Google to pay JavaScript frameworks to implement performance-first code

Nick Taylor on November 26, 2018

Photo care of Unsplash user Sharon McCutcheon Interesting initiative by Google, Google to pay JavaScript frameworks to implement performance-first... [Read Full]
markdown guide
 

Basically they're paying people to use custom-built Chrome-only features.

Which in my opinion leads to:

  1. New paradigms, hence new frameworks (or at least new major versions which are incompatible with the old ones)
  2. A step further towards a Chrome-only Internet

In short, that is terribly bad.

 

Even as a Chrome user, I do find that terrible. More web Monopoly.

 

Not sure about Chrome-only part. Opera, Vivaldi, and Brave have same features in regards to developer tools so it is a bit Chrome-BASED-only but anyways who cares. If it goes that far it could lead maybe to different kind of web as a response from other part of community. Maybe new language for web like replacement for HTML, CSS, and JS or some crazy thing like that.

 
  1. How is number one bad?

  2. Someone has to pull the blanket, or APIs with better performance will never come.

 

Microsoft still supports (horribly bad) design decisions dating from the 70s. Let me say that I really don't like Microsoft but if they did something good over the time is that they care about their customers and enabling their businesses.

Currently the web development industry is like a headless chicken running in all directions. Nobody understands what we're doing. How can we guarantee the work we make if we can't understand it? All we do right now is throw our customer's money through the window because we have no idea what we're doing.

We all need to take a step back and chill the fuck down.

Also, Google is terrible at web performance. All their apps are heavy as fuck. If you open Gmail you'll instantly pop 300 Mio of RAM. As a comparison, dev.to takes 50 Mio of RAM. How the fuck is it so hard to display a list of emails?

This grant is hypocritical bullshit. It's a booby trap into vendor lock-in.

In politics when you want to pass anything you say it's to protect from terrorism. In (web) development when you want to pass anything you say it's for performance.

Come on, Gmail is way more featureful than dev.to.
And a legacy from before easy code-splitting.
And a legacy in general.

Anyhow, that's Google being a consumer of the browser.

Google as a producer of the browser makes the ever faster (and at its time, impossibly groundbreaking) V8, and their renderer isn't too bad either.

What's your stance on JavaScript in general?

 

If there are developers out there not developing performance-first, idk what they think they’re doing but they should fix that. And google, and all companies that use oss, should alott funds for employees to donate to any project they want.

 

The rationale for non-performance-first code is that they need to get features out fast to stay relevant. Most devs who would use an open-source framework wouldn't care how fast it is, as long as it's not ridiculous. They care that it can do X, Y, and Z. Tech giants DO care about small speed improvements, though. This could provide a good counterbalance so that devs are still rewarded for delaying features to make sure their code is high-quality.

 

I guess I'm just not most devs then because I always check for performance when picking libraries.

Libraries are a different story, because you will lock you architecture to them and they are huge pieces of work you will probably not be able to optimize.

 

Not sure what performance first means, but you shouldn't optimize first. Its better to optimize later so that you can profile the performance of your overall system and only optimize what matters.

 

This isn't really true. Even at the broadest design level, there are many ways to do one thing, and some are faster than others.

IMHO it doesn't really matter in the web frontend context, which is what this thread is about (you can always easily go back to the drawing board if performance becomes a real problem). I will grant you that if you need to build a video streaming service for example, it may be worth building certain components of your backend in C++ right out of the gate.

 

With my 30 years of both open and closed source software development I can safely say that in the vast majority of cases if you're developing performance-first you're doing it wrong. Develop for correctness first, maintainability second, and performance ... somewhere about nine thousandth. The only time I've felt the urge to develop performance-first was when I was writing code that ran under CP/M and was bit-banging directly at an industrial controller. It really needed to not miss any messages coming from the large potentially explodey thing, lest the potentially explodey thing notice and expensively shut itself down to prevent said explosion.

 

The Web needs all the performance it can get. The current web projects are horrible at this chapter. Electron apps eating GB of memory, browser extensions hundreds, and the avg count of requests per site I think it got over 100?

I don't see the division happening, who does not want performance first features? Probably the RAM manufacturers.

 

In regards to division, I didn't mean a division over performance. I meant JS open source or open source in general in regards to $$$. I should have made that more clear.

 

I agree with Adam Rackis (tweet in initial post) that companies that use packages like Babel, should be contributing $$$ cash to them.

It's a drop in the pond, but I donate to @hzoo 's Patreon for this reason.

companies that use packages like Babel, should be contributing $$$ cash to them

No they should not. The might want to, they might indeed do, but they should not. Otherwise it’s not OSS anymore, it’s another kind of business sucking money from giants.

companies donating to OSS do not immediately invalidate them being OSS by any means

I think the point was that companies that use OSS shouldn't be immediately expected/required to give money to those projects. It's nice if they do, but it is their choice.

comments farther up in the thread were in agreement, but I was specifically highlighting

they should not. Otherwise it’s not OSS anymore

Well, I meant it is up to these companies.
They could, and that it fine. They are not obliged.

 

The issue of funding OSS also brings up the issue of the very recent event-stream package incident.

 

Since this is in respect to javascript community and it affects them most I think there won't be much divide, since the community is pretty open and accepting of positive changes by corporations as long as there are no hidden T&Cs that should be red flags.

There might be more competition in javascript framework world but that imo is a good thing.

If this was any other community then the story might be a lot more different.

 

There's nothing new here and no divide. It's common for large IT companies to pay for features, bug-fixes, maintenance etc of open source projects.

 
 
Sloan, the sloth mascot Comment marked as low quality/non-constructive by the community View code of conduct

Yeah, Google NDA allows almost everything save for talking about Google. Like Fight Club, but with restrictions.

I would also be glad to hear what might enforce a human being to work for Alphabet nowadays. Starving to death? Being proud of selling private data? Anything else?

 

I LOL'ed so hard about "low quality/non-constructive label".

If anybody says something that you don't agree with you can just mark this as "low quality/non-constructive" and say that a "community" did that.

That is true though. That is how community works.
I am unsure what CoC line I violated, but the community does not like whatever is against the Opinion.

How dare you speak against Google/Alphabet.

Side note: I imagine it's a badge of some sort to have worked for Google. Their employees are (perhaps anecdotally) geniuses, so to be counted among that number would be bragging rights indeed.

code of conduct - report abuse