The first reason is actually why I like arrow function over classic function definition. Code flow for me goes from definition to implementation from top to bottom, instead of what you mention about "important" code at the top. If I need something I like to know this "something" exists before trying to use it, and using arrow function along const enforces just that.

It’s refreshing not seeing a title like “STOP DOING THIS” whenever I see that, I just scroll passed the video

This IS the reason I like them. :) Put helpers functions in another file and load it

The first reason is why I use arrow functions, i have to have some structure in my code not random mess with fn declarations between usage

My opinion. 1. Hoisting -- I personally prefer to write code in this order anyway, the outer layers (i.e. the main logic) is further down, so I always jump straight to the bottom of a file if I want to know what's getting executed. I guess this just might be preference. I also come from a background of pure mathematics where everything is structured in a way where there is no hoisting, everything is ordered naturally, that is, sequentially. Regarding your example with helper functions, these can always be moved to a different file, that makes more sense if you believe there should be a separation.

this video wasnt much much about syntax but more about logic. this was actually awesome knowledge that you gave. thanks!

I agree on using the function keyword for defining small helper functions. But I still like to define them before using them.

as an engineer who has coded in many languages over 20 years, JS is the one that is the most painful to deal with BECAUSE of 'hoisting' and the changing context of 'this' depending on how it's called. Among other things, JS has so many odd quirks that make it a pain compared to many other languages.

for the first reason you mention, I think it's better to split the code to multiple files so when you use them you just import them where you use them

Reason 1 is also why I use arrow functions. Not out of puritanical belief or anything, but the backend language I work with the most is F#, which requires everything to be declared before it can be used. And when writing functional Javascript (or especially Typescript), the languages are actually extremely similar and I think it's the best functional approach. F# enforces order of files too, which can't really be done in Javascript, but I wish it could be.

Slight correction in your event handler, "this" is the same as event.currentTarget, not event.target. The target is the element that was clicked on while currentTarget is the element with the event handler.

Love your channel, man. I love that your videos are concise and straight to the point. Learn a few new things everytime I watch your content, even if it's a topic I already know well and you give me a different perspective on it.

i think that calling things AFTER defining them makes much more sense, I get super confused when the file doesn't have to respect a specific order... and seeing that so many people do terrible crap when they are not forced by linters or interfaces like typescript, I think it's very much needed to adapt to these better practices

Thanks for sharing Kyle. I prefer to use the utils file to store the helper functions and import it to the main file.

I think arrow functions do not require context switching, I may be wrong. Thus for small snippets of code I prefer using arrow functions and for larger snippets I use normal functions. As for organizing code, whether I write code at top of my file or bottom doesn't matter to me that much due to modularity, I do not write any thing else in the main script rather one main function and then calling it. But that's just my style. Honestly I do feel all the rules and guidelines for writing clean code is meaningless if your code is not performent. Most of the time this leads to unnecessary abstruction that often have many unnecessary instructions that are not needed but still called only to keep the structure "clean". For example, looping over same array multiple times in different functions where everything could be done in a single loop causing much less memory access (one of the biggest bottlenecks). Javascript is not a compiled language, but and interpreter. So from compiler diesign perspective, the optimization done by jit compilation is no where near like languages like c, where compilation in release mode throws away all the fancy abstruction and completely reorganize only performent code. Sadly dueto the dynamic nature of js and being and interpreted language this is not possible. So developers should be mindful of the cercumstences. However again it depends on how performent one wants his code to be. For a dataview kind of website it doesn't matter, but for things that requires high amount of computation it matters. As an example, working with canvas 2d if you floor the arguments before passing them to the function, you will see a significant boost in performance. The reason is simple, interpreters are not as smart as compilers when it comes to optimization. The point I want to make is that function calls have costs .

I agree with Kyle here and use normal functions for helper functions and arrow functions for callback functions. It's much easier to identify a function this way, plus hoisting is a bonus.

