Hacker News Reader

Comment by andrewmcwatters | Hacker News Reader

Thank you. I have no idea what these people are talking about. They obviously didn't actually do development back then.

> Apple took years to finally add WebRTC support to Safari, far enough behind Chrome and Firefox that it practically became a running joke among developers and even industry observers.

Really? Because no one I know uses WebRTC and would rather use ffmpeg and sockets to do serious work like that over a desktop application.

> But at the same time, the lack of support for key web technologies and APIs has been both perplexing and annoying at the same time.

Yeah, what we all need is Web Bluetooth.

View on HN

Lots of sites use WebRTC for video and audio calls. Google Meet, Discord, Slack for example. On desktop it's far more secure to use those in-browser than via desktop apps.

Yeah, and those all suck up my RAM. Color me nonplussed that WebKit doesn't follow in Chromium's footsteps.

The criticism was that "no one uses WebRTC", not that WebRTC was inefficient. Clearly millions of people do use it, so you aren't refuting the point in any way.

Perhaps if Apple had supported WebRTC earlier their engineers could have influenced the way it works to make it better. That's the main problem with ignoring standards in an org the size of Apple - you don't get to contribute to them. I don't mind that Safari runs behind the other browsers because progressive enhancement enables me to deal with that. I do mind that Apple are failing to be a meaningful part of moving the web forwards, and are basically letting Google do whatever they like.

Millions of developers are not using WebRTC. Not sure how you read my post, but clearly you wanted to make your own spin.

Are millions of developers using Widevine, too? What kind of stupid argument is that?

>Clearly millions of people do use it, so you aren't refuting the point in any way.

Are forced to use it. And millions more would be forced to use it if Safari handed it on a plate as well...

The overwhelming majority don't know, or really care, that they're using it. People don't care about tech. They care about what they see. User complaints will never be "WebRTC is bad!". They'll be things like "Google Meet kills my phone battery!", or "Teams doesn't work on my iPhone!". The solution to those complaints is usually to make the underlying tech better rather than chuck it out entirely and use something else.

>The overwhelming majority don't know, or really care, that they're using it.

Which is why as devs we have a responsibility to save them from bad tech.

That said: "PWA: because people don't care about tech" would be an apt marketing slogan.

Safari may be better than Chrome for backgrounded tabs or throttling when idle, but I can assure you its RAM usage is still quite high for many things, WebRTC included. I'm currently grappling with some OOM issues even on the latest iPhone 12 because Safari will wildly spike in RAM usage for certain things, especially around video and iframes (Nothing to do with JavaScript heap, this is webkit engine level stuff.)

Discord in particular, is an example of Electron done right. It's super performant and low impact on resources. MS Teams, made by a vastly richer company, is an example of the opposite.

It's not Electron that's eating your RAM.

Visual Studio Code was also done right but it seems to get worse with patches.

So, it's not just me that noticed it's getting worse.

I don't see how that's an argument for not supporting WebRTC...

Feel free to not use those apps. But a lot of people do want to use them.

When people say that Safari is like IE they aren't talking about 20 years ago when IE was dominant in its prime, they are talking about the geriatric IE era when it was an also-ran that you had to support but didn't implement modern features such as flexbox properly etc.

Literally hundreds of millions of people are going to make calls through Teams/Slack/Zoom & WebRTC today alone, grandpa.

> Apple took years to finally add WebRTC support to Safari, far enough behind Chrome and Firefox

"Far behind Chrome" should never be used in a sentence. Ever. Chrome releases up to 40 new "standard" features once every two months, many of them are just internal Chrome APIs with a "standards spec" in the form of a draft.

Specifically for WebRTC. It was shipped in Chrome 23, in May 2012.

Guess what? Here's the timeline:

- In 2010 Google buys a company called Global IP solutions with some proprietary tech

- In May 2011 it opensources this tech

- In October 2011 publishes the draft spec

- Just 7 months later ships it enabled by default in Chrome

Do you know when the final stable spec arrived? In May 2018, full seven years later after Chrome enabled it by default.

That's why Safari had it implemented since September 2017, because the spec was finally in a proper shape.

No idea why Firefox rushed headlong into it though.

how do specs get to 1.0? browsers heed the call to implement & do so. they work through what it takes to improve & finalize the spec.

your timeline, to me, indicates less that the spec was rushed & bad, and more that it took apple 5 years to get off their hostage taking kick & finally face the music & do what should have been done years ago, which is implement. once apple did start to implent, 1.0 came soon after: because that is how specs work. companies implement, they find what's not right, & they improve, then make a 1.0.

I would definitely be interested in a changes over time review of webrtc. my feeling is that the core ideas & protocols were largely stable, but needed tweaking & elaboration, needed conformance suites elaborated. I could be wrong. this is in contrast to something like spdy, which evolved seemingly a good number of times in major ways, before ultimately getting parlayed into quic+http3.

Reply on news.ycombinator.com