Redux is fantastic.
Some of you might disagree, so let me tell you why.
Over the last few years, Redux has popularized the idea of using message-passing (also known as event-driven programming) to manage application state. Instead of making arbitrary method calls to various class instances or mutating data structures, we now can think of state as being in a "predictable container" that only changes as a reaction to these "events".
This simple idea and implementation is universal enough to be used with any framework (or no framework at all), and has inspired libraries for other popular frameworks such as:
However, Redux has recently come under scrutiny by some prominent developers in the web community:
Reading some Redux example code I wrote four years ago and I have no effing idea what's going on there22:47 PM - 04 Nov 2019
398
2771
Andrew Clark@acdlite
Redux (complicates things)23:50 PM - 14 Nov 2019
If you don't know these developers, they are the co-creators of Redux themselves. So why have Dan and Andrew, and many other developers, all but forsaken the use of Redux in applications?
The ideas and patterns in Redux appear sound and reasonable, and Redux is still used in many large-scale production apps today. However, it forces a certain architecture in your application:
Andrew Clark@acdliteThe good part of Redux is the reducers.
The bad part is using a single immutable atom for the whole UI and *synchronously* updating it when literally anything changes. People like this because it’s conceptually simple, but it’s a huge architectural constraint to impose on a UI.15:51 PM - 03 Aug 2018
As it turns out, this kind of single-atom immutable architecture is not natural nor does it represent how any software application works (nor should work) in the real-world.
Redux is an alternative implementation of Facebook's Flux "pattern". Many sticking points and hardships with Facebook's implementation have led developers to seek out alternative, nicer, more developer-friendly APIs such as Redux, Alt, Reflux, Flummox, and many more.. Redux emerged as a clear winner, and it is stated that Redux combines the ideas from:
However, not even the Elm architecture is a standalone architecture/pattern, as it is based on fundamental patterns, whether developers know it or not:
Rather than someone inventing it, early Elm programmers kept discovering the same basic patterns in their code. It was kind of spooky to see people ending up with well-architected code without planning ahead!
@brian_d_vaughn Redux is half of a pattern. The other half is a wild west mix of middleware, indirection, and unnecessary abstractions.13:27 PM - 05 Dec 2019
In this post, I will highlight some of the reasons that Redux is not a standalone pattern by comparing it to a fundamental, well-established pattern: the finite state machine. This is not an arbitrary choice; every single application that we write is basically a state machine, whether we know it or not. The difference is that the state machines we write are implicitly defined.
I hope that some of these comparisons and differences will help you realize how some of the common pain points in Redux-driven applications materialize, and how you can use this existing pattern to help you craft a better state management architecture, whether you're using Redux, another library, or no library at all.
What is a finite state machine?
(Taken from another article I wrote, The FaceTime Bug and the Dangers of Implicit State Machines):
Wikipedia has a useful but technical description on what a finite state machine is. In essence, a finite state machine is a computational model centered around states, events, and transitions between states. To make it simpler, think of it this way:
- Any software you make can be described in a finite number of states (e.g.,
idle,loading,success,error) - You can only be in one of those states at any given time (e.g., you can’t be in the
successanderrorstates at the same time) - You always start at some initial state (e.g.,
idle) - You move from state to state, or transition, based on events (e.g., from the
idlestate, when theLOADevent occurs, you immediately transition to theloadingstate)
It’s like the software that you’re used to writing, but with more explicit rules. You might have been used to writing isLoading or isSuccess as Boolean flags before, but state machines make it so that you’re not allowed to have isLoading === true && isSuccess === true at the same time.
It also makes it visually clear that event handlers can only do one main thing: forward their events to a state machine. They’re not allowed to “escape” the state machine and execute business logic, just like real-world physical devices: buttons on calculators or ATMs don’t actually do operations or execute actions; rather, they send "signals" to some central unit that manages (or orchestrates) state, and that unit decides what should happen when it receives that "signal".
What about state that is not finite?
With state machines, especially UML state machines (a.k.a. statecharts), "state" refers to something different than the data that doesn't fit neatly into finite states, but both "state" and what's known as "extended state" work together.
For example, let's consider water 🚰. It can fit into one of four phases, and we consider these the states of water:
liquid-
solid(e.g., ice, frost) -
gas(e.g., vapor, steam) plasma
Water phase UML state machine diagram from uml-diagrams.com
However, the temperature of water is a continuous measurement, not a discrete one, and it can't be represented in a finite way. Despite this, water temperature can be represented alongside the finite state of water, e.g.:
-
liquidwheretemperature === 90(celsius) -
solidwheretemperature === -5 -
gaswheretemperature === 500
There's many ways to represent the combination of finite and extended state in your application. For the water example, I would personally call the finite state value (as in the "finite state value") and the extended state context (as in "contextual data"):
const waterState = {
value: 'liquid', // finite state
context: { // extended state
temperature: 90
}
}
But you're free to represent it in other ways:
const waterState = {
phase: 'liquid', // finite state
data: { // extended state
temperature: 90
}
}
// or...
const waterState = {
status: 'liquid', // finite state
temperature: 90 // anything not 'status' is extended state
}
The key point is that there is a clear distinction between finite and extended state, and there is logic that prevents the application from reaching an impossible state, e.g.:
const waterState = {
isLiquid: true,
isGas: true, // 🚱 Water can't be both liquid and gas simultaneously!
temperature: -50 // ❄️ This is ice!! What's going on??
}
And we can extend these examples to realistic code, such as changing this:
const userState = {
isLoading: true,
isSuccess: false,
user: null,
error: null
}
To something like this:
const userState = {
status: 'loading', // or 'idle' or 'error' or 'success'
user: null,
error: null
}
This prevents impossible states like userState.isLoading === true and userState.isSuccess === true happening simultaneously.
How does Redux compare to a finite state machine?
The reason I'm comparing Redux to a state machine is because, from a birds-eye view, their state management models look pretty similar. For Redux:
state+action=newState
For state machines:
state+event=newState+effects
In code, these can even be represented the same way, by using a reducer:
function userReducer(state, event) {
// Return the next state, which is
// determined based on the current `state`
// and the received `event` object
// This nextState may contain a "finite"
// state value, as well as "extended"
// state values.
// It may also contain side-effects
// to be executed by some interpreter.
return nextState;
}
There are already some subtle differences, such as "action" vs. "event" or how extended state machines model side-effects (they do). Dan Abramov even recognizes some of the differences:
@FransBouma Ignoring the CS education shaming, I *was* aware of what a state machine is when I worked on Redux. However it hasn’t been a great way to explain it to people in practice, which is why it’s not in README. Additionally, reducers don’t model transitions so it’s not 1:1 analogy.23:28 PM - 19 Nov 2018
A reducer can be used to implement a finite state machine, but most reducers are not modeled as finite state machines. Let's change that by learning some of the differences between Redux and state machines.
