[ickelbawd]

I have some sympathy for this viewpoint. And I think Alex’s heart is in the right place—Nextjs is a dumpster fire and react server components are when react finally jumped the shark for me.

Maybe if we had HTML6 we wouldn’t be in this scenario. HTML5 was great but form building on the web (without JS) is a second-rate experience. And it’s even more miserable once JS is in the mix, but hey developers can provide a much better UX for end users than HTML and CSS alone could possibly provide.

Sorry Alex, but without JS the web would have died a decade ago as phones took over. It’s only JS that keeps us in the ring.

[Devasta]

HTML can never be improved, that is at the very foundation of its current form. We are 20 years into the takeover from the W3C and still can't even do a PUT request without JS, nevermind anything approaching even one tenth of the functionality of XForms.

The web of semantic documents died years ago, it's an application platform now and that means JS. HTML is naught but a payload carrier.

[forgetfreeman]

All true, but an error doesn't become a mistake until you refuse to correct it. Perhaps it is finally time to let the web-as-an-application-layer paradigm die?

[aragilar]

https://alexanderpetros.com/triptych/ will hopefully address that.

[Devasta]

I wish them luck, but the last time this was attempted it was rejected as it apparently doesn't make sense that you would even want to do it in the first place.

https://www.w3.org/Bugs/Public/show_bug.cgi?id=10671#c16

[aragilar]

https://github.com/whatwg/html/issues/3577#issuecomment-2329... and the links inside appear to imply there's interest?

[pjmlp]

I was really disappointed that XHTML effort was killed by the HTML5 folks, we could be so much better, but it is as it is.

[Devasta]

The founding members of the WHATWG really are the Thomas Midgley's of the software world. The world will spend the next century of software development trying to repair their screwups.

[pjmlp]

Ironically that work is now being made obsolete with apps regaining the role of native apps, or the plugins revenge via WebGL/WebGPU/WebAssembly.

[amadeuspagel]

The author actually links to the NextJS page for server components as "a tacit acknowledgement that their shit stinks". I guess this is how you acknowledge progress if you want to write an article in the tone that is expected on HN and mastodon now.

[andrewflnr]

He's not opposed to JS entirely (he specifically praises a handful of JS-based tech like ServiceWorkers), more so the all-JS single page type of app. To be fair he definitely could have been more clear about what he's talking about and what lines he's drawing.

[nchmy]

He has dozens of other exceptional essays on his site, and plenty on social media and conference talks. Check them out.

[ec109685]

All the desktop apps he cited that are a threat to native are SPA’s so it doesn’t seem like he hates them everywhere.

I agree though it’s very odd that desktop web apps are serviceable, while they universally are horrible on mobile.

[andrewflnr]

Because their performance falls below acceptable levels specifically on (relatively slower) mobile devices. It takes a looong time for him to get around to it, but performance (specifically UI latency) is the crux of the whole article, as I read it at least.

[marcus_holmes]

Yeah, I read this (and some of his linked articles) as pointing out that Apple are specifically killing browser performance and features so that PWAs cannot compete with apps on iOS.

Specifically because they can't claim 30% of revenue from web apps.

[JimDabell]

It’s wishful thinking. Apple is being used as a scapegoat for the web’s failings here. PWAs would be so successful if it weren’t for Apple! is the easy but wrong answer.

It’s possible to deploy a PWA to Google Play so that it’s installed on Android just like a native Android app. If PWAs were as good as native apps, why would anybody build web + iOS + Android instead of just building web + iOS? Even if opening a PWA made an iPhone explode into flames, it would still be worth deploying the PWA to Google Play to eliminate Android development effort. But plenty of native Android apps are still being built every day. That’s not because of Apple. That’s because users prefer native.

I am a huge fan of the architecture of the WWW. I think it’s the most successful API and platform in history for a reason. But there are a lot of reasons why people genuinely prefer native apps, and it doesn’t just boil down to mean old Apple refusing to support things or users being brainwashed. People still choose native over web when Apple is not a factor at all. Every single native Android app is a testament to that. So when people also choose native over web on iOS, why should we think that this is down to Apple and not for the same reasons they choose native over web on Android?

It doesn’t help that the web platform veered off into web components either. When front-end developers were jumping on board React, Vue, Angular, etc., architecture astronauts working on the web platform spat out an unholy mess full of footguns. I prefer to work as close to the web platform as possible, so every so often I try again with web components before remembering how painful they are.

The people working on web components should take a long hard look at why front-end developers keep choosing frameworks that either eschew web components altogether or abstract them so far away the user doesn’t have to think about them. Telling quote from the Svelte 5 announcement last week:

> Equally, component composition is more awkward in Svelte 4 than it should be […] This is because in 2019 it seemed likely that web components would become the primary distribution mechanism for components, and we wanted to align with the platform. This was a mistake.

— https://svelte.dev/blog/svelte-5-is-alive

I really hope the web does better, but in order for it to do so we need to be more honest about its failings. If you are a front-end developer and you feel optimistic about the future of web development, I invite you to try writing something reasonably complex using web components instead of your normal framework.

[pjmlp]

Not only Google, Microsoft is also a big pusher of PWAs, or used to be.

[photonthug]

> Specifically because they can't claim 30% of revenue from web apps.

This feels true, but it’s just one horn of the dilemma. The other is where websites wreck their own UX intentionally to herd users towards apps, presumably because pushing ads isn’t as lucrative as grabbing location data, contact lists, or whatever else can be legally exfiltrated and sold.

It’s weird to frame this problem as if it were all the fault of whatever trends in js development or browser development (even though I agree those trends often seem foolish and messy, at least to the outsider). The whole situation the author bemoans just looks like another symptom of surveillance capitalism, enshittification, etc, call it what you will. Even with a perfect web, users will still get pushed towards apps if apps are the best way to exploit the user. There’s no other calculus of platform aesthetics or excellence going on here

[marcus_holmes]

Agree. I think it's entirely possible that both Apple and other businesses are pushing people to apps (by destroying their website productivity), for different reasons.

[aragilar]

I doubt he's opposed to JS, given the number of new JS APIs he pushed Chrome to implement as part of Project Fugu, he seems more opposed to specific frameworks.

[DeathArrow]

>Sorry Alex, but without JS the web would have died a decade ago as phones took over. It’s only JS that keeps us in the ring.

Not really. Web would still be there but for websites not web apps. The web was designed for hypertext, a medium that is just supposed to deliver information not build apps upon.