloom filter.

madhato 4 years ago |

I made a simple Wordle clone in ~70 lines without using Hyperscript.

https://dougmcleod.ca/share/wordle.html

dane-pgp 4 years ago |

This isn't just a beautiful demonstration of Hyperscript, it could also be a great introduction to what it can be like when writing for the web platform, or to coding in general. Well done!

fouc 4 years ago |

I'm confused. Hyperscript is supposed to be an alternative way to writing JSX, good for skipping the babel/transpile step. Such as the React h() function. For example:

    h('header', { id: 'my-header', className: 'header' },
      h('div', { className: 'logo' },
        h('a', { href: '/' }, 'mtsknn.fi')
      )
    )

Hyperscript.org doesn't seem to be related to this at all?

https://mtsknn.fi/blog/hyperscript-hyperior-alternative-to-j... https://dev.to/fromaline/hyperscript-the-hidden-language-of-...

speg 4 years ago | |

You are correct. Hyperscript.org is a companion library to htmx; they are totally different from the JSX Hyperscript.

rkimb 4 years ago |

Nice work! I spent some time playing with this today and used it to build a solver algorithm. It's a greedy search through the possible combinations that prioritizes words based on letter frequency and uses the hints provided to whittle down the search space.

Though not enforced by the Hyperwordle clone, the script plays the game on "hard" mode by default where all letter hints must be included in subsequent guesses. Ironically, this constraint made the algorithm more efficient, easy mode tends to take longer and fail more frequently!

Anyone have a better algo yet? https://github.com/rgkimball/wordlebot

tardyp 4 years ago |

Wordle will become the new todolist

e12e 4 years ago |

Nice. Is there a link to the 60 (63) lines of source code? It's not obvious on mobile at least?

Ed, nevermind:

> I highly suggest going to the site and viewing the source!

For chrome on Android:

  view-source:https://arhamjain.com/hyperwordle/

olah_1 4 years ago |

> build 80% of Wordle in a different language

I have been trying to do the same thing! I was looking for projects on github that I could adapt to a new language. Sadly, most of them require a server. It seemed to me that this game could pretty easily just run on the client alone.

I will dig into yours and see how it can be adapted to different languages too :)

mcintyre1994 4 years ago | |

I think wordle does just run indefinitely on the client. It has a word list in its JS which has the list of winning words in order, so you can see the upcoming ones. And the current word is somewhere in localstorage as well. Obviously not much fun if you poke around there too much, but my guess is it just checks the current date and indexes the words array and sets all the state from there.

recursivedoubts 4 years ago |

holy smokes, that's some seriously sweet hyperscript

_Bruno42 4 years ago |

Doesn't work for me, getting a JS error when I'm entering a word:

hyperscript.org@0.9.4:1 'first <:empty/> in first .guess' is null

phkahler 4 years ago |

Doesnt validate input as words,

jFriedensreich 4 years ago |

every key press takes 4 seconds to update the ui, if this is supposed to be a showcase for hyperscript, its not doing too well.

jFriedensreich 4 years ago | |

if it helps for reproducing: i pressed keys too fast wich was interpreted as a double tap in safari and triggered a zoom in. After zooming back to normal, the lag was even worse.

recursivedoubts 4 years ago | |

what machine/browser are you on?

the hyperscript runtime is defintely... pedestrian

jFriedensreich 4 years ago | | |

i tried on ios safari, which is admittedly not a great browser

ajusa 4 years ago | | |

Yeah, that must be a Safari thing. I pulled out my Nexus 5 (circa 2013) and everything feels instant. I don't have any working iOS devices lying around to test on unfortunately.

superlopuh 4 years ago | | |

My Firefox on specced out Intel MBP lags noticeably

stefs 4 years ago |

it's buggy and accepts bogus words

pygy_ 4 years ago |

Hyperscript already has an established meaning in this space, as a pure JS-based syntax alternative to JSX.

See https://github.com/hyperhype/hyperscript and all its descendants, like Mithril, HyperApp or Sinuous.

Could you please think of another name ?

anentropic 4 years ago | |

It seems to be this one:

https://github.com/bigskysoftware/_hyperscript

https://hyperscript.org/

but not sure it's anything to do with the OP though

pygy_ 4 years ago | | |

Look at the source, it is written with that language/framework.

I seems they want to pull a Substack. Color me unimpressed.

speg 4 years ago | | |

He used the framework, not named it.

pygy_ 4 years ago | | |

Oh, my bad... Sorry for the noise.

@dang if you happen to read this, can you hide/flag the thread ?