Difference: finite & extended states
Typically, a Redux reducer's state will not make a clear distinction between "finite" and "extended" states, as previously mentioned above. This is an important concept in state machines: an application is always in exactly one of a finite number of "states", and the rest of its data is represented as its extended state.
Finite states can be introduced to a reducer by making an explicit property that represents exactly one of the many possible states:
const initialUserState = {
status: 'idle', // explicit finite state
user: null,
error: null
}
What's great about this is that, if you're using TypeScript, you can take advantage of using discriminated unions to make impossible states impossible:
interface User {
name: string;
avatar: string;
}
type UserState =
| { status: 'idle'; user: null; error: null; }
| { status: 'loading'; user: null; error: null; }
| { status: 'success'; user: User; error: null; }
| { status: 'failure'; user: null; error: string; }
Difference: events vs. actions
In state machine terminology, an "action" is a side-effect that occurs as the result of a transition:
When an event instance is dispatched, the state machine responds by performing actions, such as changing a variable, performing I/O, invoking a function, generating another event instance, or changing to another state.
This isn't the only reason that using the term "action" to describe something that causes a state transition is confusing; "action" also suggests something that needs to be done (i.e., a command), rather than something that just happened (i.e., an event).
So keep the following terminology in mind when we talk about state machines:
- An event describes something that occurred. Events trigger state transitions.
- An action describes a side-effect that should occur as a response to a state transition.
The Redux style guide also directly suggests modeling actions as events:
However, we recommend trying to treat actions more as "describing events that occurred", rather than "setters". Treating actions as "events" generally leads to more meaningful action names, fewer total actions being dispatched, and a more meaningful action log history.
Source: Redux style guide: Model actions as events, not setters
When the word "event" is used in this article, that has the same meaning as a conventional Redux action object. For side-effects, the word "effect" will be used.
[state, dispatch] = useReducer()
Is it too late to have everybody call the object you pass to dispatch "event" instead of "action"?
dispatch(event)
dispatch(action)
Probably not. A boy can dream.00:55 AM - 19 Jan 2020
Difference: explicit transitions
Another fundamental part of how state machines work are transitions. A transition describes how one finite state transitions to another finite state due to an event. This can be represented using boxes and arrows:
This diagram makes it clear that it's impossible to transition directly from, e.g., idle to success or from success to error. There are clear sequences of events that need to occur to transition from one state to another.
However, the way that developers tend to model reducers is by determining the next state solely on the received event:
function userReducer(state, event) {
switch (event.type) {
case 'FETCH':
// go to some 'loading' state
case 'RESOLVE':
// go to some 'success' state
case 'REJECT':
// go to some 'error' state
default:
return state;
}
}
The problem with managing state this way is that it does not prevent impossible transitions. Have you ever seen a screen that briefly displays an error, and then shows some success view? If you haven't, browse Reddit, and do the following steps:
- Search for anything.
- Click on the "Posts" tab while the search is happening.
- Say "aha!" and wait a couple seconds.
In step 3, you'll probably see something like this (visible at the time of publishing this article):
After a couple seconds, this unexpected view will disappear and you will finally see search results. This bug has been present for a while, and even though it's innocuous, it's not the best user experience, and it can definitely be considered faulty logic.
However it is implemented (Reddit does use Redux...), something is definitely wrong: an impossible state transition happened. It makes absolutely no sense to transition directly from the "error" view to the "success" view, and in this case, the user shouldn't see an "error" view anyway because it's not an error; it's still loading!
You might be looking through your existing Redux reducers and realize where this potential bug may surface, because by basing state transitions only on events, these impossible transitions become possible to occur. Sprinkling if-statements all over your reducer might alleviate the symptoms of this:
function userReducer(state, event) {
switch (event.type) {
case 'FETCH':
if (state.status !== 'loading') {
// go to some 'loading' state...
// but ONLY if we're not already loading
}
// ...
}
}
But that only makes your state logic harder to follow because the state transitions are not explicit. Even though it might be a little more verbose, it's better to determine the next state based on both the current finite state and the event, rather than just on the event:
function userReducer(state, event) {
switch (state.status) {
case 'idle':
switch (event.type) {
case 'FETCH':
// go to some 'loading' state
// ...
}
// ...
}
}
You can even split this up into individual "finite state" reducers, to make things cleaner:
function idleUserReducer(state, event) {
switch (event.type) {
case 'FETCH':
// go to some 'loading' state
// ...
}
default:
return state;
}
}
function userReducer(state, event) {
switch (state.status) {
case 'idle':
return idleUserReducer(state, event);
// ...
}
}
But don't just take my word for it. The Redux style guide also strongly recommends treating your reducers as state machines:
[...] treat reducers as "state machines", where the combination of both the current state and the dispatched action determines whether a new state value is actually calculated, not just the action itself unconditionally.
I also talk about this idea in length in my post: No, disabling a button is not app logic.
Difference: declarative effects
If you look at Redux in isolation, its strategy for managing and executing side-effects is this:
¯\_(ツ)_/¯
That's right; Redux has no built-in way of handling side-effects. In any non-trivial application, you will have side-effects if you want to do anything useful, such as make a network request or kick off some sort of async process. Importantly enough, side-effects should not be considered an afterthought; they should be treated as a first-class citizen and uncompromisingly represented in your application logic.
Unfortunately, with Redux, they are, and the only solution is to use middleware, which is inexplicably an advanced topic, despite being required for any non-trivial app logic:
Without middleware, Redux store only supports synchronous data flow.
Source: Redux docs: Async Flow
With extended/UML state machines (also known as statecharts), these side-effects are known as actions (and will be referred to as actions for the rest of this post) and are declaratively modeled. Actions are the direct result of a transition:
When an event instance is dispatched, the state machine responds by performing actions, such as changing a variable, performing I/O, invoking a function, generating another event instance, or changing to another state.
_Source: (Wikipedia) UML State Machine: Actions and Transitions
This means that when an event changes state, actions (effects) may be executed as a result, even if the state stays the same (known as a "self-transition"). Just like Newton said:
For every action, there is an equal and opposite reaction.
Source: Newton's Third Law of Motion
Actions never occur spontaneously, without cause; not in software, not in hardware, not in real life, never. There is always a cause for an action to occur, and with state machines, that cause is a state transition, due to a received event.
Statecharts distinguish how actions are determined in three possible ways:
- Entry actions are effects that are executed whenever a specific finite state is entered
- Exit actions are effects that are executed whenever a specific finite state is exited
- Transition actions are effects that are executed whenever a specific transition between two finite states is taken.
