News for hackers. Stuff that matters.
Re:Is Turbopack 10x Faster than Vite? by qudat on Tuesday November 01, 2022 @08:09AM (#33420097)
Vercel and the VC money they generated is probably one of the worst things to happen to FE OSS. I avoid next like the plague.

To make matters worse, now the react and next teams are actively colluding with each other, effectively allowing vercel to have a competitive advantage on feature development for react.

It might seem great now but this is the hairline crack that will destroy the react ecosystem.

I wouldn’t be surprised if vercel tries to buy react and its team.

Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Tuesday November 01, 2022 @08:40AM (#33420347)
Vercel is an evil company in disguise.

I’ve understood this and stopped using Next with all the enabled by default and intentionally difficult to disable metrics collection system they have in Next.

Having it enabled by default, without asking is already evil enough.

Refusing to have a setting you can easily commit into the repository to turn it off is next level.

They trick people with the “just set this env var” to turn it off, and rely on people doing it wrong, forgetting to do it in certain environments, etc. it’s intentionally complicated, when a frigging Boolean setting in next.config.is would be better. They refuse to do it. That’s what evil companies do.

Re:Is Turbopack 10x Faster than Vite? by Dork1234 on Tuesday November 01, 2022 @09:29AM (#33420905)
So which framework should you use? It is easy to say XYZ is bad, but what do you use instead, ASP.NET?
Re:Is Turbopack 10x Faster than Vite? by o_m on Tuesday November 01, 2022 @09:40AM (#33421084)
Laravel looks like the most mature fullstack framework. .NET and Java might be nicer on the backend, but they are a decade behind the competition on the frontend. The javascript community doesn't seem to want a fullstack CRUD framework, the consensus seem to be that every project should make its own ad hoc framework, like the PHP community did 10+ years ago. This resulted in a lot of spaghetti code, and javascript, even with their advanced libraries, isn't much better. The downside of not betting on javascript is that "everyone" knows javascript by now.
Re:Is Turbopack 10x Faster than Vite? by iKlsR on Tuesday November 01, 2022 @10:03AM (#33421390)
I chanced on https://adonisjs.com/ the other day. It's modeled to be somewhat similar in ecosystem, batteries included and some syntax/layout to Laravel. Haven't really tried it yet but the community seems to be very active, there is also https://adocasts.com/ for eg.
Re:Is Turbopack 10x Faster than Vite? by egeozcan on Tuesday November 01, 2022 @10:59AM (#33422114)
> The downside of not betting on javascript is that "everyone" knows javascript by now.

I'd say: everyone thinks they know javascript. Learning a new language is easy (especially with C-like syntax) but not the library. The language having nearly zero standard library makes it change under your feet on every different context. Deno started the initiative to supply what you get in the browser everywhere, but the browser being a GUI environment, the missing stuff need to come from somewhere, like the Deno.listen[0]. .NET is not much better with one million different runtimes with different platform support and capabilities.

[0]: https://deno.land/[email protected]?s=Deno.listen

Re:Is Turbopack 10x Faster than Vite? by CrazyPyroLinux on Tuesday November 01, 2022 @09:44AM (#33421147)
> ASP.NET?

I do - .NET (since 5) has been great. And you've got stuff like Blazor for the frontend if you want to avoid JS entirely.

Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Tuesday November 01, 2022 @01:40PM (#33424567)
First and foremost, I don't go SPA/React by default on every project, so that reduces a lot my need of the kind of framework I think you're thinking about.

Now, if I do need a lot of interactivity in the frontend, I use Inertia with Laravel, which gives me a real full stack, robust, battle tested, well integrated and very productive solution.

Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Tuesday November 01, 2022 @01:41PM (#33424580)
And Adonis is great if you want to stay in jsland
Re:Is Turbopack 10x Faster than Vite? by M4v3R on Tuesday November 01, 2022 @09:33AM (#33420967)
Can you provide a source for 'They trick people with the “just set this env var” to turn it off'? I'm reading their documentation about telemetry and it says something different, am I missing something?

[0] https://nextjs.org/telemetry

Re:Is Turbopack 10x Faster than Vite? by hu3 on Tuesday November 01, 2022 @11:54AM (#33422944)
The documentation is purposefully deceiving. It makes you believe, in good faith, that disabling telemetry is per project when in reality it is per machine.

I just tried it myself on a brand new Next project:

    $ npx create-next-app@latest
    
    ...

    $ npx next telemetry disable
    Next.js' telemetry collection is already disabled.
    Status: Disabled
    You have opted-out of Next.js' anonymous telemetry program.
    No data will be collected from your machine.
    Learn more: https://nextjs.org/telemetry
Note the part that says "from your machine".

An issue was crated about it but they just closed without responding: https://github.com/vercel/next.js/issues/31360

Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Tuesday November 01, 2022 @01:42PM (#33424601)
THIS
Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Tuesday November 01, 2022 @01:34PM (#33424472)
Sure.