On a real production project I almost certainly will have each of my functions in a separate file (excluding anonymous functions and callbacks) and when passing callbacks I usually want to bind the parent context like it is default with arrow functions, so I'm the opposite: I use the function keyword only when an arrow function does not cut it. Which is rare.

You make a lot of good points there, especially the ReferenceError which I honestly hadn't come across yet, because I work with nested functions almost exclusively. Maybe it also has something to do with the fact that to me as an old C coder defining a function before use is as natural as breathing. The reason why I love arrow functions so much is because unlike all the syntactic sugar around functions, classes and methods todays programmers grow up with, assigning an arrow function is very close to the bare bones of functions, which, when all's said and done, are very little but an address to jump to in memory. I must say, I disliked JavaScript for the longest time for its lackadaisical typing and conventions, but ever since I began looking at it from the bare-bones perspective, it has fast become my favourite language, even before Lisp and my beloved C

Why would you define the utility function in the same file you call them ?

To me, hoisting is one of the most terrible features of JS, and it should be removed and burned. The idea that your code can reference things that you haven't even defined yet is just asking for spaghetti code. Removing the possibility of hoisting is one of the biggest reasons *in favor* of arrow functions, in my opinion. Module imports can clean it up if you don't like scrolling through the definitions to get to the main logic.

i m a full stack developer.. I know this concepts.. but refreshment of this concepts and the explanation in your videos gives my mind a kind of comfort and boosts confidence for unknown reason.. keep up the good work friend.. your video was inspiring for me to keep working as dev !!

Ok, so now I have two things I disagree with you: not using semicolons, and now the arrow functions :P To me, having to define the functions before they're used is not an issue, but a plus. It helps when you work on a team where usually a few devs are messy with their code, so these small things that force you to be a bit more organized are really helpful. Same goes for some people's neverending creativity for using 'this'. With arrow functions we avoid several errors that cause bugs. So in the last few projects I've worked on, we use the opposite strategy: everything's an arrow function except for some specific cases. As for the syntax, it's just preference/getting used to. Anyways cool video, it's nice to see different approaches / trains of thought.

thank you so much, videos like this one (efficient, dense, clear) help me fill in the significant gaps i have in my knowledge

Wow. Awesome to finally see a Kyle video I don't agree with after many years

I had the "this" problem two days ago and could not understand why the arrow function didn't work, and the function did. Thanks for explaining it! I ended up just telling the linter to ignore my function, lol.

WDS is the mentor we all need.

Thanks for this. As a tech lead I use your entire channel to teach the “Dunning-Kruger Effect” and this video expands the syllabus.

I thought I was alone with this concept. I hate using const over function. Most of the reasons you said but primarily because when I'm scanning code for a variable or function I don't want everything to look like a variable. And hoisting, I don't want to worry about where I'm defining my function. The funny part is it's not even really any shorter of syntaxes. Const name () => is is literally one less character than function name() . You can omit the brackets in arrows but I often use them anyway.

Good points, but this convention breaks as soon as you need to type a function in TypeScript (think of FC). Plus, I've found that relying on hoisting generally results in poorly organized codebases unless you're working on a tiny project w/ a single file (where arguably convention wouldn't matter that much anyway). If you think "what if I want to run some code in a utils.js file where I export my helper functions?", then I would say look back at your architecture and think whether you really want to run arbitrary code in a utilities file that's meant for "utilities". Idk, these are some of my reasonings to use arrows everywhere (except in OOP & decorators). Everyone should use the approach they feel most productive in as long as they know the differences between them 💙

I don't use arrow functions with const/let but didn't even think why. Thank you for telling me why 😂

The people in the comments trying to refute the first point of the video are forgetting one thing: In a team, one programmer should avoid making the dreams of other programmers impossible. In some cases (especially very small libraries) it totally makes sense to put the public API at the top and less important functions at the bottom.

Great vid, I’m an sdet coming from c# and transitioning to Js, your vidz really helped me

Amazing! It's nice to see an expert writing standard function syntax for the exact same reasons I also do.