Fun fact: this is why statecharts are said to have the characteristic of both Mealy machines and Moore machines:
- With Mealy machines, "output" (actions) depends on the state and the event (transition actions)
- With Moore machines, "output" (actions) depends on just the state (entry & exit actions)
The original philosophy of Redux is that it did not want to be opinionated on how these side-effects are executed, which is why middleware such as redux-thunk and redux-promise exist. These libraries work around the fact that Redux is side-effect-agnostic by having third-party, use-case specific "solutions" for handling different types of effects.
So how can this be solved? It may seem weird, but just like you can use a property to specify finite state, you can also use a property to specify actions that should be executed in a declarative way:
// ...
case 'FETCH':
return {
...state,
// finite state
status: 'loading',
// actions (effects) to execute
actions: [
{ type: 'fetchUser', id: 42 }
]
}
// ...
Now, your reducer will return useful information that answers the question, "what side-effects (actions) should be executed as a result of this state transition?" The answer is clear and colocated right in your app state: read the actions property for a declarative description of the actions to be executed, and execute them:
// pretend the state came from a Redux React hook
const { actions } = state;
useEffect(() => {
actions.forEach(action => {
if (action.type === 'fetchUser') {
fetch(`/api/user/${action.id}`)
.then(res => res.json())
.then(data => {
dispatch({ type: 'RESOLVE', user: data });
})
}
// ... etc. for other action implementations
});
}, [actions]);
Having side-effects modeled declaratively in some state.actions property (or similar) has some great benefits, such as in predicting/testing or being able to trace when actions will or have been executed, as well as being able to customize the implementation details of executing those actions. For instance, the fetchUser action can be changed to read from a cache instead, all without changing any of the logic in the reducer.
Difference: sync vs. async data flow
The fact is that middleware is indirection. It fragments your application logic by having it present in multiple places (the reducers and the middleware) without a clear, cohesive understanding of how they work together. Furthermore, it makes some use-cases easier but others much more difficult. For example: take this example from the Redux advanced tutorial, which uses redux-thunk to allow dispatching a "thunk" for making an async request:
function fetchPosts(subreddit) {
return dispatch => {
dispatch(requestPosts(subreddit))
return fetch(`https://www.reddit.com/r/${subreddit}.json`)
.then(response => response.json())
.then(json => dispatch(receivePosts(subreddit, json)))
}
}
Now ask yourself: how can I cancel this request? With redux-thunk, it simply isn't possible. And if your answer is to "choose a different middleware", you just validated the previous point. Modeling logic should not be a question of which middleware you choose, and middleware shouldn't even be part of the state modeling process.
As previously mentioned, the only way to model async data flow with Redux is by using middleware. And with all the possible use-cases, from thunks to Promises to sagas (generators) to epics (observables) and more, the ecosystem has plenty of different solutions for these. But the ideal number of solutions is one: the solution provided by the pattern in use.
Alright, so how do state machines solve the async data flow problem?
They don't.
To clarify, state machines do not distinguish between sync and async data flows, because there is no difference. This is such an important realization to make, because not only does it simplify the idea of data flow, but it also models how things work in real life:
- A state transition (triggered by a received event) always occurs in "zero-time"; that is, states synchronously transition.
- Events can be received at any time.
There is no such thing as an asynchronous transition. For example, modeling data fetching doesn't look like this:
idle . . . . . . . . . . . . success
Instead, it looks like this:
idle --(FETCH)--> loading --(RESOLVE)--> success
Everything is the result of some event triggering a state transition. Middleware obscures this fact. If you're curious how async cancellation can be handled in a synchronous state transition manner, here's a couple of guiding points for a potential implementation:
- A cancellation intent is an event (e.g.,
{ type: 'CANCEL' }) - Cancelling an in-flight request is an action (i.e., side-effect)
- "Canceled" is a state, whether it's a specific state (e.g.,
canceled) or a state where a request shouldn't be active (e.g.,idle)
To be continued
It is possible to model application state in Redux to be more like a finite state machine, and it is good to do so for many reasons. The applications that we write have different modes, or "behaviors", that vary depending on which "state" it's in. Before, this state might have been implicit. But now, with finite states, you can group behavior by these finite states (such as idle, loading, success, etc.), which makes the overall app logic much more clear, and prevents the app from getting stuck in an impossible state.
Finite states also make clear what events can do, depending on which state it's in, as well as what all the possible states are in an application. Additionally, they can map one-to-one to views in user interfaces.
But most importantly, state machines are present in all of the software that you write, and they have been for over half a century. Making finite state machines explicit brings clarity and robustness to complex app logic, and it is possible to implement them in any libraries that you use (or even no library at all).
In the next post, we'll talk about how the Redux atomic global store is also half of a pattern, the challenges it presents, and how it compares to another well-known model of computation (the Actor model).
Cover photo by Joseph Greve on Unsplash
Top comments (64)
So just to verify my own understanding, Redux is half a pattern because it doesn't fully match up with Finite State Machines.
I think the comparison between the two is "loose" and not a strict distinction and in that sense yes Redux is half a pattern. Or more correctly it shouldn't be directly compared to FSMs, as it isn't a full representation of FSMs.
I think its possible to get Redux to be a "full pattern" that correctly replicates FSMs, but I don't think its something we would want. There's already a verboseness to Redux without specifying how side effects are handled and or how strict transitions are handled.
Also Redux maybe half a pattern, but this is half a full post 😉 (sorry couldn't help myself haha)
PS. I use NgRx, which provides effect handling and cancellations, but this isn't a Redux/NgRx feature, it an Observable/rxjs feature.
Hey Brad 👋🏻. I think what @davidkpiano is trying to convey on Redux to be a "full pattern" is instead of just thinking about the new state; it should embrace side effects as well because any non-trivial app has to deal with side effects. Right now side-effects are treated as second-class citizens and pushed to middlewares to solve the problem. Having this "disconnect" makes it a half pattern.
The same could be said about how Redux doesn't handle displaying the view, or how it handles taking user actions, which are usually React's job.
Redux is a lib, not a framework in itself, it leaves the side-effects up to other libs. Its obviously a design decision that went well with React at the time. React handle the UI, Redux handles the state, async stuff handled by other libs.
Now that React has "moved on" to being able to manage state by itself via hooks, contexts, etc. Looking back it seemed like Redux wasn't "good enough", which may or may not be true, but hindsight is 2020.
The thing with comparing Redux to any other pattern is there is going to be a disconnect, as its job isn't to be a "full pattern", its job is to manage state, and it handles that well.
But the purpose of an application shouldn't be managing state, that's a collateral.
So the fact that redux is a collateral is the failure of an ecosystem.
After all, it was made just to solve a react problem, and to solve the redux problem, you need middleware.
Perhaps there's a need for a complete solution that deal with all aspects of modelling UI.
You have your head deep down CQRS and are not seeing the forest for the trees anymore.
Your right that managing state is collateral of the pattern React produces, where React handles the UI and now something else must handle the state.
