Create a gist now

Instantly share code, notes, and snippets.

What would you like to do? -- Click ↑↑ fork ↑↑ to play!

A tweet-sized, fork-to-play, community-curated collection of JavaScript.

How to play

  1. Click the Fork button above to fork this gist.
  2. Modify all the files to according to the rules below.
  3. Save your entry and tweet it up!

Keep in mind that thanks to the awesome sensibilities of the GitHub team, gists are just repos. So feel free to clone yours and work locally for a more comfortable environment, and to allow commit messages.


All entries must exist in an index.js file, whose contents are

  1. an assignable, valid Javascript expression that
  2. contains no more than 140 bytes, and
  3. does not leak to the global scope.

All entries must also be licensed under the WTFPL or equally permissive license.

For more information

The site hasn't launched yet, but for now follow @140bytes on Twitter.

To learn about byte-saving hacks for your own code, or to contribute what you've learned, head to the wiki. is brought to you by Jed Schmidt. It was inspired by work from Thomas Fuchs and Dustin Diaz.

/* Rules:
(1) anonymous function or IIFE // make sure
(2) may execute immediately // to annotate
(3) <=140 bytes // your code
(4) no globals // so everyone
(5) permissive license // can learn
(6) have a good time! // from it!
Version 2, December 2004
Everyone is permitted to copy and distribute verbatim or modified
copies of this license document, and changing it is allowed as long
as the name is changed.
// [REQUIRED] A name for your library.
// This must match /^[a-z_]\w*$/i
"name": "140bytes",
// [OPTIONAL] A description of your library, phrased as a verb predicate.
// The gist description is used by default.
"description": "Explain the rules.",
// [OPTIONAL] Up to 5 keywords used for indexing.
"keywords": [

Nit: Twitter allows more than 140 bytes. (Multi-byte characters still only count as one character.)


140bytes commented May 18, 2011

Yeah, I wish they hadn't done that. CJK tweeters get an unfair advantage on Twitter.

Save your entry and tweet it up!

When tweeting up, is any particular manner recommended?

a) #140bytes
b) the code #140bytes
c) just the code
d) ?

jed commented May 21, 2011


Until the site is properly up, a link to the gist with "+@140bytes" is good.

If the code is meant to be assigned to a variable, does the code need to be an self-invoked anonymous function that returns a function, or can the snippet just be a function?

For example:

var method = /* my code snippet injected here */;

var method = (function() {
    return function() {
        // functionality here

// OR:

var method = function() {
    // functionality here

jed commented May 23, 2011

@eliperelman the latter. you only need to do the former if you need to keep some kind of persisted state across calls. see the parseURL function for an example.

Thanks, where do you want completed URLs to be posted?

jed commented May 23, 2011

no need, all you need to do is fork and you'll show up there →

joside commented May 25, 2011

"(2) may be self-executing" ??? What's that?

jed commented May 25, 2011

@joside this is a valid entry: function(){}()

joside commented May 25, 2011

@jed sure this is valid, but this function is definitly not self executing.
Maybe it would be better to say this is a "function invoked directly after declaration"

jed commented May 25, 2011

@joside cool, fixed!

That pattern is called IIFE nowadays.

p01 commented May 29, 2011

The annoted. js and index.js in this gist can be misleading regarding function definition. As seen in , the gist can use a named function because such gist can perfectly be used in ways that do not leak in the global scope.

jo commented May 31, 2011

This package.json is not JSON btw.

I think the inclusion of a test.js to provide usage examples should be encouraged by the instructions.

nesk commented May 31, 2011

↑ I approve this message ! ↑

jed commented May 31, 2011

cool, thanks for all the feedback. i've made some changes, let me know if there's anything i missed.

jo commented May 31, 2011

@jed, thats great!

How about adding a a doctype to test.html, if only to trigger standards mode?

<!DOCTYPE html> 
<div>Expected value: <b>undefined</b></div>
<div>Actual value: <b id="ret"></b></div>
  // write a small example that shows off the API for your example
  // and tests it in one fell swoop.

  var myFunction = function(){ /* the code here should be identical to the entry. */ }

  document.getElementById( "ret" ).innerHTML = myFunction()

Shouldn't there be a test.js that can be copy/pasted into any console (browser, Rhino, Node, etc) instead of constraining it to the browser with test.html?

jed commented Jun 1, 2011

@jed I see you’ve edited test.html, adding the doctype – thanks! However, you’re still closing the body before inserting the <script>, which is syntactically invalid. For inspiration, feel free to have a look at my proposed test.html in my comment above :) As you can see, you could just get rid of the opening and closing body tags, as long as there’s a body-only element (in this case, <div>) preceding the <script>. Alternatively, you could just move the closing </body> tag down.


140bytes commented Jun 1, 2011

cool, fixed.