1. Last paragraph of that link you posted

2. https://github.com/vercel/next.js/issues/8851

3.https://github.com/vercel/next.js/issues/10713

4. Many similar discussions to this one here on hackernews.

Re:Is Turbopack 10x Faster than Vite? by adoxyz on Tuesday November 01, 2022 @11:42AM (#33422732)
To disable Telemetry simply run `npx next telemetry status` or set

``` NEXT_TELEMETRY_DISABLED=1 ```

in your .env file. I mean it couldn't be any easier.

Re:Is Turbopack 10x Faster than Vite? by joshmanders on Tuesday November 01, 2022 @01:17PM (#33424194)
You seem to have missed this part.

> They trick people with the “just set this env var” to turn it off, and rely on people doing it wrong, forgetting to do it in certain environments, etc. it’s intentionally complicated, when a frigging Boolean setting in next.config.is would be better. They refuse to do it. That’s what evil companies do.

Re:Is Turbopack 10x Faster than Vite? by adoxyz on Friday November 04, 2022 @11:09AM (#33468010)
I'm sorry I don't see how they're tricking people or relying on them do it wrong... seems like you're really reaching there.
Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Tuesday November 01, 2022 @01:35PM (#33424498)
And you're just falling into their trick. Just check how this works. And how easy it is to forget to do this on every environment you're deploying to, or for every other developer in your team.

This is exactly the evil part I'm talking about.

Re:Is Turbopack 10x Faster than Vite? by adoxyz on Friday November 04, 2022 @11:12AM (#33468060)
What is the trick? They provide an easy way to disable telemetry and multiple ways to do it. It's up to your team to follow your organizations processes. If you don't want telemetry enabled, tell them to disable it and have checks in place to make sure it's disabled.
Re:Is Turbopack 10x Faster than Vite? by 0xblinq on Thursday November 10, 2022 @12:15AM (#33542533)
You don't get it.

There's an easy way to do this, which is having a config setting. They don't want to put it easy, and provide all these more convoluted ways, because what you describe is an "ideal" which doesn't reflect real life at most workplaces. That's the whole trick. "It's easy, just put this environment variable and ensure everyone on your team does it on their laptops and ensure every staging and production environment also does it and setup checks and scripts to automate it and alarms and improve your processes... it's just that easy".

Can you explain what's wrong with having a friggin boolean in a config file? Like VSCode does? Like any other non-evil piece of software does?

That's what I mean with being evil. That's what I mean with tricks.

Of course it can be disabled, and it "seems easy". That's the trick.

Re:Is Turbopack 10x Faster than Vite? by wdb on Tuesday November 01, 2022 @10:15PM (#33430695)
Always a good idea to add some network policies for egress traffic
Re:Is Turbopack 10x Faster than Vite? by afraid_to_speak on Tuesday November 01, 2022 @08:49AM (#33420431)
Is it really surprising? Facebook has never been known as an ethical company and is openly hostile toward users, why would you expect its workers to be any different?

These are people that care more about a high paycheck than the damage they do enabling genocide [1]. I honestly lost all respect for everyone in the company, they care more about optics in regards to "BLM" or "standing with Ukraine" with useless banners but I guess if you're poor and a minority it's alright to be slaughtered for user engagement.

It's disgusting.

[1] https://en.wikipedia.org/wiki/Rohingya_genocide#Facebook_con...

Re:Is Turbopack 10x Faster than Vite? by yen223 on Tuesday November 01, 2022 @10:14AM (#33421505)
It only took two comments to go from talking about web bundlers to Rohingya genocide. This website sure is something.
Re:Is Turbopack 10x Faster than Vite? by OutsmartDan on Tuesday November 01, 2022 @08:57AM (#33420526)
Before next, server-side React was a nightmare. They've simplified it a bit by providing Next.

It would be wise to decouple Vercel, the company, and Next, but Next as a framework itself is great to work with at scale.

Re:Is Turbopack 10x Faster than Vite? by the_mitsuhiko on Tuesday November 01, 2022 @10:01AM (#33421348)
The increased coupling of next to vercel does not appear to be an accident but part of the business model.
Re:Is Turbopack 10x Faster than Vite? by qudat on Tuesday November 01, 2022 @10:14AM (#33421508)
I totally agree with you, Next has introduced some innovative ideas and created a framework on top of react that rivals the quality of the react codebase itself.

However, this is the foot in the door that creates incentives that prioritizes their business model over creating a vibrant and healthy ecosystem.

Now that the react dev team and next dev team are closely working together, to the point that the react team is part of the nextjs conf announcements, it really starts to show the wolf in sheep's clothing.

This is not an open ecosystem anymore, it's being run by FB and VC money.

Re:Is Turbopack 10x Faster than Vite? by CharlieDigital on Tuesday November 01, 2022 @09:46AM (#33421176)
What I find most distasteful about this Turbopack move is that it can be seen as a mechanism to create leverage by owning the ecosystem not unlike how Apple does with their platforms and Meta is trying to do with VR.

Both target.com and walmart.com are running Next.js (view source → `__NEXT_DATA__`).

The disingenuous comparisons against Vite's performance seem innocent enough, but there's an underlying cynicism I cannot shake that the true objective is to own the ecosystem so they can create leverage against the community of users -- both commercial and non-commercial -- in the future.

I think I would personally feel very differently had they announced that they are investing in Vite, for example.

Re:Is Turbopack 10x Faster than Vite? by strangescript on Tuesday November 01, 2022 @10:24AM (#33421629)
"I wouldn’t be surprised if vercel tries to buy react and its team."

Meta would have nothing to gain and everything to lose from that, highly unlikely they would be interested.

Re:Is Turbopack 10x Faster than Vite? by swyx on Tuesday November 01, 2022 @11:57AM (#33422993)
> I wouldn’t be surprised if vercel tries to buy react and its team.

why buy it when you can just wait for people quit and come over :)