Category: Tech

Some things you want to use a lot of the time (clothes). Some things you want to use sometimes (tire pressure gauge) and somethings barely ever (this hat). I hear you ask: “But what about the popular Javascript state management tool Redux Carly? When do I use that?”

Well, it’s your lucky day because I am here to provide some clarity on when implementing Redux is a good idea and when to skip it. It is a tool that many people have strong opinions on, so let’s dive into it!

If you are unfamiliar, Redux is a Javascript library written by Dan Abramov and Andrew Clark (of React fame). React already has built in state management (useState, useContext etc), but Redux was specifically designed to manage complex state in larger applications, and it can be used outside of React (such as in Angular).

The createStore function is the heavy lifter in the Redux library. From the docs (I added comments):

import { createStore } from 'redux'

//initialize a reducer here
function todos(state = [], action) {
  switch (action.type) {
    //initialize some state 
    case 'ADD_TODO':
      return state.concat([action.text])
    default:
      return state
  }
}

//initialize the store  
const store = createStore(todos, ['Use Redux'])

// update the store
store.dispatch({
  type: 'ADD_TODO',
  text: 'Read the docs'
})

console.log(store.getState())
// [ 'Use Redux', 'Read the docs' ]

The createStore function takes three arguments: a reducer function, the initial state, and any store enhancer functions. The reducer function is intended to return the next state tree, which is rendered relative to the supplied action and state. The state tree which is returned is an object called the Store. The enhancer function (or functions) are intended to allow for the utilization of various middleware tools.

So, similar to useReducer, the way to change/affect the Store is by dispatching an action onto it. However, unlike useReducer, this is the only way to affect the state in the Store. By doing this Redux encourages very strongly directed flow of information in the app, enforcing that all state changes are directed to the global Store using the global dispatch function.

But, you might be saying, “But what if I just want to move my useReducer up into the top level of my app? How is that different from Redux?” I mean, that is essentially how Redux works. However, with Redux you have a clear pathway for unidirectional state flow at a global level in your app out of the box, in addition to the ability to implement lots of cool middleware like the action-logger, which helps you debug complex state. So if you have a need for predictable, clearly enforced data flow in a complex app, don’t reinvent the wheel, and try out Redux.

However, there are some tradeoffs. Dan Abramov himself says that the library is “not designed to be the most performant, or the most concise way of writing mutations. Its focus is on making the code predictable.” Its intent is to help engineers untangle complicated, shared state in complex applications. If you are writing an app with relatively simple state, then Redux is likely entirely unnecessary.

Ultimately, useReducer, useContext, and Redux can all be used in conjunction to manage state in large, complex ecosystems, as its intended purpose was to impose predictability and stable, unidirectional dataflow.

Sources:

1. https://redux.js.org/faq/general#:~:text=Redux%20is%20most%20useful%20in,worked%20on%20by%20many%20people
2. https://changelog.com/posts/when-and-when-not-to-reach-for-redux
3. https://medium.com/@dan_abramov/you-might-not-need-redux-be46360cf367
4. https://blog.isquaredsoftware.com/2017/05/idiomatic-redux-tao-of-redux-part-1/#how-redux-actually-works
5. https://www.robinwieruch.de/redux-vs-usereducer/
6. https://stackoverflow.com/questions/54576865/when-to-use-native-react-usereducer-hook-and-how-it-differentiate-from-redux

You know when you load a site, and random parts of the page seem to take much longer to load than others? In a React site, this is often because one component sits within another. In order to load the child component, whatever data that is needed to render the parent is fetched first, thus causing a delay in loading the child.

function MyProfile({person}) {
 return(
  <MyProfileTheme theme={person.theme} > 
    <MyFace name={person.name} />
    <MyInfo info={person.info} />
  </MyProfileTheme>
)
}

However, when it comes to addressing this issue, there are some considerations. If the developer wants the code to be easily maintainable, they would want the data fetching for each component to be contained within the component itself, as opposed to being coupled to the root component.

export function MyProfileTheme({theme}) {

 const individualTheme = fetchMyTheme() 
 return <IndividualTheme/>

}

But, if components are nested within each other, we get a cascade of fetch requests, as mentioned previously, known as waterfalling. For each request, the client has to make a fetch request, wait for the server to return the data, and then as soon as that is complete, make another. This can definitely affect performance!

Last year, the React team at Meta began developing the concept of moving React components into the server, so that the client only makes one request to the server.

A new type of tag is added to a file name to indicate to React that the file is to be rendered on the server.