It seems to me that almost everybody uses arrow functions for the same reason: to give a structure to their code. It's like telling a story without giving spoilers.

as solo dev its okay but when you work with code that someone else wrote, hoisting is pain in the ass because in comparison, if there are no strict rules where we should place functions, hoisted functions is immediate "peek definition", because they can be lower in file or higher, whilst arrow functions enforce you to define them before using, that makes them not significantly easier but quite a bit, also no "this" after dealing with class components in react i truly started hating hoisting

A lot of people commenting about putting helpers in another file... A couple things: 1) Definitely do that. I think Kyle already knows that's best practice. But for the sake of brevity, he makes his point here not mentioning it. 2) If you DIDN'T do that, and kept all helpers in the same file, I think what Kyle says makes perfect sense -- important logic at top, function definitions below. 3) Either way, for DECLARING functions, I agree that the traditional function syntax is clearer -- yes even when declaring them in separate files/modules. 4) Arrow functions are awesome, but 99% of the time I use them I would agree with Kyle here -- it's in the context of anonymous lambdas.

Interesting take. Like a lot of the commenters, I define/declare functions at the top whether using the function keyword or arrow functions. I'll even look at my code below and go back up and cut/paste the functions in the order in which they are called. It just works with my brain that way. But I think one advantage to putting your "important code" at the top is that once the functions are written and tested, you might argue that there is no need to see them because when writing the "important code" and using tested/proven functions, you don't need to be concerned with the implementation. However, that is one of my issues. I always concern myself with the implementation. Slows me down sometimes. 🙂 Another thing I will say is that sometimes it takes a second to see whether something is a function or a const variable if they are all bunched together so it becomes more important for readability to use white space properly and have some order to defining const someFunc => oneliner; vs const someVar = value;

Your use of hoisting is what the Clean Code book teaches. It kinda reads like a newspaper article where the important stuff is on top, the more you go down the more details you get. For example if I'm going to explain how to make a pepperoni pizza, I can say: "Roll the dough out, spread tomato sauce, grated cheese and sliced pepperoni on top in that order. Then put the pizza in a preheated oven on 200 celcius for 30 minutes. Slice the pizza and eat when sufficiently cooled down." This is already pretty verbose but technically for most people this is enough to know what we're making and how. The "helper functions" would then in order that they appear underneath the main function. So in this case: "The dough is made by..." "The tomato sauce is made by..." "You grate the cheese by..." "The cheese is made through..." "Slice the pepperoni in thin 2mm..." "The pepperoni is made by..." "The pizza is sliced in 8 different equal parts" I could put the oven preheating details also in it's separate section but making it too generalized actually takes away descriptive power. So it's important to keep track of your level of abstraction, don't mix the different levels too much and don't go too deep that you get lost. When you open the file you immediately see what the code is doing. You scroll down to see how it is being done exactly.

This is really helpful. I am always trying to find ways to use arrow functions. You make a great point here. Thanks.

I know about hoisting but I don't rely on it. I write regular functions as much as possible on the "top-level" to take advantage of the `function` keyword as you mentioned, and at the top of the script so that it "flows" naturally

I just watched the intro part and I'm already sold. 😊

Thanks, kyle It was a great comparison

Finally now I'm not alone with this opinion

The main reason I would use const arrow is the fact that they can't be modified. If you do (with a function) formatNumbers = [1, 3] somewhere bellow It would "destroy the pointer to the function itself".

As a react developer I always use arrow functions, but examples in this video is cool, it's good to know this stuff

I use function expressions at work because that's what's used there and it's not important enough to rock the boat. But I really dislike them in most cases. I find regular function declaration syntax much easier to read. I use arrow functions a lot, but as in the video, mostly when defined inline as a callback, or if I need to close over `this` (very rare). Hoisting is sometimes annoying but mostly not an issue since most of my functions will be in module files rather than defined where used.