Idk if I'd go as far to say "this is a failure of the ecosystem", when such a system is so successful at what it does. You don't need to be the perfect solution, just good enough.
Its easy to point at flaws, point at solutions to those specific flaws, and then state the original system is a failure. As no system is perfect, but the real question is what kind of benefit do you get out of the solution, and is it worth it to move? Are the "failures" that bad?
I'm not sure what a "complete solution" looks likes, idk if anyone does. I'd even say such might just be the "perfections fallacy framework".
"when such a system is so successful at what it does"
I wouldn't call such a system so successful, it does solve a specific problem, albeit also creating more problems. But that could be said about anything.
That's why its a failure of the ecosystem, each system in itself is good and fine, but the sum of all them is less than great.
" Its easy to point at flaws, point at solutions to those specific flaws, and then state the original system is a failure. "
Pointing failures is a bold action when everyone is blinded by hype.
"As no system is perfect, but the real question is what kind of benefit do you get out of the solution, and is it worth it to move? Are the "failures" that bad? "
That's the exact question I started asking, is it worth using React, yes, the failures are bad, in as much as I'm seriously considering parting ways with using react to build user interfaces.
Indeed, I'm already did that, I only use Inferno now, because I migrated to another entire ecosystem, I'm either using ReasonML or Fable with Elm approach for constructing user interfaces and managing messaging and state.
React was only doing the rendering, not even components and JSX I was using, only functions and render.
"perfections fallacy framework"
The entire idea of HTML+JS+CSS is a perfection framework, perhaps I just wanted a canvas and webassembly to make applications, because HTML was made for documents.
Ecosystem failure, yes it is.
Pointing errors is easy, pointing the solution is actual hard.
This industry is built on this idea of disrupting but it won't release the bone, there's no meat anymore in HTML...
(time to bring back VRML, just kidding, but seriously considering)
"Pointing failures is a bold action when everyone is blinded by hype."
I'm not sure what being "bold" with claims has to do with anything beyond ego. Nor am I sure if its reasonable to say React is "blinding by hype" at this point. At a bare minimum, hype pays the bills since React is still popular and being asked for in the job market. At a worse case its a straw man argument ("I'm right because everyone else is blind!")
"The entire idea of HTML+JS+CSS is a perfection framework, perhaps I just wanted a canvas and webassembly to make applications, because HTML was made for documents."
I don't think you understood what I meant by "perfectionist fallacy framework". The perfectionist fallacy refers to the idea "something perfect exists". I never claimed HTML+JS+CSS is a "perfect framework", nor did I claim it itself is a framework.
Furthermore, you claim "the failures of React are bad", then provide 1 example, where "React was only doing the rendering...". Which by itself sounds exactly like what React should be doing. I'm not sure if that is suppose to provide support to the claim "React is a failure", or just provide support for the alternatives you directly mention your moving toward. Unless of course your trying to compare apples to oranges, where you compare React by itself to a more "full" solution.
"This industry is built on this idea of disrupting but it won't release the bone, there's no meat anymore in HTML..."
I don't see how HTML has anything to do with disruptive tech. Or more relevantly, what frameworks do to disrupt really anything other developer viewpoints and egos.
"That's why its a failure of the ecosystem, each system in itself is good and fine, but the sum of all them is less than great."
I can see the ecosystem being multiple parts being itself an issue, but I can also see it as an advantage for those with the manpower and capabilities to leverage each part correctly together, and interchange each part. To label the entire ecosystem a failure, and together "less than great" only works if you give solid examples for all cases. At a bare minimum, it should be agreeable the React ecosystem is successful when used correctly, and its possible you can screw it up with the freedom it gives.
"I'm not sure what being "bold" with claims has to do with anything beyond ego."
It has to do with going against the crowd, that has nothing to do with being motivated by ego.
Its just a strategy, following the safe path versus going slicing the vegetation into the unknown, which has risks, but also benefits.
"Nor am I sure if its reasonable to say React is "blinding by hype" at this point"
Are we past this idea of using React everywhere, so we can start considering where its applicable ? I think not yet.
See all this discussion I caused only because I considered possibly it is not a good answer, and even considered the entire web as not the answer for the problem.
"hype pays the bills"
great argument, paying the bills, couldn't care less about anything technical do we ? better stick with ego motivation, at least it makes sense, anything could pay the bills, even flipping burgers, are we discussing merits of technology choice or what?
"React is still popular and being asked for in the job market"
Oh, the appeal for popularity, that's also a fallacy, its not an argument.
"I'm right because everyone else is blind"
Its not about being right, the moment I start choosing the technology I use based on the problem I need to solve, when others choose based on hype/popularity/job_market, I'm a bit more right, at least in my specific context/use.
I could be wrong or not, but its a bigger mistake to delegate important choices to the crowd.
Its the same argument as the 'ol you can't go wrong by choosing IBM/Microsoft/Oracle or whatever, yeah, technically no, but realistically you are also not right in your choice, popularity is the worst argument for choosing anything.
"I don't think you understood what I meant by "perfectionist fallacy framework". The perfectionist fallacy refers to the idea "something perfect exists". I never claimed HTML+JS+CSS is a "perfect framework", nor did I claim it itself is a framework. "
No, I understood it, I am seriously claiming that HTML/JS/CSS is not perfect, which should've been obvious, people just assume that without even realized, good think you didn't, because that's not what people usually do, it was just a counter-point, perhaps badly executed.
And HTML/JS/CSS are a framework, by definition, they are outside your code, and which totally dictate how your code is organized, they are not libraries, the "web" is a framework, a very bad one for making applications, but a good one for documents, which it was intended to.
And its not about being perfect, the problem with the web tech is that it is the lowest common demonominator, we don't want perfection, that doesn't mean we should accept the lowest possible solution either.
Something better must be possible, I'm betting my business on it.
Some people seem to be too invested to even consider the possibility of something better than the web...
That's their problem, not mine.
" "React was only doing the rendering...". Which by itself sounds exactly like what React should be doing. "
If React only did that... It also has hooks and try to manage state, one of the reasons I ditched it, I only need rendering, not even VDom, what if this idea of having the VDom is a wrong one?
Then that could be considered as a failure, my failure in using React to being with. To be fair, that's not React's failure, but kind of is the ecosystem failure if you go with the crowds (and they are part of an ecosystem) and just use what others are using because of popularity, that was my mistake, it won't happen again.
That wasn't supposed to be an example of were React fails. that's what I mean by ecosystem. React on its own is not bad, but it does create more problems to be solved, that's accidental complexity.
The web is rotten with accidental complexity.
My lens of failure is very simple, is a tool fit for that purpose, React is fit for its purpose, although with too much overhead.
What I claim failure is the composition of the entire ecosystem of modern web-applications, because its clearly not made for that purpose.
