Tuesday, April 23, 2024
HomeGolangThe self-fulfilling prophecy of React

The self-fulfilling prophecy of React

Printed: July 27, 2022

Up to date: August 1, 2022

Scan a wide range of front-end job postings on nearly any dev job board lately, and also you’ll most probably discover one qualification talked about ubiquitously: React. (“Or comparable front-end framework,” the extra open-minded employers would possibly add. They may even listing just a few examples. However React will certainly be there.)

Nonetheless, within the yr 2022, I’ve to marvel why. As a result of (and for those who simply got here right here for the quotable sizzling take, right here it’s, proper up entrance):

React isn’t nice at something besides being standard.

Don’t get me improper: React is good. Even I, a famous and unapologetic React non-fan, readily admit that. Give me a selection between React and no frontend framework in any respect, and given any challenge of adequate dimension or complexity, I’ll fortunately take React, thanks very a lot.

React laid the groundwork for a lot of what different frameworks are doing and have executed. Vue 3 and its composition API are clearly impressed by React hooks. Lots of the conventions of Svelte come from React. Nuxt, a Vue framework I’ve used and love, takes a lot of its concepts (and 75% of its title) from Subsequent, a React meta-framework. For that matter, the complete component-based mannequin of frontend design and improvement owes a lot of its present state (ha) to React—although it ought to be famous others additionally helped pave the best way.

However that’s all sort of my level: very similar to a genre-defining film, album, or online game from a earlier period, React’s greatness is extra in what it meant on the time—and in what it impressed—than what it at the moment is in the present day, absent that context (no pun meant).

Okay, I concede that analogy was a little bit of an overreach. There’s admittedly a lot much less of a distinction between React and its contemporaries than there may be between, say, Citizen Kane and an important film from trendy instances, or between the unique Legend of Zelda on the NES and Breath of the Wild on the Change. That stated:

React has aged. And I don’t assume most individuals—notably these utilizing it usually—understand how a lot or how poorly.

While you dwell solely on this planet of React, you doubtless solely see the enhancements. It looks like React is getting higher and higher on a regular basis. In some ways, it’s. However that perspective shields you from realizing React’s velocity, and its ceiling, have lengthy lagged behind that of its comparable colleagues.

React does what it does properly, but it surely doesn’t do something higher than different frameworks.

Your best option for the job

Let’s say you’re the CTO of a brand-new tech startup, or a solo entrepreneur beginning out constructing some new net software program product of some variety.

You’ve acquired a greenfield challenge forward of you; you possibly can select no matter expertise you wish to construct your factor. No constraints. How do you make the decision on the front-end framework you’ll use for the subsequent a number of years, if not the complete lifespan of your product?

(Assume right here that your challenge is of adequate dimension and complexity that going with out a front-end framework is solely not a good selection.)

There are many issues to think about, together with:

  • Efficiency
  • Studying curve
  • Bundle dimension
  • Scalability
  • Neighborhood and assist
  • Monetary backing
  • Developer expertise
  • Hireability

I’d argue in all of these areas besides perhaps one (emphasis on perhaps), React would really be a dangerous name.

Let’s undergo the listing.


There are many alternative metrics by which you’ll be able to measure efficiency, however no matter what your particular focus is, React just isn’t within the prime tier. Vue, Svelte, Stable, Inferno and a bunch of others usually present markedly higher efficiency than React. Relying in your wants, you may go much more minimal with one thing like Alpine or Petite Vue (although I don’t assume they fairly need to be lumped in the identical classification because the others).

React’s efficiency relative to the sphere is fairly well-documented, so I don’t really feel we have to go into it right here. However suffice to say: in case your aim is to construct probably the most performant factor you possibly can, React isn’t what you’re contemplating.

Preact, a minimal model of React, is worthy of consideration on this case, but it surely’s totally different sufficient from React that it could or is probably not a 1:1 drop-in substitute, so I don’t depend it as the identical factor. Additionally: whereas Preact’s efficiency is considerably higher than React’s, it’s not essentially higher than a few of the different contenders within the subject talked about above.

Studying curve

Assuming you understand nothing about any front-end framework, React is certainly not the simplest to be taught or to make use of.

JSX is a gotcha-riddled kludge to clumsily shoehorn HTML right into a JavaScript operate return. The one factor worse than utilizing JSX with React is not utilizing JSX with React.

Moreover JSX, React itself has its personal bevy of distinctive conventions and trip-ups (to not point out two fully totally different syntaxes which might be not totally interoperable).

In React, many issues different front-end frameworks deal with for you or make trivially straightforward require guide intervention or important boilerplate (or each).

Nearly each framework was initially constructed for the overall inhabitants—besides React, which was initially constructed for Fb engineers. (In equity, React has developed to a common market product since then, however its origins nonetheless have a robust impact on it even in the present day; a lot of its early choices and optimizations are nonetheless unavoidably in place.)

I’m certain I don’t must carry up useEffect right here. Or how one can’t use some commonplace HTML attributes in JSX (as a result of JSX doesn’t know the distinction between a React prop and an HTML attribute). Or memoization. Or how there aren’t actual conditionals, so you need to depend on short-circuit operators. Or how stopping infinite loops is your accountability—despite the fact that there aren’t actually loops, so you need to use array strategies…you understand what, let’s simply transfer on.