Yea totally agree with all points👌 after my initial arrow function hype I moved back to normal functions in many cases

Definitely the most important reason for understanding whether to use ES5 or ES6 function syntax is with 'this' binding, especially when it relates to creating methods. That is actually important for the functionality. While I appreciate and understand your 'main reason' for better layout flexibility with hoisting, it is simply personal preference and isn't as important imo. Either way, thanks for the reminders. 'This' is and will always be confusing forever.

In short, use arrow functions only for: 1. anonymous callbacks 2. event handlers I completely agree.

Never thought of functions this way. Thank you!

I actually hate hoisting lol, and my helper functions are always in a separate file anyway. But it’s all about taste in the end of the day

For reason 1. If I have many helper functions and especially if I want to use them in many different scopes or in different files. I like to have them in an object, possibilty in a separate file that can be imported. This way, I have a structure where I don't have to scroll past the function expressions all the time as I can keep them in a separate file or collapse them all together, and this works really well with arrow functions. But it really depends on the complexity of the project naturally. E.g.: ``` export const helperFunctions = { helper0: () => { console.log("I help") }, helper1: () => { console.log("I help too") }, helper2: () => { console.log("I help three!") } } ```

Thanks for excellent video and Happy New Year!

Why do you want hoist arrow functions, they're introduced in javascript for the specific cause. I don't understand the argument here.

If you use typscript like almost everyone else, then that means your code will be transpiled into es5 anyways, hence your functions will be actual functions regardless of the declaration method and be hoisted anyways. i got your point tho

My preference is to use arrow functions only where there's an immediate return, or if I need the outer this. `function a(b) { return b+1 }` is less typing than `const a = (b) => { return b+1 }`, but more typing than `const a = (b) => ( b+1 )`. In general, this allows the reader to quickly make assumptions about the code - if you see `=> {`, you know to look for a `this` referencing the outer block

I am a Vue dev and in Vue 3 I exclusively use arrow functions. I a helper function is required at the top it's placed in a file of it's own. I prefer to group blocks based on what they do rather than just put "important" functions at the top.

You just introduced OOP thats great you are learning.

Thank you! I 100% agree.

I discovered THIS the hard way during a implementation of a binary search class object. It was quite interesting.

I’m almost always dealing with other people’s code right now so my own preferences are pretty mercurial, entirely down to how I’m feeling on whatever small project I’m working on. I’m usually consistent within a project, but across projects it can vary. I think I usually use regular functions when just defining them in a module because when I’m naming them arrow functions don’t save much if any typing. That’s about it.

Kyle, have a great 2023!

Thanks Kyle for making me a better JavaScript developer.

I assume you've never worked on a massive code base where you end up in scope hell from legacy code you or 20 other developers wrote. There is a place for everything, and let and anonymous functions are a godsend.

I prefer function keyword because it is clear that it is a function, not a const variable. Readability. It gets confusing when everything starts with const.

When you use typescript, it doesn't matter if you use arrow functions or functions, it wants you to declare them before calling them.

I’ve been doing const arrow functions ALWAYS because it’s just the way we do the internet now…I have missed my hoisted functions so much!!! So much more readable. thank you for permission to mix and match depending on the use case 😊😊

That's a good reason, I also like to have details of implementation at the end of the file below everything else. But there are ways around functions declaration (is that a right name for using function keyword?). When (if) you use react components you can just put helper functions at the bottom, as the component will be ran later than those, so they will exist at that point. Also those helper functions can just go to a separate file (which they probably should). So I usually still use const/let functions.

I always default to arrows unless I need to create a new context. I'm not bothered by functions clogging up the top of file because they probably belong somewhere else if you have that many of them.