App.server.js

(as opposed to Root.client.js)

An important benefit of writing components in this way is that if hefty libraries or dependencies are utilized in a component which is rendered on the server, these dependencies do not have to be rendered by the client.

import { ClunkyDependency} from "somewhere"

function MyProfile({person}) {

const importantThing = ClunkyDependency()

 return(
  <MyProfileTheme theme={person.theme} > 
    <MyFace name={person.name} />
    <MyInfo info={person.info} />
  </MyProfileTheme>
)
}

If MyProfile is a Server Component, then ClunkyDependency is not passed to the client in the bundle. Yay!

However, there are some constraints to utilizing server components, namely that it, in an of itself, cannot be interactive. Functions like useState or event handlers cannot be utilized within a server component. However, they can still import client side components which are interactive, but all the interactive logic must be contained within the child client component.

Additionally, any props passed from the parent server component to the child client component must be serializable, ie then can be encoded into JSON.

export function MyProfileTheme({ theme }) {
  const individualTheme = fetchMyTheme();
  return (
    <>
      <IndividualTheme>
        <ThemeHeader
          function={() => doSomething()}
          title={<div>Header Text</div>}
        />
      </IndividualTheme>
    </>
  );
}

For example, if MyProfileTheme is a sever component, then we cannot pass the ‘function’ prop to the ThemeHeader (functions are not serializable). However, we can pass JSX!

Server Components vs Server Side Rendering

Server components are different from SSR. Server side rendering is when the server renders React into HTML quickly, allowing for a first contentful paint of the page, while the Javascript is still being fetched/applied. This means that the initial visuals of the page load rapidly first, while the interactivity/etc loads next.

Conclusions

All in all, server components provide a helpful option for improving performance and reducing bundle size.

1. https://www.youtube.com/watch?v=TQQPAU21ZUw&ab_channel=MetaOpenSource (Data Fetching with React Server Components)
2. https://dev.to/nas5w/what-is-the-difference-between-server-side-rendering-ssr-and-react-server-components-5339

What’s the difference between var, let and const? Good thing you have this handy blog post to guide you!

The primary differences can be singled down to

• SCOPE
• REDECLARATION
• REASSIGNMENT
• and HOISTING

Scope:

Scope refers to the context in which the variable is able to be referenced.

• var:  local (within a function)* or global **
• let:  block scoped
• const:  block scoped

*Imagine a function, with is an inner block (such an ‘if’ block). If a variable is reassigned within this inner block, and then called outside of the block (but still within the outer function) it will retain its reassigned variable.

**If a variable is declared with var in the global context, it actually becomes a property of the window object.

Reassignment:

The action of assigning a new value to a variable.

• var: can be reassigned
• let: can be reassigned
• const: cannot be reassigned***

***does not mean immutable! When the variable for an object is assigned, the properties and attributes of the object can be changed. But the reference in memory will only be able to point to one place.

Redeclaration:

The re-initialization of a new variable.

• var:  can be redeclared
• let: cannot be redeclared
• const: cannot be redeclared

Hoisting:

Referencing a variable before it is declared.

• var:  hoisted value= undefined
• let: hoisted value= uninitialized and will lead to error
• const: hoisted value= uninitialized and will lead to error

 

The Beginning

In a time, long, long, long ago, in 1995, there was a browser called Netscape.

Ah, old memories.

The minds behind this browser had begun to realize that the simple static webpages of the internet were restrictive, and lacked the ability to respond dynamically to user interaction or input. Netscape wanted to respond to this need by generating reactive functionality within their browser, but they felt the contemporary language of the web, Java, was ill-suited to the task.

So they gave Brendan Eich, the inventor of Lightscript and clearly someone capable of working under pressure, 10 days to generate a programming language to perform this task. And lo and behold, he delivered!

Lightscript (the precursor to Javascript) was developed. Microsoft also developed it’s own language, reverse-engineered off of Lightscript, with significant differences to the Netscape version. These two versions of JS battled it out in something old people remember as the ‘browser wars’ (just kidding you guys). Eventually this was fixed when the ECMA standards specified a single version of the language and we all lived happily ever after.

But who cares? Well, Brendan probably cares. And you should too, because the capabilities that Javascript provided to the browser transformed the way we experience the Internet.

The Age of the Static Page

Imagine only being able to click on buttons, that simply took you to another page. Imagine inputing information in a search box, that when entered also took you to another page. If you wanted to further filter it, you had to reload the page. Every time you wanted to see different data, or have something respond on the page, you had to re-load the page or navigate to a new page.