Of course React doesn't have anything to do with that, its just that I even consider if the "purpose" of what React achieves was even necessary. That's the question, and there's where's the biggest failure, we are using one thing made for one purpose to another purpose and claiming it is good, when there are clearly examples (like the post) which shows the problems.
When you decision, which should've been no brainer get this reaction...
I've just decided to go against the idea of needing React, not because it is a failure, but because this entire approach for solving this problem is a failure.
"I can see the ecosystem being multiple parts being itself an issue, but I can also see it as an advantage for those with the manpower and capabilities to leverage each part correctly together, "
You have to think hollistically for that, the problem is not the system being multiple parts, I can't context the idea of specialization.
The problem is that the system was made for other purpose and we got stuck in a local maxima.
That's why I said it is not great.
"To label the entire ecosystem a failure"
But it is a failure of being fit for purpose, because we didn't knew the purpose in 2004 when this insanity of using the Web document model to create applications, it all started with AJAX, and even earlier, with DHTML.
No, I don't need to give solid examples for all cases, I could do it, but then this will turn into a master's dissertation, and I have better things to do.
"At a bare minimum, it should be agreeable the React ecosystem is successful when used correctly"
I will agree with that, the system can be used succesful, albeit you have to either ignore the problems it creates, or live with them. I decided I had enough.
"freedom it gives"
There's more freedom in not being restricted by it actually, that's why I went back to "VannilaJS".
For such a long analysis, the article completely misses the point about Redux being an event store that implicitly leverages several good patterns that are seen in highly-available distributed systems.
it forces you to separate your write model (dispatch) from your read model (props). this is known as CQRS.
the reducers are actually referred to as projections in event sourcing. reducers/projections are just aggregate data accumulated from events and are considered part of the read model. the strategy of doing the reduction on each event allows you to present aggregated data in "realtime". the alternative would be to iterate through large sets per read or write. it's a per-event, "realtime" .reduce()
Redux + Thunks was terrible. But Redux + Sagas more closely represent event sourcing as you see it in distributed systems. with events thought of as streams (or aggregates in the Domain-driven Design world). Sagas are actually referred to in the event sourcing world as either process handlers or sagas and are responsible for complex stream handling.
event stores and blockchains use roughly the same structure conceptually: Append only File
except with blockchains, each new insertion is hashed with the previous hash in order to keep the log cryptographically verifiable and immutable-- and is needed for Byzantine Fault Tolerance (so each node can be verified very quickly as a good actor or not).
smart contracts -- the reader contracts have similar logic to reducers. write contracts all look like event dispatch.
consensus algorithms do exist outside of blockchain tech. and with data structures that are considered high availability (they cluster and replicate), you will see that they are all AoF or KV stores or a combination of both.
most databases under the hood actually have a very similar event sourcing structure referred to as WAL -- write-ahead logging.
the typical event payload is the same in all these systems {eventType, metadata}
Just take a look at Redux w/ Sagas, Greg Young's Event Store, any blockchain, and even Kafka. They are all fundamentally similar. And since they all source from a log file, they can all replay state.
Also, Redux is an opt-in and can be as specific as you want based on how you tag your events. Your entire app doesn't need to respond if you slice your state right. And the structure itself has proven to be efficient. Event sourcing systems can handle millions of concurrent streams. Sure, people might argue there are optimizations in infrastructure that are NOT in Redux, but the structure itself is pretty ideal for streams if you write good handlers and partition your aggregates in a sensible way. The people who wrote Redux Saga totally see it.
I'm actually surprised that Abramov doesn't know the source inspiration for the thing that made him famous in the community.
There's two parts to this article, Rodolfo. Be patient.
Excuse me? I'm sure you think I'm rude, but this article is unnecessarily long and didn't touch on much other than how confused people are (which surprisingly include people on the React team). My 1 comment has far more insight into the implementation than your entire article. Why don't you just take the criticism? It's obvious you're not crystal clear on this either and you went down a rabbit hole and will probably end up using a lot of the insight I just gave you (and I'm sure what other commenters have given you).
In fact the only reason I signed up was to comment at how bad the article is because I felt that it was prolific enough that it does people who want insight a disservice. Will you really make an awesome part 2 that isn't an inane, self-serving, ambiguous CS analysis of a practical pattern that will help clarify it for people and give them insight on when and how to use it? I really hope so because you have a great opportunity here to teach.
Hint: Immutability and finite state machines will probably confuse people more. Save that for part 3 or a deep dive on structures. Event sourcing is a pretty misunderstood pattern and sub patterns have emerged.
Redux is definitely not used for event sourcing, nor is event sourcing applicable for most of the applications Redux is used for.
Thanks for your comments, though!
You don't understand the pattern. Don't mislead an entire group of people.
You are storing events. There is a literal "store" prop passed into Provider. Therefore it is a "source" of events and not just ephemeral event-driven programming like you see in video games.
It is event sourcing.
You just proved to me that you don't understand event sourcing.
You don't store state with ES (except for snapshots etc). You store events. That is impractical for most apps.
I think you're just Googling event sourcing and trying to disprove me. Because you can most definitely store "state" with the pattern and it is done quite often. They're called projections. State storage is your read model, actually. For people who just need to read from streams, they don't need projections in-line with the event store. They do aggregations in another service. We just see the entire flow in Redux.
I mean you're kind of proving to me you haven't thought of this pattern much and you're about to try and teach people.
Go write a post then.
I am working on it.
But I'm not prolific and it is something I know that gets even the event sourcing community heated. Only reason why I chimed in is because you are prolific and I think you have a chance at giving a ton of people an aha moment that isn't in the wrong direction.
If you want to research for your follow up blog post, check out how similar these are to Redux and Redux sagas:
In particular, the event handlers:
github.com/commanded/commanded
Projections:
eventstore.org/docs/projections/in...
Discussion on sagas and handlers:
stackoverflow.com/questions/342846...
Can't deny how close the patterns match up. I mean the EventStore Projection API is basically what we know as a reducer. Event Store was made back in 2012.
But I will say that in infrastructure, people misunderstand and do all sorts of weird event sourcing patterns. If you had a good evented model in your infrastructure that represented state pretty well, complex stream transforms in Redux (and Redux itself) starts to disappear because the work has been done for the UI.
The only reason why we needed Redux in the first place was because we needed to model some type of complex state representation that the backend didn't give us. If the state is computed in the infrastructure and the UI as a service just receives it, where is the role in Redux other than acting as a global store or a pubsub? And even then, with GraphQL and Apollo + cache, Redux's complexity and importance starts to disappear.
Time to eat some much needed humble pie Rodolfo; you want to push your point of view so hard...