With a pure-JavaScript test bed, you could potentially allow running tests inline (within the same page that's displaying the gist) when is launched. Arbitrary HTML markup is not the most portable way to capture tests, but it probably wouldn't be a big deal to parse out the tests from HTML...

jo commented Jun 2, 2011

I propose to add another constraint:

  • the function should not access the global scope

This way the function would have to receive globals like document or window as parameter. With this seperation dogma 140bytes would be better testable, smaller and more reusable.

jo commented Jun 2, 2011

I would also prefer a pure test.js. (And maybe it could be better called example.js.)

jo commented Jun 2, 2011

@jed: thanks for starting this über awesome project!

the function should not access the global scope

thats I huge impact! It means I cannot even access Math or Object, without them getting passed to me first! I don't see how this will make entries more testable or smaller( well, maybe). They could be more reusable if you work with frames, and can pass in different docments.

I think this constraint is too much to ask for.

Can core JS objects (e.g. Math) be considered global? Also, does "access" imply "pollute" or does it really mean "read/write"?

jed commented Jun 3, 2011

given that Object === window.Object and so on for these things, i would definitely consider them global.

i'm mostly concerned with unintended side effects due to global pollution, such as leaving temporary variables everywhere.

jo commented Jun 3, 2011

@Kambfhase: I admit I haven't thought of Math and Object and even forgot Date. Shame on me!
I totally agree that the existence those global objects should be implied. Could you agree with

  • the function should not presuppose global objects except the standard global objects

as listed in eg.

I am especially focused on document and window which are not available outside the browser.

I am especially focused on document and window which are not available outside the browser.

true. But if a snippet uses document it most likely has no use outside the browser. And instead of using window, ppl can use self which is the very same, but shorter plus it doesnt fail in ES5 compared to this.

jed commented Jun 3, 2011

there is still much to be done with the site, but for now i am still slammed with client work.

i had this idea, though, which may be relevant: provide a dependencies array in the package.json, which has the IDs of other entries. then, when we build an app containing several entries, your entry would be curried so that the first n arguments are the dependencies themselves.

jo commented Jun 3, 2011

maybe you're right. My underlying intention had arising while trying to validate the gists server side. Maybe thats the only need for that constraint and has no use outside of that specific case.

Any ideas on how to perform server side validation for 140s using document?

jo commented Jun 3, 2011

@jed, those dependencies are great!

Any ideas on how to perform server side validation for 140s using document?

@jo, Envjs might help.

jo commented Jun 3, 2011

@atesgoral: thank you, I will have a look

Any ideas on how to perform server side validation for 140s using document?

Take the "selector function", it uses document. It mainly exists as a shim for older browsers. Is there even a use-case for it on the server? I highly doubt that.

I have mixed feelings about the dependencies array. I thought that each 140bytes entry should be monolithic. If you want to use a snippet you can then simply copy and paste it. Everything bigger can go to .

I have mixed feelings about the dependencies array.

I second that sentiment.

Once you start creating dependencies among libraries/modules, it can quickly turn into a maintenance nightmare. Some of these libraries are still in-flux: their interfaces, or even their behaviour may change. Also, it creates a barrier for newcomers: It creates a constant need to know what's available to use as dependencies. I also have other negative gut-feelings that I cannot articulate with words right now...

I also suggest introducing some kind of " stamp" to certify gists as meeting the requirements and quality standards.

There’s some trailing whitespace in the test.html file. Wasted bytes!

atk commented Jun 22, 2011

Could we somehow clear up the unfilled gists on the right side? or wait for the website

atk commented Jul 1, 2011

I'd like to see something like "golf request" here where developers can enter small snippets with little more than 140bytes so the community can help them to get rid of the last remaining spare bytes.

jed commented Jul 3, 2011


the best bet is to go ahead and submit a work in progress. entries don't need to be within 140 bytes right off the bat.


tsaniel commented Aug 23, 2011

I wonder if there is a 140bytes for css.

p01 commented Aug 23, 2011

@tsaniel: I'd love to see that. I made a couple of tweet-sized user CSS including a TurnOffTheLight for YouTube and Vimeo, and a "noise" filter for Facebook

tsaniel commented Aug 27, 2011

@p01: I'd love to see that too. The 140bytes for javascript lets me learn lots of javascript tricks, I hope the css one will do so.

tsaniel commented Sep 2, 2011

I think we should remind people to name their gist's description.
As it's hard to know whether one really did something or just fork.

jed commented Sep 2, 2011

i dunno, i think the tags on the site are probably good enough, since descriptions are truncated at a few words here anyway.

uhhhhh... How do I get my gists on the website?
Is it automatic? My snippets aren't there :(

atk commented Mar 30, 2012

Since we don't want spam entries, jed usually updates them manually. Since he's currently rather busy as js1k judge, please be patient until he finds the time.

tsaniel commented Aug 2, 2012

Nobody's around...

@tsaniel Every possible function that can fit into 140 characters has been written. Nothing more to see here...