Sounds real fun. 

Obviously this sucked. Lightscript, and eventually Javascript was developed to address this issue. Eventually, a technique called Dynamic HTML was developed which allowed browsers to move things around the screen, and even change in response to user interaction. In addition to this, Microsoft developers had developed a technology called XMLHttpRequest  in 2000 but after the web development bubble burst and web development in general slowed dramatically. Asynchronous server communications were around, but they were not particularly prominent.

“And some things that should not have been forgotten were lost.”

However, in 2005, a developer named Jesse James Garrett wrote an essay describing a methodology his team utilized, where several technologies were used to mimic the responsiveness of desktop applications. He coined the combination of technologies that enabled this responsiveness AJAX.

AJAX is Born

AJAX stands for Asynchronous Javascript and XML. In Jesse’s own words:

“The Ajax engine allows the user’s interaction with the application to happen asynchronously — independent of communication with the server.”

Information can be displayed on a webpage, whilst data can still be being fetched or processed in the background. This meant that changes could be made to the page or part of the page, and the whole page would not have to be reloaded. This made pages more interactive and performant.

The term AJAX refers to the technologies which make this kind of webpage possible.

How specifically do these technologies function? Well, in order to communicate with another server via Javascript, you need a JS object to provide that functionality. The object we utilize is the aforementioned XMLHttpRequest.  One of the most popular APIs for working with this object to communicate with serves is called the Fetch API. But! Before we talk about the Fetch API, we need to talk about Promises, the foundational modern JS methodology for asynchronous operations.

“I promise to return the result of this asynchronous operation, come hell or high water!”

Promise Me This

Intro to Promises

A promise, I should remind you, is (and I stole this definition pretty much straight from MDN so don’t sue me):

• a JS returned object which represents the result of an asynchronous operation,
• to which you attach callbacks,
• which are then executed depending on what the the state of the promise is when it completes.

Check out this code:

getSomeData("urlWhereDataLives").then(doThisFunction);

Our function getSomeData is an asynchronous one. While that request to the url is pending, the function getSomeData returns a promise. The promise (which is a JS Promise object) has a then() method, which we invoke and pass a callback function. When the data we requested in getSomeData finally arrives, it is passed to the callback function we passed to then().

If you want to handle errors you can pass it a second function:

getSomeData("urlWhereDataLives").then(doThisFunction, oopsErrorHappened);

If there is an error, the second callback is invoked, oopsErrorHappened. However, this is not as common as using the catch() statement:

getSomeData("urlWhereDataLives").then(doThisFunction).catch(oopsErrorHappened);

This seems simple, but in the words of David Flanagan,

“Promises seem simple at first….But they can become surprisingly confusing for anything beyond the simplest use cases.”  – Javascript: The Definitive Guide 

Real talk, David. Real talk.

Promise Chaining

So getSomeData was an asynchronous function. What if we want to parse the data returned by it in another asynchronous operation? Enter Promise chaining:

getSomeData("urlWhereDataLives").then(parseThisData).then(doSomeFunction)

The function parseThisData initiates an asynchronous operation with the results of getSomeData. Then when parseThisData is complete, the results are passed to doSomeFunction.  In this way, the results of one asynchronous operation are passed to another by using the function then() attached to the new Promise objects created when the asynchronous functions are invoked.

Did that make sense? I hope that made sense because this train is rolling on! Choo choo!

Promise Terminology

There are some very specific terms for talking about Promises.

A promise can have 3 states:

• pending – the Promise has no value, it’s chilling
• fulfilled – Sucsess! Some kind of result value has been assigned to the promise
• rejected – no result could be assigned.

When the Promise leaves it’s pending state, it is resolved*.

• resolved – the Promise is either fulfilled or rejected

*According to the Stack Overflow answer which can be found in my references, this is not *technically* accurate, but for the sake of working with Promises, I will call it good enough.

Now lets talk about a common API for handling async operations!

Go Fetch!

The Fetch API is a lovely interface that saves us from the scary oogey-boogey man that is the XMLHttpRequest API, which looks like this:

function makeRequest() {
    httpRequest = new XMLHttpRequest();

    if (!httpRequest) {
      alert('Giving up :( Cannot create an XMLHTTP instance');
      return false;
    }
    httpRequest.onreadystatechange = alertContents;
    httpRequest.open('GET', 'test.html');
    httpRequest.send();
  }

Ugh. Even David Flanagan calls it “old and awkward”.