Bundle dimension

That is much like velocity, however I felt it was value distinguishing between the 2. It could be one factor if a big bundle despatched over the wire translated to raised efficiency as soon as downloaded, however that’s not the case right here.

React undoubtedly received’t lead the pack in relation to small bundles. Once more, a lot has already been written on this level, so I received’t harp on it.

I will point out, nevertheless: some used to argue React would load from the cache more often than not, so its bundle dimension was irrelevant. That was all the time a spurious declare to start with, however to no matter extent it was true, it’s now not the case in trendy browsers, as a result of safety upgrades that forestall cache sharing between domains.

Right here it additionally bears mentioning that Preact is a chance, however once more: Preact is totally different sufficient that it’s not assured to be a seamless drop-in, and likewise not dramatically totally different, size-wise, from different choices.


Whereas React actually has the most firms and apps demonstrating its capacity to carry out at scale, I don’t assume it’s truthful to say amount equals high quality.

Each main front-end framework, from Vue and Svelte to Angular and Ember and past, has demonstrated capacity to carry out at massive scale. They all have a powerful Nascar-like array of company logos on their homepage.

React doesn’t have something particular right here; it simply has probably the most examples.

If that’s your thought of security, so be it. But it surely’s not truthful to say that React actually demonstrates higher capacity right here. It’s simply been given extra alternatives.

Neighborhood and assist

I contact on this (albeit particularly to Svelte) in one other publish, however to summarize: the concept React has the biggest neighborhood—whereas technically true—doesn’t essentially make it a better option.

An enormous neighborhood generally is a draw back, too, particularly within the case of a so-called “unopinionated” framework equivalent to React. It may possibly imply too many packages to select from, and too many alternative, competing opinions that you just’ll should determine between and take a stance on.

However even moreover that, greater doesn’t essentially imply higher. There’s a degree the place the neighborhood is sufficiently big, and the product doesn’t essentially proceed to get higher past that.

To start with, extra participation builds higher issues. However there’s a degree of diminishing returns (and a degree the place communication value begins to decelerate the challenge moderately than velocity it up). Past that time, there’s no proportional relationship between the variety of folks collaborating in a neighborhood, and the standard of it.

It may be very comforting to be part of the most important neighborhood, as a result of it looks like further safety. The extra folks there are within the boat, the much less doubtless they’ll permit it to sink, or one thing alongside these strains.

However sooner or later, there’s not actually a distinction. Past a sure dimension, a good greater ship with much more folks on it doesn’t actually change something.

You don’t must be within the greatest boat. You simply must be in one which’s sufficiently big to climate the storms, and there are lots of these obtainable.

Monetary backing

Some persons are very involved (maybe for good cause) that their framework of selection might vanish at some point, if it stops being supported and/or maintained. React’s Fb connection is taken into account an indication of belief to these with this viewpoint, however at what level does a challenge have sufficient backing which you can cease worrying?

Angular is backed by Google. Vue is without doubt one of the most profitable and well-funded open-source tasks in historical past. Vercel is at the moment using not less than two Svelte maintainers (together with Svelte’s creator) to work on the challenge full-time. Stable has over 100 contributors and not less than six main company sponsors.

Clearly, backing just isn’t a difficulty amongst main front-end frameworks. So this isn’t actually any level in React’s favor, both.

Developer expertise

Whereas being simply the most-used front-end framework, and the one with probably the most consciousness, React positioned behind each Stable and Svelte by way of satisfaction on this yr’s State of JS Survey outcomes. React additionally positioned behind Svelte, Stable and Vue by way of curiosity (which, for React, was beneath 50%).

React’s satisfaction and curiosity have been declining steadily for years, whereas its utilization has flatlined.

State of JS is only one survey, in fact, and also you get totally different solutions relying on the way you phrase the questions. However different, comparable surveys inform an analogous story. React got here in far behind Svelte and solely barely forward of Vue within the StackOverflow survey as properly, for instance.

Anecdotally: Scott Tolinski talked about on a latest episode of Syntax how one among his builders left for a React job that paid nearly double—however that it was nonetheless a troublesome determination, as a result of this developer cherished working in Tolinski’s SvelteKit codebase a lot.

So whereas builders are clearly each conscious of and utilizing React, it doesn’t appear truthful to say it’s the factor they’re most considering or happy with.


That is the one space the place React definitively comes out forward. If you must rent a dev who is aware of your factor already, React is clearly the selection.

I argue, nevertheless, this isn’t an important cause to decide on a framework.

On condition that your app could have bigger bundles, decrease velocity, and better complexity, as we’ve mentioned—all for little to no profit in different areas—selecting React only for the sake of getting new devs up and operating quicker is short-term achieve for long-term loss. (Tech debt, in different phrases.)

These gambles can repay, however they may not. They usually provide you with just a few weeks now at the price of many months down the street. So whereas that is certainly an actual profit, and an space the place React inarguably comes out forward, it’s not essentially a superb factor to prioritize.

