I find this a totally bizarre complaint. I've spent the past few months working on Svelte stuff and I've seen people on HN make this same complaint about Svelte's templating language with {#if} and {#each}. Who cares? What is so wrong, exactly, with "reinventing a concept that's already in the language"? It does not make code any harder to understand or to write, and it does not harm performance (in this case, quite the opposite).

I would much rather have a reactivity model where I plug in completely standard concepts and patterns (a for loop) than one where I have to deal with a bunch of framework-specific, complicated ones (hooks). That Solid's reactivity primitives are familiar is an advantage, not a disadvantage.

So you end up with a secondary full featured language usually with worse IDE support, worse error messages, more surprising issues, etc. You need to understand the scoping mecanisms and if things go wrong hope there is a debug tool available.

And in the end those templating languages do not prevent you from mixing UI responsibilities from the rest of your code.

If you want a reactive model you can have one. I personally prefer explicit messages like calling setState.

The idea that this type of thing should be happening anywhere near the view rendering loop is the exact reason I've not had a great time picking up React codebases.

By the time you're rendering data into markup, the data should be in the exact state you need it. No further filtering or data mangling or sorting. That type of data manipulation should happen at the point of data change and then it shouldn't happen again until the data changes again.

The simplistic templating languages in Vue/Svelte/Alpine/whatever-comes-next force you to pull your data manipulation back to somewhere more appropriate, with Vue even throwing a warning if you try to filter within v-for construct.

Because React is JS, people are let loose to do wildly inefficient operations and do them over and over and over whenever _anything_ in that component changes.

It’s nice to have a concept “ground truth” in data and props and then computeds that sort of tie it all together.

You can do same/similar with React hooks, just not as clean or obvious.

I do, however, think that working with React changes your mental model somewhat, and when I'm working with React I catch myself doing a lot more data wrangling close or in the rendering loop than I would in any other modern framework. Certainly since class components have fallen out of favour, you're working with a function designed to be run hundreds of times, while Vue and Svelte both provide clear patterns to deal with data at the point of change, then separately deal with updating the display of that data as required.

It takes using something like MobX to really push a React codebase to a data-driven model and that means many inexperienced developers fall into the common pitfalls far more easily than if they're using an alternative framework imo.

Gilad Bracha calls these "shadow worlds" https://gbracha.blogspot.com/2014/09/a-domain-of-shadows.htm...

Having templating DSLs in other frameworks isn't a deal breaker, but it's a pro of React that I appreciate.

Personally I've never found a problem I wasn't able to solve in Svelte.

Well, what can you tell me about TypeScript type inference for this custom DSL?

https://www.solidjs.com/docs/latest/api#%3Cfor%3E

Nothing, inherently. Just like there's nothing inherently wrong with having extremely clear and simple rules for how to use hooks, and lint rules to identify when you're not following those rules. Nothing inherently wrong with either, some people just have strong distaste for one or the other.

Once you deal with larger amounts of data and need virtualised rather than fully-materialised lists, you start using different things in React as well. The fact of the matter is that if you care about performance at all, the simple ways are just insufficient, and the native language constructs were designed for procedural programming, not reactive interface rendering, which requires fundamentally incompatible semantics. It’s not even fair to claim that React uses regular JavaScript idioms—VDOM, hooks, the entire shebang is all about eschewing regular JavaScript idioms because they don’t scale. (OK, so there’s also the matter of transient state like scroll positions, element focus, and form field values; it’s not fair to say that React does all these things purely for performance’s sake, as the naive immediate mode approach would also break such functionality.)

I want to push back somewhat on this practice. Our computers are fast enough now, and the browser implementations optimized enough, that they should be able to handle thousands of materialized list items without breaking a sweat. Sometimes you really need virtualization, e.g. if the underlying data source has millions of records. But if the data can be fully materialized, then the implementation is simpler, and the user can take advantage of things like find in page. Virtualization is a convenient way to avoid the inefficiency of unoptimized VDOM-based rendering (e.g. with React, and yes, I know there are other optimizations available in React), but fine-grained updating (as in Solid) is even better.