As others have said, the flow of code as is executed is generally first defining things and then using them, so having functions before executing them makes sense. Now it’s fine to have a different preference but you can’t escape the reality of how code is executed entirely. If you import a function from another file, for example, it has to be at the top of the file. At best you can partially change the order by utilizing hoisting but not entirely, so you still end up having to jump around a file in certain cases. As someone who values consistency I find it much easier to simply go with the flow of code execution and generally expect entry points to be at the bottom of a file. Then there is the whole thing when working with a team and/or an existing code base, you generally will want to match the style of the project, so always keep that in mind. :)

Makes sense! I knew the fact but never really thought about it until now. :)

Great example! 🧑🏼‍💻

I devide my code in sections, one being helpers functions, all declared with arrow functions. Later I have one update function, which all actions go through, so I’m never in doubt whether I’m looking at a helper function or the core of it, because there is only one core update function. The rest is helpers and sometimes i move to another file and have them all imported. Nice and clean 😊

Hi! Thanks for the video, but I may have misunderstood the first reason because English is not my native language, but it doesn't matter which type of function declaration you choose, arrow or old-style declaration, in any case, hoisting won't work with const or let. That's why the first example is not quite true in your statement. The second reason is the main difference between the two function declarations, because the function call depends on the context in which this function is called.

I love these cleaner code vids.

Another game changing video! Thanks!

This is why I don't like JavaScript (especially in backend). Billions of ways to do a single thing is a double edged sword. Sure it allows more than one way to do a task. However, it adds unnecessary complexity to the codebase.

Totally agree with you on this.

I prefer to read a function definition first and then I look up how it's used, so the arrow functions don't pose a problem for me.

arrow functions inside functions (no "this" hoisting needed) & normal functions in the global scope is how I generally do things.

Nailed it

Really Nice video. Thanks for the clear explanation!

Woah! Thank you for this!

Super helpful, thx!

KYLE THANK YOU AS ALWAYS YOU ROCK

Thanks bro, been distracting myself with code

Another reason some might prefer regular functions over arrow functions for callbacks is that you can name an anonymous regular function. Makes it a bit easier in a stack trace to quickly see the source of the error. Personally not for me, but I did occassionally find it useful when I had to write IE-compatible code back when callbacks were king and promises (let alone async await) weren't around yet. Thank god those days are gone.

1. Use function is more semantic 2. There is no need to use the arrow function to define the functions always in modern JavaScript or TypeScript, especially if the bundle file is not concatenated directly. 3. Because of the scope of the variables in the function, the use of 'function' is clearer. 4. Readable and Maintainable...

Everyone is entitled to their opinion, and if this works for you cool. But if code like you've written were to cross my path professionally on any team I've lead it worked on it'd get kicked back from code review. Hoisting can be a pain to debug if something goes wrong. And while you might not have a problem a jr dev who comes in behind might not be the same.

I always prefer arrow functions. Only in classes it is not allowed.

At work we just use modules for commonly used helper functions and import them. For things like onClick functions we use arrow functions defined at the top of the react file underneath the useState declarations.

Use => for callbacks and single liner codes without { } and return. const sum = (a, b) => a + b;

Usually helpers grouped in some Object thus you will also have an error about initialization. Try to use hook document.load.

I'm with you on using the function keyword for its intended use of... functions. But I'm coming with a serious and long programming background in C-type language.

Abssollutely. Many devs make hype instead of correct usage

This is really a personal preference situation. My brain likes consistency. It’s so difficult for my brain to process differing syntax for the same thing. If I have functions and arrow functions in the same code, then I start to loose track of what’s going on and have to put in more brain power. Its much easier to see => and know that what I have is 99% likely to be an arrow function. For context: I used to be an avid function hoister whenever possible. Context Edit: 1. I don’t used “this” often

Very interesting ! They teach to use arrow which to me is confusing

1) Avoiding hoisting is a good thing. It's not a difficult concept to understand but one that has generally only added confusion and errors in JS. 2) If none of those functions are important, then why were they written? The example here is a case where it is both too trivial to support the argument AND a more complex example would be better done with modules, as others have noted.

Happy New Year Kyle! ❤

I prefer the syntax of the normal functions. That's the main reason for me.