XMLHttpRequest: “Hey, you better watch the way you talk about me!”

Fetch has a nice, happy 3 step process for making HTTP requests. I will quote my favorite JS book, “Javascript: The Definitive Guide” to summarize them:

1. Call fetch(), passing the URL whose content you want to retrieve.
2. Get the response object that is asynchronously returned by step 1 when the HTTP response begins to arrive, and call a method of this response object to ask for the body of the response.
3. Get the body object that is asynchronously returned by step 2 and process it however you want.

And that’s it! Nothing bad ever happens and that’s all you ever need to think about! Yay!

Thank Goodness! All my problems are solved! This must be Heaven!

Just kidding. It’s usually not that simple. Let’s dive in, shall we?

It’s also important to note that the Fetch API is Promise-based, and there are two asynchronous steps.

1. Call fetch(), passing the URL whose content you want to retrieve.

Give me your basic authorization credentials in your HTTP request header NOW!

Calling fetch initiates an asynchronous operation. The Promise returned by fetch() resolves to a Response object.

Huh?

Well, this sentence just means exactly what it says: when the Promise is resolved, it is now a Response object, thanks to the magic of the Fetch API. We like the Response object, because it offers us all kinds of fun stuff, including two very important methods for parsing data we get from HTTP requests: json(), and text(). More on that later though.

Very often the API that you are requesting stuff from needs you to request it in a specific way. You may need to include authorization credentials, or…other stuff.

So when you call fetch, you might need to provide two arguments: one being the API url, the other being the necessitated header information.

//make new Headers object
let headers = new Headers();

//set the header
headers.set("Authorization", 'Basic ${btoa('${myUsername}:${password}')}');

//include it in fetch request
fetch("myAPI/whereIGetStuff", { headers: header }).then(doSomeStuff)

2.Get the response object that is asynchronously returned by step 1…and call a method of this response object…

Sweet! I got my data from the API in my cool Response object! Now what?

Most commonly the response from a server is a JSON object. So we might want to parse it as a JSON object like in the following code, calling the json() method of the Response object…

fetch("myAPI/whereIGetStuff.com")
    .then(response => response.json()) 
    .then(output => console.log(output))

Or we might want to check the status:

fetch("myAPI/whereIGetStuff.com")
     .then(response => console.log(response.status))

3.Get the body object that is asynchronously returned by step 2 and process it however you want.

When we call fetch() to get data from some browser, the second step, the parsing of the Response into JSON or text, is itself an asynchronous operation.

So another Promise object is created, which we then pass another callback function to invoke when this second Promise is resolved:

fetch("myAPI/whereIGetStuff.com")
.then(response => response.json()) 
.then(myData => console.log(myData))

This callback function is where we finally get to handle our data!

Goodbye Yellow Brick Road

Now, we have come so far!

You were introduced to the history, and therefore necessity of asynchronous code, along with its utility in promoting interactivity in webpages.

You were introduced to the fundamental technology of asynchronous code: the Promise.

And finally you were introduced to one of the most popular and useful API’s for working asynchronously: the Fetch API.

Now forth, and happy asynchronous development!

References:

Marcos Sandrini, JavaScript: its history, and some of its quirks, Nov 3 2021, https://msandrini.medium.com/javascript-its-history-and-some-of-its-quirks-6d5da5e8a47

Aaron Swartz, A Brief History of Ajax, Dec 22 2005, http://www.aaronsw.com/weblog/ajaxhistory

Jesse James Garrett, Feb 18th 2005, Ajax: A New Approach to Web Applications, https://courses.cs.washington.edu/courses/cse490h/07sp/readings/ajax_adaptive_path.pdf

Using Promises – JavaScript | MDN. Developer.mozilla.org. https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Using_promises. Published 2021. Accessed December 16, 2021.

Promise – JavaScript | MDN. Developer.mozilla.org. https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Promise. Published 2021. Accessed December 16, 2021.

Fetch API – Web APIs | MDN. Developer.mozilla.org. https://developer.mozilla.org/en-US/docs/Web/API/Fetch_API. Published 2021. Accessed December 17, 2021.

Flanagan D. Javascript: The Definitive Guide. 7th ed. Sebastopol: O’Reilly; 2020.

What is the correct terminology for javascript promises. Stack Overflow. https://stackoverflow.com/questions/29268569/what-is-the-correct-terminology-for-javascript-promises. Published 2015. Accessed December 17, 2021.