Redux is absolutely not event sourcing (I used both extensively)
The only thing in common is that there are "events", something extremely common and found in pretty much all codebases (and the dreaded DOM)
No frontend keep all the events from last week around for the logged user, it's slow enough as it is. If you don't keep all the events, then it's not sourcing, just event driven (like almost all UI paradigms from the last decades)
Trying to find similarities between very different architectures might help you learn new things, but there's not point pushing this tool down people's throat.
You wanna mansplain me more? Maybe you can explain how React works, too. Insult me because you think I'm using analogies as learning tools (although that is absolutely completely valid)?
Actually, I think you should eat some humble pie. I am bringing up a legitimate comparison and I have used both extensively as well.
Your argument doesn't make sense and is just a nitpick. You in fact DO keep all the events in context of the lifetime of your frontend. If you do not choose to persist them after the browser closes or refreshes, that's your choice. But the implementation itself matches event sourcing. You are using an event store, running projections off of them, and using those projections in your read model. You can even replay your events to regenerate your state. If it looks like a duck, talks like a duck?
Also, there are event streams that terminate in event sourcing, you know. You could think of a browser session as a unique event stream with a prefix "ui-1", "ui-2", etc.
But again, the main point is: in context of the UI's session lifetime, the events are not ephemeral and you are indeed keeping all the events.
As for speed, the implementation of Redux might be slow, but updating state per event is one of the reasons why people use event sourcing in infrastructure-- to do relatively complex and flexible compute "realtime" without having to iterate through large sets.
The "slowness" is due to the growing event store which is a problem in ALL event sourcing implementations. In infrastructure, you typically cold storage "really old events".
If you don't have the flexibility to see the valid comparison, I doubt you've used both as much as you say or you haven't really thought about it. The similarities go FAR beyond analogies and learning helpers.
In fact, if you took the event storage itself and persisted it, you now have a way to recover complete session state through replay. Though UI state isn't so complex or critical that you can't just persist the state itself versus doing a replay from the events. It would be useful if you wanted to track clickstream and state transitions beyond 1 session. Which some people might want to do.
As I said in a previous post, other than purely UI state, we needed Redux to model complex domain state that the infrastructure should have already given us. If the we already had that state, we wouldn't need to mirror what many scaled infrastructures use for our frontend. The reason you remove Redux is you already HAVE the state you need. But I think the pattern itself is a pretty important entry drug (although contentious because people do it in so many different ways) concept that could help launch people who know React and have only worked in React to understand how distributed systems manage state and data.
React is already very good training in SOLID. And as a result, a lot of the component patterns you see fundamentally match up closely with supervision tree models in distributed systems. Now THIS next statement is me using an analogy, but if you diagram a supervision tree and don't label it and just describe it in terms of supervision tree definitions, then tell people to label it in terms of React Components, they will surprisingly know how to.
@Rodolfo you’re my hero 🤙🏼
" If the state is computed in the infrastructure and the UI as a service just receives it, where is the role in Redux other than acting as a global store or a pubsub? And even then, with GraphQL and Apollo + cache, Redux's complexity and importance starts to disappear. "
Well, you are not wrong about this.
That has nothing to do with the fact of things being event store or not, its just that you don't need to store collateral state if you model the data for the consumer perspective, instead of the producer, this should be obvious.
When all you have is an "EventStore", everything looks like events.
The only problem with Eventstore and CQRS is that FSM are actually the generalization, not the other way around.
With the David actually got right, and you didn't, Rodolfo.
Saying redux is like event sourcing, is like saying redux is like a state machine. It's definitely possible to model redux like a state machine, but there's no inherent constraint in redux that promotes this.
The only constraint redux has is that reducers are pure...there's no constraint on actions. Actions can be modeled as commands or events.
Then again, if they're events, where are the commands? And if they're commands, where are the events?
Also where is CQRS? Aggregates?
"Event-driven" does not mean "event sourcing".
After studying more about virtual finite state machines I would say that these patterns that are "good"(citation needed) for distributed systems are just specific instances of something more generic.
I mean, they do work, but they are not as general as they could be, they are specific cases of a more general pattern.
I don't think the author missed "point about Redux being an event store", that point is actually irrelevant for understanding FSMs as they are the more general case, which was the entire point of the article, and I don't think the poster is wrong.
This is what's causing the clash, you are doing something like bottom-up, when the reasoning was top-down.
Clearly a very well thought out post, with some interesting points.
I'm struggling to agree with the main theme here, though; I can't understand why the fact that Redux doesn't explicitly follow the state machine pattern is a problem. Redux is a library for state management, right? Some types of state might be best represented as state machines, but is this the only way to represent state? Probably not. And I'm guessing that state machines aren't always the best way to manage state, either.
State machines are one type of abstraction which one might apply over state, and whatever-pattern-redux-should-be-labelled-with is another. Pick your poison. Yes, you can introduce bugs into an application with Redux, some of which might be prevented with a state machine pattern, but I imagine the reverse is true, also.
What I personally like most about Redux is it forces state into a subset of the language which is easy to manage. Pure functions and values are simple, and easy to reason about.
The main problems I see with Redux are:
Regarding 1., I don't see why Redux should need to contain any async logic at all. Put it elsewhere. Regarding 2., there are better ways to do this and it's not really the point of using Redux at all.
This is the main problem I'm trying to highlight. Fragmented logic because of an incomplete pattern.
Sure - but if Redux took a more holistic approach, and went so far as to implement state machines, then it'd be a different library. There are options out there which do give you state machines in javascript, for instance XState - which you develop (as an aside, I think it'd be better if you state quite explicitly that you're the developer of that library, for the sake of full disclosure).
In the example you give of Reddit's loading state, we might equally view whether or not there were no results as an item of derived state. It can be derived from two items of required state; the set of results, and the loading state of the application. Something like:
Not saying this is a better approach, just a different one, and a different solution to the same problem. Different perspectives will yield different results and it's hard to give very clear guidance as to which is better in all situations, and I think that calling Redux half a pattern is a little misleading. It's half of the pattern you decided to compare it against, or, it isn't a pattern which could be called state machines.
I didn't want to do this because I just wanted to talk about Redux and state machines, instead of comparing/promoting my library. I didn't think that would be fair.
Of course! You can definitely write your logic like this, without explicit state machines, and cover all the edge cases, and have it work just as well as an explicit state machine approach. It just becomes much harder to "reason about" and understand how the app flows from a higher level. But if you don't care about that (and you just want to ship stuff the way you know best), then that's fine.
No, it's half of a pattern because it provides lightly opinionated "scaffolding" (atomic global state, reducers are pure functions with no constraints except for purity, side-effects via middleware somehow) but the other half of the pattern is what the user needs to implement themselves. They need to "fill in the blanks" and that leads to them constantly half-inventing state machines, whether they know it or not.
Like I said, the comparison against state machines isn't an arbitrary choice. Our UIs are state machines.
