Hacker News Reader

Comment by langitbiru | Hacker News Reader

So I just read about "Choosing Boring Technology" before coming into this thread. I guess React.js is the boring technology (already 5-6 years) one and Svelte is the new and exiting one! :)

View on HN

I’ve been holding off from React for a long time, but also after reading the “boring technology”, I came to realize React is in fact - here to stay.

It'll be gone in 5 years. It's getting too big and too slow. Also, the react team managed to add hooks and useEffect in a way that makes it very easy to accidentally send your app into an infinite loop and not know why.

> It'll be gone in 5 years. It's getting too big and too slow

Funny, that's what everyone said 5 years ago. I'll bet on React to be here for another decade, no problem - there's just way too much momentum. In a way, it's Rails all over again (that's a positive).

Rails burned out in under 10 years. The reason I say React will be gone in another 5 years is because it's already been popular for 5 and the muttering has started. Do you remember Adobe Flex? React is following a shockingly similar path and timeline.

Still writing Rails today, nothing seems to come close, except maybe Phoenix sortof

> Rails burned out in under 10 years

It may not be the glory days, but Rail's is still going strong 13 years later. React, however, has reached a popularity many times larger than RoR ever did.

https://trends.google.com/trends/explore?date=all&geo=US&q=r...

There has never been a project (I can think of), other than the Linux kernel, with the resources that companies and developers have put behind React.

> Do you remember Adobe Flex

Flex never came close to either RoR or React popularity, and was propelled mostly by the dying wave of ActionScript programmers from the aughts after cell phones killed the Flash game market.

So is Flex[0]. What's your point? It's not popular anymore and hundreds if not thousands of code bases are being converted from both of these technologies into something else.

[0]: https://trends.google.com/trends/explore?date=all&geo=US&q=r...

> What's your point?

Seriously? Adobe Flex was a closed-source and proprietary commercial offering capitalizing on the prior decade's popularity of Flash.

React is open source with a vast and robust ecosystem of even more open source to go along with it. It couldn't be more night & day.

> Adobe Flex was a closed-source and proprietary

That confirms it: you aren't old enough to remember Flex. It was open source.

Not until version 3 in 2008, after which it became Apache Flex. Maybe you should try to get both your facts and attitude straight.

Please don't get involved in flamewars on HN. Personal attacks in particular will get you banned here, so please don't do those.

https://news.ycombinator.com/newsguidelines.html

Please don't get involved in flamewars on HN.

https://news.ycombinator.com/newsguidelines.html

Sorry. That was not my intent.

> React, however, has reached a popularity many times larger than RoR ever did.

> https://trends.google.com/trends/explore?date=all&geo=US&q=r...

I'm not sure this chart is very good evidence of this claim. There are so many other things named "react" -- Nike React shoes, Kids React videos, reactionary politics, and so on -- that it's a bit like claiming Trump.js is the most popular JavaScript library in the world because of this chart: https://trends.google.com/trends/explore?date=all&geo=US&q=r...

Hooks definitely need to be deeply learned, and has one of the steeper learning curves in React. But at its core, its solving a hard problem (lifecycle side-effects), in a way that is more functional, and produces more maintainable code. Thats a fair trade-off imho. And honestly, once you understand them, the infinite loops are deterministic and understandable.

If you want a component-based view library that is written in js, and you are happy to take on the task of assembling a build-system, styling system, routing and of course state-management, I can't see much of an architectural improvement you could make over React.

If you want batteries included - go with Vue.

If you want the Ruby/spring opinion and batteries included go with Ember.

Svelte looks kinda cool. But it seems to sit deep in the browser spec, and in my NAIVE opinion, it doesn't seem hold the component abstraction as something even valuable. It just wants you to script away. Which, is just movement within a trade-off space - if it suits you go for it.

The component abstraction is primary with Svelte. Each file is a component and only one component. This means the only ceremony you need to create a component is to create a new file. It's a perfect solution in my opinion.

I went through the svelte examples and I was exited until I reached the If block example:

    {#if user.loggedIn}
      <button on:click={toggle}>
        Log out
      </button>
    {/if}

To me such template language is a huge step back from JSX. JSX is JavaScript, thus you get the full power of the language (JS/TS) with the full support from your editor (type checking, auto-completion, etc.).

I hate template languages myself. But in Svelte’s defence, their template language in version 3 seems to be well-designed and streamlined.

It doesn’t have the ad-hoc nature of “oh crap we forgot to add if-then-else” of Angular or “let’s create half-a-dozen incompatible javascript-like DSLs” of Vue.

Yeah I think I basically agree. Being able to simply write idiomatic javascript to compose elements is just so awesome.

Though I would assume Editor support of Svelte can/is being built.

My daytime job isn't primarily programming anymore, and frontend development has never even been my job. I used to do embedded and backend development, and am now in more of a ops/architect/automation consultant (freelance). I still need to code on a regular basis, and Go and Python are my go-to languages these days.

On a regular basis, I encounter the need for a GUI for some API or backend I write, so I tried to get into a more modern Javascript eco system. I have created websites and even full-on forum software in a long forgotten past, so I have some basic old-school webdev and basic Javascript knowledge (from back when directly manipulating the DOM was still the way to go). So take this as a bit of an outsider look on things, who is confronted with full-time Javascript developers on a daily basis.

When I started looking into react, vue, ... it always overwhelmed me. Whatever I started doing, according to someone, it's bad practice and should just do Y instead of X. Nobody seems to agree on best practices, and there seems to be a whole lot of 'magic' going on in all those frameworks, and knowing where to start is an absolute mess. It's confusing as hell.

Now I quickly went through the tutorial and examples on the Svelte site and it felt pretty simple, I can reason with how it works behind the scenes, and can read the generated Javascript to verify that that's really what's going on. It makes the whole thing seem more manageable in my head.

As far as "the new exciting one" goes - what appeals most to me personally is its simplicity. One of the reasons Go also appealed to me. It was exciting for a while, but ended up being the boring 'it just works' solution which completely replaced C and C++ in my life (that's not a jab at these languages, they have their uses, I just don't need them anymore)

I've been doing web development professionally for two decades but I feel exactly the same way. It couldn't be any simpler!

To be clear, this is Svelte version 3, and the first commit was Nov 15, 2016. So the framework is technically 3+ years old.

That said, I don't think boring is about age, but about real-world exposure

To be more in line with the article they're referring, its more about "known unknows" and "unknown unknowns"

Reply on news.ycombinator.com