It depends a little on the complexity of the rendering for each item, and where you are fetching the data from, but when you’re into thousands of records you’re very likely to need at least some partial rendering. Suppose each record’s data is one kilobyte (I’ve seen far lower and far higher), then one thousand records is already one megabyte, which for many people will take multiple seconds to transfer, so you’ll still want to load the visible records before fetching more, or do streaming parsing of the records as they come in. And that’s ignoring the backend’s performance on fetching records, which must be taken into account too.

People are certainly often too eager to reach for virtualised lists, or worse still lazy loading without reserved scroll height, but even at a thousand records with simple rendering they’re still probably generally warranted—fast computers can cope with comparatively little visible difference, but on slower ones (especially older and cheaper phones) you’ll easily feel the difference. Memory usage can also be a concern for larger quantities of data and DOM (1000 × 100KB = 100MB).

I’m saying all this as one that scorns React and VDOM stuff in general as unnecessary performance overhead, and likes to use Svelte and plain JavaScript and things like that (or better still, to eschew JavaScript); and who worked on Fastmail’s webmail, which certainly uses such progressive-loading lists, on a precise-DOM-updates framework that cares significantly about runtime performance. React and its ilk are certainly particularly prone to using virtualised lists as a crutch to work around their shortcomings.

All that being said: yeah, I wish things like Discourse would stop doing aggressively lazy loading when there aren’t even several hundred comments in a thread. Render just the things on screen to begin with, if you must (though it’d be better to just send real HTML from the server and let the browser take care of all this, even if it complicates your JavaScript loading), but then load all the rest straight away so that I’m not penalised just because I’m on the other side of the world from the server, and let the browser handle in-page search.

Speaking of both Discourse and screen readers, before my stint at Microsoft, I wrote a Windows screen reader, which tried to detect client-side page navigation by watching for the URL (minus the fragment) to change. Discourse's infinite scrolling implementation broke this heuristic, because Discourse would use the history API to update the URL as the user scrolled. Not sure if I or they were in the wrong there.

Interesting that you feel that Discourse penalizes you for being far away from the origin server. When Discourse was new, one of the founders blogged about how their heavy use of client-side JavaScript made the application better for users far away from the origin server:

https://eviltrout.com/2013/01/06/turbolinks-and-the-prague-e...

Maybe the author had a point, but it sounds like Discourse still relies too much on frequent round trips.

Edit to add:

> a precise-DOM-updates framework that cares significantly about runtime performance

That's Overture, right? I wonder how it compares to Svelte and Solid.

For the UI part of it: what I would use would depend on my requirements (is it a list, is it a grid, how is it to be interacted with, &c.) and what was already in use (React, Svelte, plain JavaScript, other). I personally would often be inclined to implement it from scratch, because I’m typically not impressed with most library options (they have a tendency to be heavy, limited, and slower than they need to be) and am familiar with exactly what needs to go into it to make it as perfect as is possible (it’s not a particularly large amount of work, but it is fiddly in places and must be done correctly or it’ll be awful), but that’s not a course of action I would recommend for most developers.

My specific use case is a React application, and I have found it easier to implement a dedicated listener system, than try to fit things into component states.

I'm right there with you, but when React invents a whole markup language inside of JavaScript, it's not in much of a standing to make purity criticisms.

It's about as pure as you can get while having any sort of html-ish 'templating' whatsoever.

So, I'd say it's exactly in the right place to be making purity criticisms. They've taken the only approach that preserves the integrity of the code and doesn't involve build time magic.

    function MyComponent() {
        return For({
            each: [1,2,3,4],
            children: x => <div>{x}</div>,
        })
    )

But coming back to JSX, a custom function call syntax is OK and pure because it has a one to one mapping on line number?

I don't know if there's much point in discussing purity since it's badly defined and mostly in the eyes of the beholder, but it always smelled like one hacky syntactic sugar to me.

It has no intrinsic semantics, and maps pretty much directly to actual javascript (which you can write directly or use an alternative helper for — hyperscript being a common one).