I do still think it's relevant to mention that you've developed that library. Not to say that it makes you biased; it makes you informed on state machines and the strengths/weaknesses of that pattern.
Anything that's half a pattern could be considered a whole of some other pattern. The boundaries between patterns are completely arbitrary and we can draw the lines where we like.
I wholly disagree that using derived state is less reasonable than state machines. The two aren't mutually exclusive in any case, but both patterns have their supporters and theory, both patterns are likely equally able to be reasoned about. Indeed both are likely to be little more than constraints which can be applied to state management, and it's quite possible that both make an application's state a more reasonable proposition.
But it is! The proof is that derived state can determine "what state are we in", and finite state machines can also determine "now that we're in this state, what can happen next?" That part is valuable information, and very hard to determine without some sort of abstract model.
Indeed you're correct; though to reiterate what I said in the last reply, both are restrictions we can apply onto state which are likely to make it more reasonable. I don't think they're anything like mutually exclusive.
I had meant for derived state to be little more of an example of a different pattern which you could measure Redux and/or state machines against. My point was intended to be that these types of comparisons depend on the reference points of those comparisons; Redux is 100% of one pattern and 50% of another, and 0-99% of yet more patterns.
Interesting post and made me clarify my thoughts about Redux. I would argue that actually Redux is trying to be more than it should and this is a big source of confusion to people trying to learn it.
For example, whether you use a state machine or not should have nothing to do with the state itself. Why do you need a reducer? Just have your own state machine logic and update the state using events from the state machine.
Redux should just really be about updating state with events, reducers muddy the waters. I think this muddiness comes from its origins of use with react and specifically the idea of uni-directional data flow. Practically, state management should be agnostic of uni-directional data flow, but this is entrenched in Redux design.
Middleware and async actions also lead people down the wrong path. Mostly this is just code you should have in plain code you write. Eg, loading data should be outside of state management. Load data, update state. Don't complicated with middleware.
I prefer libraries that do one thing well. A state management system should just be about events, state and subscribing to changes. Redux has confusing concepts like reducers, async actions and middleware, not to mention actions are events. It would be much better split into two parts - state management and state machine, although arguable whether you need a library for the second part.
I would also argue that the loading state of an api request does not belong in the global state. This is local to a single component, putting local un-shared state in the global state makes things more complicated.
Completely agree. I'll cover this more in the second article. Nothing belongs to global state because nothing truly is global state.
That's my main gripe with redux.
I think no more than 20-30% of state should be global: logged user and its organization, resource cache and that's pretty much it.
It's insane to have to go through all the selectors for some fairly local piece of state changed.
Partial update of state in immutable pure functional applications is pretty much open research.
We actually need machinery to deal with it, you can't just update the global state every time.
You just happen to now have the same problem again, instead of React and DOM, you reintroduced it with your entire application global state in place of the DOM, and now need some form of React to deal with it.
Perhaps the entire perspective was wrong.
Virtual Finite State machines is what I'm looking into, they contain the mutable state, the are pure in the way that state+event change state, and they allow for side effects, because without side-effects:
"Haskell is a completely useless language because in the end a program with no effects, there's no point in running it, is a useless program.
And you know, you have this black box, and you press go, and it gets hot, but there's no output.
So why did you run the program" - Simon Peyton Jones
youtube.com/watch?v=iSmkqocn0oQ
The low-level API is fine... that it enforces a single atomic global store is where unnecessary complexity grows.
More on that in part 2!
I inherited a react/redux app here at work, and I found the organizational structure really hard to follow due to how everything was coupled. It had code smells all over it. The ideals were good, but the implementation left a lot to be desired.
Code is for people to read, so if patterns are confusing I disregard them and ignore "the experts". Amusingly the best and most understandable software pattern i've happened on in my career has been a mix of all these patterns.
C# has part of the answer with extension methods. You define data inside classes, and then create transformers (extension methods) between types (basically your entire codebase is a giant LINQ library). If the data being operated on is immutable, and you undo all the mistakes the OO architects created by getting rid of nulls, using discriminating unions, not using exceptions (or wrapping them at the lowest possible levels) and much more (fsharpforfunandprofit.com/fppatterns/) you end up with really modular code.
For the UI you use a hydrid model, where you make your datastructures implicitly serializable (as lispy as possible) with react-ish UI generator functions to render atoms of UI (again you end up with a LINQ library of functions to build UI very similar to react but without needing the jsx syntax). If you memoize the important bits, then your UI doesnt get recreated over and over. I then use a global event pipe for all the events. It all works great.
The only problem is, I've met precisely zero people that program this way, by stealing a bunch of patterns and mashing them together. FP abstractions sometimes get so into the weeds, that it's hard to read, so you have to strike a balance between all these aspects, by continuous refactoring and no upfront design. This is obviously really hard with more than one developer as the bad habits everybody has breaks all this stuff in a myriad of ways.
Rich Hickey has a talk where he states "simple is hard". That's what I do. I make everything so stupidly simple, it's modular as hell. Then I go to work, and get inundated with OO spaghetti code, and it saps the fun right outta life...
Do you think that Redux falls into the bracket of complex and easy, rather than difficult and simple? (to use the categories defined in Rich Hickey's talk)
For me, it falls very much within the bucket 'simple and difficult'. It has a single representation of state, which is a value. State can only be updated by reducers, which are pure functions. Reducers can only be triggered by actions, which are just data.
I'm struggling to see what part of that is complex, and I've spent a lot of time trying to think about whether or not Redux is simple or complex. How it gets integrated with React components might be a little more complicated on the surface, but again this is basically always done using higher order functions, which are an abstraction of sorts, but relatively straight forward.
Please do correct me where I'm wrong because I'm not trying to just pick holes in what you say here, I agree with the rest of your comment entirely, I'm just trying to objectively question where Redux falls along the simple-complex spectrum.
I found that Redux became a lot more appealing to me personally once I sat down and properly read the paper 'Out of the Tar Pit'. I like the idea of silo-ing state and restricting its manipulation to a subset of a language.
Hmm good question. I think the conceptual model is definitely simple, but as you say, since it's only a piece of the pattern, many devs opt for obtuse boilerplate (ignoring DRY), rather than buckling down to build higher level abstractions to achieve KISS and make code easy to reason about and navigate/digest. They are afraid to write utility code, or code generators (because it's perceived as hard) to essentially be as lazy as possible and write the least code. So they are lazy and wasteful instead and accrue tech debt and increase illegibility. Most devs don't write their code for humans to read with the simplest conceptual model. I love plain functions. Simple data, single argument, no side effects, simple data out. Most code should literally look like an excel spreadsheet. Classes are table rows, lists of classes are tables, formulas are static mapping methods to new tables. Unidirectional and simple.