Plus, if you understand React already, studying one other front-end framework shouldn’t be laborious for you. All of them have their very own nuances and quirks, however all of them use roughly the identical psychological mannequin. Any good developer who is aware of React can and can be productive with any framework.

I perceive within the enterprise world, tradeoffs must be weighed. Dev velocity is a crucial consideration. All the objects above are. And also you would possibly determine the choice with decrease velocity, bigger bundles, and better complexity continues to be worthwhile for different good causes and benefits.

However keep in mind that selection additionally offers you completely no tech benefit over your rivals. They’re all (principally) utilizing React, too.

Secure bets don’t repay huge.

So why does React keep on prime?

As a result of we don’t all the time worth the strongest selection as a lot as we worth consensus. (The truth is, you may make an argument that consensus is the strongest selection, in some ways.)

React retains getting picked…as a result of React retains getting picked.

There was a time earlier than React was the default, when folks selected it for different causes. It solved issues that they had, or it was new and attention-grabbing, or no matter else the rationale might need been. However no person was selecting it, in these early days, with any actual inkling that it’d assist them break into the job market, or as a result of it will be the protected wager.

That point is gone. Now, principally, folks select React purely for familiarity and security.

Corporations hold selecting React as a result of they know there’s an enormous pool of candidates who realize it; candidates continue learning React as a result of they know firms are hiring for it. It’s a self-sustaining cycle.

React is the default, and absent a robust cause to decide on in any other case, folks normally simply keep on with the default.

whereas (reactIsPopular) {
  reactIsPopular = true

“No one ever acquired fired for selecting React” just isn’t how the unique saying goes, but it surely would possibly as properly be at this level. It’s the protected choice. It won’t be everyone’s favourite, but it surely doesn’t should be. It most likely received’t be anyone’s least favourite both. It’ll get the job executed.

Absent a robust enterprise want, either side of the market—these hiring and people attempting to get employed—are incentivized to coalesce. React, for higher or worse, is the widespread floor they’ve landed on.

When and the way does this alteration?

That is the half I’m actually considering, having watched the final a number of years of improvement (pun meant).

I don’t have a solution. As I discussed earlier than, React has flatlined by way of utilization, in response to State of JS. (This doubtless means it’s really rising, because it’s the identical slice out of an ever-expanding pie. But it surely’s remained a constant 80% of the pack for 3 years now, which doesn’t look like a robust signal of continued adoption.)

In some unspecified time in the future, although, the cycle breaks. And I don’t know what catalyzes that.

It might be a protracted, sluggish chipping away on the mountain. Tendencies can usually look sudden looking back, however they normally shift steadily over time. Possibly as different frameworks show themselves an increasing number of (and, for that matter, because the variations between them matter much less and fewer from an onboarding standpoint), firms start to decide on different choices the place they’ve the luxurious.

Whereas I don’t assume most firms will do that, there’ll undoubtedly be some that hit a efficiency ceiling with React, and migrate to a extra performant choice primarily based on enterprise want. If your organization lives and dies by cellular utilization, for instance—particularly in locations with comparatively low-powered units or networks—you would possibly (actually) not be capable of afford React’s efficiency tradeoffs.

That received’t be true for many firms, although. It doesn’t make sense emigrate most elderly tasks. The bulk received’t have efficiency points sturdy sufficient to outweigh the price of a serious refactor. That’s wonderful and completely comprehensible. I wouldn’t suggest most companies migrate away from React, provided that it’s working properly sufficient for them. The associated fee doesn’t repay.

So it may be that the winner has been topped, and there can be no successor. On this situation, both React will proceed to evolve as wanted to remain on prime of the heap; or by the point React is dethroned, we could have moved previous front-end frameworks totally (maybe as a result of browsers generally or JavaScript specifically have expanded to the purpose that we don’t want them). Or, not less than, into some new post-framework period.

One different thought right here: it’s doable we’re already transferring previous React, however we simply can’t see it at a excessive stage but. Corporations hiring now are most probably to be the businesses that made their selection on framework already, a very long time in the past. Simply as Covid testing tends to lag behind precise, present ranges by just a few weeks, so present hiring developments might be lagging behind adoption benchmarks.

I don’t know what the way forward for front-end appears to be like like. No one actually does. But it surely’s a really protected wager that React will proceed to be the factor for a really very long time.

If you happen to’re studying front-end improvement within the hopes of getting a job, or attempting to stage up your profession, and React is a information hole for you, it looks like the most secure wager you may make. React isn’t going wherever anytime quickly.

That stated, nevertheless: I hope an increasing number of builders will discover different choices. I hope firms think about different candidates for brand spanking new greenfield tasks. A lot of the enjoyment I’ve found in front-end in recent times has been due to Vue and Svelte. In my private opinion and expertise, React will get the job executed, but it surely doesn’t make the work enjoyable.

I do see a few of this taking place, as extra builders start to marvel what all of the hype is with different frameworks, and understand precisely how obtuse React has really been this whole time. I hope extra will proceed to discover the opposite choices on the market, if for no different cause than higher range of prospects for front-end builders sooner or later.



Please enter your comment!
Please enter your name here

Most Popular

Recent Comments