This argument is silly because it inevitably becomes a pissing contest of who can be most like vanilla JavaScript. In that case, why use JSX? Just write hyperscript calls instead. Why use React Router/React Context helpers? Just wrap your components using vanilla function providers instead. Why use React Hooks, which inevitably look like magic to a JavaScript veteran because the library inherently hides away some global state? I hope you can see what I'm getting at here.

I had/have your bias, but from playing with it I found a couple things:

1) Like React, you can swap out the template feature for a function call (or subcomponent). e.g. instead of

  return (
    <button...>
    ... 
    </button>
    <For each={state.todos}>
    ... 
  );

  function displayTODOs<T>(todos: T[]): any {
    let arr: any[] = [];
    for(let [i, todo] of todos.entries()) {
      const { done, title } = todo;
      let elem = (/\* JSX \*/);
      arr.push(elem);

    }
    return arr;
  }
  ... 
  return (
    <button ...>
    </button>
    {displayTODOs(state.todos)}   
  );

What I like about the React monoculture is that it's one less thing I have to care about. I can focus on the other aspects of my programs, beyond turning JSON into HTML.

I haven't used SolidJS so I'm not going to put it on blast. However, hearing people compare its reactivity model to Knockout JS gives me the heebie-jeebies, because Knockout projects were horrific to reason about (and test) beyond a certain scale.

Isn't this optional? Can't Solid use regular JSX loops?

For reactive control flow to be performant, we have to control how elements are created. For example, with lists, a simple map is inefficient as it always maps the entire array.

This means helper functions.

I don’t see the big deal here. Error boundaries, suspense, context, very popular routing libraries all have used components to encapsulate functionality. That’s to say first party and third party functionality in the React ecosystem have gone down this path.

So, what you have here does not remove complexity from problems, it moves the complexity from one place to another.

So, when you have a simple task, and a straightforward framework, what you see is “it’s easy!”. …because when you use the complex framework you get a bunch of “solutions” to problems that don’t exist on your problem.

That’s why it appears overly complex.

…but for a complex problem, when all you have is a simple framework (like <For…>) you have to implement the complexity yourself, which makes you view the framework as feeble and under whelming.

So, you are just moving the complexity from one place to another; the question is, is the complexity of react really something most people need, or can a framework like solid solve the 90% of simple problems most developers have?

It’s hard to tell.

Most new frameworks excel at solving simple problems because it makes for cute demos.

Is solid any different?

That’s my question. How does it work at scale, for large complex projects? Is there a whole design system implemented in it? Who’s using it and for what?

The claim that it’s “not complex” doesn’t help.

All that means is there are probably a crap load of things it doesn’t include I’ll have to do myself.

There is no magic bullet that removes complexity from tasks.

React is a complex beast, and a nice clean framework to replace it would be welcome.

…but you have to approach this kind of discussion honestly.

Hello world examples are a dime a dozen.

From the moment I saw a post about immer.js, I was sold because it seemed like an obviously better solution for the vast majority of cases where I would otherwise grab Immutable.js, a library that I wanted to like but inevitably struggled against.

This.. isn't quite as revelatory. I'm not saying it's not all that you claim it is, it's just that from a glance, I don't see how this somewhat different approach addresses the problems I run into often with React in a major way (beyond the claimed performance boost).

We have found that continuing to use immutable.js Map and List but using plain JS objects, not Records is sort-of a sweat spot. But one needs to enforce immutability with Flow/TypeScript read-only types and use the latest immutable.JS to make it work.

I wonder if I could have your opinion on my library, and why half the code, native performance is actually more complicated in the long-run. I don't know the answer at the moment. Documentation isn't complete... it is just web components. eg hello world just becomes a function like const component = hello('Andrew'); document.body.append(component); in the example on github. Anyway, the todo shows a more typical real world example I guess.

Actually, that is only half the reason. You can do whatever you want in my library (github.com/thebinarysearchtree/artwork) and it doesn't have any kind of virtual DOM or whatever Lit does, because you just create elements with JavaScript. The second reason, that everyone just assumes is the default, is that React has to use HTML-like templates and not just JavaScript.