My gut blames encapsulation as well. If one can't easily open an engine, and mod it, then we get these weird unreadable codebases like Dan Abramov was referring to. Everybody is solving a small slice of the problem and a lot of the problems internal to the project are hidden and not exposed to the community. I noticed this when I re-implemented entire utility libraries that I later noticed Microsoft had in their own reference source (often with the same signatures), but had not shared. So we get all these npm snapshots of pieces of the problem and people repeating the same solutions all over the place and they're hard to hook together and model cohesively. A big pet peeve of mine is that all code should be public and modifiable from anywhere. Swim not dive.
I suspect encapsulation doesn't scale well, because when one is 200 levels up from a problem (the npm packages folder depth is insane), perspective is lost. My gut says that if there was an R&D group that built an app from the ground up modeling everything like a physicist would, be it temporal logic (reactive extensions), asynchrony (promises), UI component models like react, the mapping of state to UI with redux, the low level rendering of graphics, etc. All those things would need to be in the same codebase to evolve together (as a modelling exercise). The fact that our dev environments don't come with hygienic macros or simple code generators built in (controversial I know), means we can't create DSLs to solve all our issues with minimal bloat when the libraries don't quite capture our requirements.
Unfortunately the difficult nature of code sharing and collaboration fragments this and separates concerns that perhaps shouldn't be separated. Not sure much of this diatribe helps anything though D:
PS: I am not a DRY extremist, I lazily model problems by how many times I see code repeating. Ad-hoc, on the fly with little forethought. Only when my bad code spidey sense tingles, do i step back and think about things. The repetition hints at missing conceptual models and informs what abstractions I create. This is really hard to do collaboratively however, as it requires training people to use a ruthless refactoring algorithm while they code. I still haven't had any luck teaching people how to do this at scale. I've never really taught people anything en-masse, so i'm sitting here on my high horse when perhaps I should be trying to convert a few people at work and see what comes of it. :)
Redux is simple like assembly, which is not good.
Actually, patterns are an anti-pattern that demonstrate clear as the day that the language is failing at modelling what should be modelling clearly and in a syntactical and understood way.
As someone who studies programming language development, every time I see pattern, I see a broken programming language or library.
Now knowing that, you are never going to be seeing patterns the same way for your entire life.
We shouldn't need to have patterns, we should be able to express an idea directly.
English doesn't have patterns, does it ? it has syntax, and semantics, but no pattern (well, it has some, but mostly they are legacy, we just ignore them anyway)
" I've met precisely zero people that program this way. "
Meet me, you have 1 person now.
Come over to F#+Fable, we have more persons, and we are trying to do a thing here.
Thanks for this, David. I know these has been some shade cast in some of the comments (yOu ArE mIsSiNg ThE pOiNt), but I love this thinking outside the box and saying "redux can be this", and I love using finite state machines in other code basis, so there's a good chance I am going to march to the beat you are banging out.
I'm very much looking forward to your next post.
One question, though: I've always daydreamed about working in Orlando. Do you work remotely from there or do you "go to work" in an office there? I didn't think there were many dev jobs down there.
Great post! I went from "why is he calling actions events" to "why weren't actions called events in the first place??" I love the idea of using a
statusenum instead of a combination ofloading,error,data, etc.One thing... Unless I misunderstand what you're saying about canceling an async thunk request, you can just ignore the success action if the fetch has been canceled:
Here's a working codesandbox: codesandbox.io/s/redux-thunk-cance...
Looking forward to part 2!
Very well written. I’ve just come up the very same idea that redux middleware as state transition functions and shared it to my colleagues then I came to this articles.
Keep up your good works.
P/S: On the other hand, I also think that a middleware is really a controller which is missed in MVC model. A very limited controller which is responsible for UI triggered changes and API triggered changes
Okay, please keep using Redux however you'd like, thank you for your comments
I do find it fascinating how there are very drastic responses to this post so far I must say. Some people agree that Redux is inherently "wrong", and other argue its inherently "right".
I'm like many, waiting for part two :D
Not sure about you, but in the types of apps that I write, I can not see any advantages of having a "single source of truth". In general, the complexity is always either at the page level, not the app level (the app is essentially just a router with some small global data, and an entity cache), or within some sub components on each page.
So I don't need a "single state atom" for anything. What I do need is to manage the state transitions of those complex components. Redux does nothing for me here. XState does tons.
I suspect most people are using redux as a global entity cache, or worse...what you do to store "global" data.
They talk about Single root of "thruth", but the elephant in the room is the insidious internal mutable state on every React component that's not pure and/or has hooks.
Wasn't the entire point of modelling state/events/transitions knowing when to do actions, like rendering the screen.
So if there was a proper way of doing virtual finite state machines, you wouldn't need react to reconcile the three, you could just go and change the DOM directly with impunity.
The DOM would be the single source of truth for the visual representation of the applications, and it wouldn't be slow because you wouldn't be randomly poking and mutating it without reason.
It would actually be faster, as it would only change precisely in "reaction" to state+events, ironic for something called "React" that doesn't react to anything, it should be called Render.
This is great....very well-written and clear. I gave up on Redux early because of my confusion about async actions....having to install some kind of third-party middleware just didn't feel right. Looking forward to the second part.
Hello David,
I have been reading FSM related posts (including yours) with great interest since a few years now, but I've never taken the step to include this pattern in my projects. I feel like there's more boilerplate and that it's a bit overthinking for simple UI component.
Anyway: "[Transitions] can be represented using boxes and arrows" makes me wonder if you've ever thought about implementing FSM using category theory and algebraic data structures?
Nice article! The moment I saw Redux - I started thinking on building state machines. That resulted in this nice library: github.com/mocoding-software/redux...
It allows build finite state machines using Redux. Please check this out.
Interesting article, looking forward to the next part!
Have you seen gist.github.com/andymatuschak/d5f0...?
In github.com/spotify/mobius, which I interpret as similar to the ideas you're describing, we decided to not mandate modelling 'state' as an FSM, though we think that FSMs are sometimes the best model (cf github.com/spotify/mobius/wiki/Def...).
Hi Petter, yes I've read that a while ago, interesting implementation! Spotify's Mobius (thanks for showing it to me!) also looks interesting, I'll take a deeper look at it.
This complexity vaporates when we stick to the much simpler event, eventhandler pattern. Events emit data and subscribers handle the notification. Components are small and only do one thing. For example, a Person component works with a Person model and A person object. That's it. This component is responsible to get the person object, subscribe to the result in populate the person model. Later it may contain a FindPerson function which has a well defined input and output interface. The person component grows over time and may provide notifications at any time. Because events are loosely coupled, what happens next is of no concern to this component. Simple concept, used by every browser in the industry why not follow what already works? Event s have been around for 50 years. They are the original observer pattern.
Some comments may only be visible to logged-in visitors. Sign in to view all comments.