Inkbug commented Aug 2, 2012

@tsaniel @atesgoral What about my function (three comments up) - can anyone shrink it?


tsaniel commented Sep 27, 2012

I can't even reply to any gist, anyone knows what's going on?

atk commented Dec 13, 2012

The new layout breaks some of our concepts... :-(

dvhk commented Aug 2, 2013

Sorry, I don't use twitter, but here's a Lindenmayer system and some basic visualisation:

<title>L-system test</title> <script>

// A Lindenmayer system, see
// Input:rules as a lookup table, start String, and iteration count. Outputs result String
// Example 1: Algae from the Wikipedia page above: f({"A":"AB","B":"A"},"A",3)
// Example 4, note that non-changing symbols also need rules: f({"F":"F+F-F-F+F","-":"-","+":"+"},"F",3)
function f(r,s,i){for(j=i;j--;){var a=s.split('');for(k=a.length;k--;){a[k]=r[a[k]];}s=a.join('');}return s;}

// A shorter-than-256-bytes simple turtle graphics SVG visualisation for the Lindenmayer system output String.
// Usage with Example 4 above: g( f({"F":"F+F-F-F+F","-":"-","+":"+"},"F",3) );
function g(s){var i=[0,3,0,-3],j=[-3,0,3,0],d=9,a=s.split(''),v='';for(k=a.length;k--;){if(a[k]=='F'){v+=' l'+i[d%4]+' '+j[d%4];}if(a[k]=='+'){d++}if(a[k]=='-'){d--}}document.body.innerHTML+="";}

Please enable JavaScript!

Revlin commented Sep 29, 2013

Why did you call the gist LICENSE.txt? Is there a story behind that?

Revlin commented Sep 29, 2013

Alright! I just re-enabled the javascript: uri thingy in firefox and tried

javascript:(function(){var a="-webkit-",b='transform:rotate(1turn);',c='transition:4s;';document.head.innerHTML+='<style>body{'+a+b+a+c+b+c})();


Revlin commented Sep 29, 2013

Yeah, I'm new to this, but I have a bunch of ideas... I think a link to should be included in the tweets, because it gives some coherence to the tweeting, which are now totally random disconnected fragments of javascript that no one other than 140bytes participants will know what they're about or more importantly, where to go to find out what they're about. Yet, there are lots of js/es programmers out there that would be interested in this if they saw an entry tweet and were able to follow a link to the home page.

That's my argument, now the technical challenge: No matter what the size of a URI/URL twitter now minifies (or expands) it to 22 chars, allowing only 118 remaining for the actual code. I kind of think this is acceptable, because this 1) only make the challenge harder, which is the point, right, hardness? and 2) the chars reserved for the link can be considered part of the task, as it helps to promote the task(s) as a whole

atk commented Sep 30, 2013

@Revlin: The point of this challenge is fun and learning. The difficulty of it is balanced between "too easy to be fun" and "too hard to do anything of interest". So making it harder will not improve the challenge, quite the opposite. In addition, I don't think that we need the promotion. We are here for our own entertainment, not for marketing purposes.

Revlin commented Sep 30, 2013

@atk Thanks, dude. I'll start over with all that in mind.

I guess I can be self-defeating and waste 22 chars on including the link myself, without that necessarily being a rule :)

So, according to the instructions all I have to do is fork this repo and my fork will show up on the main page (whether or not I've made any changes)? I'm a bit confused about that... how important is the package.json in this process?

Revlin commented Sep 30, 2013

So I see a button on the main page, "Log in with GitHub," but clicking it has no effect. Is this the way to go to "Save your entry" as suggested in the instructions?

My first crack:

atk commented Oct 1, 2013

Since the last github change, Forks don't show up on the main gist (unless you select "Forks").

The automatical updates of the homepage ( don't work. Jed and I never found out why. Also, we usually only include stuff that has an actual use or at least a cool effect. The package.json should be valid JSON, otherwise the importer will stop (again).

jed commented Oct 3, 2013

hey all. i don't think the rules need to be changed, but i understand that our UI isn't optimal in that case (the platform has changed underneath our feet a few times since launch two years ago).

as for stuff showing up on the site, it doesn't happen until one of the admins stars the gist. this is to prevent spam and make sure that entries are not dupes and are of sufficient quality/creativity.

dluciv commented Nov 25, 2013


Is tweet suitable to submit the gist, or I should reformat it in some other way?

xem commented Dec 5, 2013

Hi all,

Last month, subzey and I invented a system allowing to execute 190 JS characters in one single tweet.


I described it in detail here:

Though, It uses 140 characters, not 140 bytes...

So do you think this system is still suitable for entries?


ghost commented May 24, 2014

Also you can count the bytes with


ghost commented May 24, 2014

I forked it and wrote a simple browser detection script in < 140bytes
tweet link:

Am I missing something or is it really hard to make two forks of the same gist on github? I've been struggling with it for an hour now. Even used the github API to try and fork it to no avail.

As the original site is dead now I created a list of all forks with description. Any feedback is welcome!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment