Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Chrome Won

andreasgal.com

Post reply on HN
Print view View original

571–580 of 583 posts

Re: Chrome Won

#571

Post by stupidcar »

pcwalton wrote: post #518
bokan wrote: post #512

Earlier quoted context omitted.

The question isn't whether you should or shouldn't try to improve, it's where to invest your resources. Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from…

> Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from dozens of domains on any given page load. We need to fix the web platform so that apps don't need to…

I have never implemented a custom layout in JS because the browser's layout is slow. My JS-based layouts are always slower than using CSS.

I do layout in JS because the existing layout capabilities don't cover my use cases. I create complicated visualisations in SVG, and SVG's layout capabilities are so pathetically anaemic that I have no choice but to take over and use a mixture of D3 and custom JS to arrange things.

Similarly, I don't do synchronous layouts because I'm a moron who doesn't understand performance, but because I need to measure the size of various data-bound elements and feed those dimensions back into my custom multi-pass layout algorithm.

AFAIK, Houdini is an actual attempt to solve both of these problems: It will let run my custom layout algorithm in a worklet, and give it access to the font metrics and element sizes it needs to run efficiently.

The Houdini people seem to be the only web platform people who actually "get it" with respect to these kind of problems. Everyone else seems to be either ignorant or dismissive that these problems even exist.

Re: Chrome Won

#572

Post by roca »

pcwalton wrote: post #518
bokan wrote: post #512

Earlier quoted context omitted.

The question isn't whether you should or shouldn't try to improve, it's where to invest your resources. Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from…

> Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from dozens of domains on any given page load. We need to fix the web platform so that apps don't need to…

I think that's a bit unfair to Houdini. As others noted here, one of the main goals for Houdini is to make CSS cleanly extensible, so every time some cool new feature comes along you don't have to choose between a) waiting for it to be standardized and implemented in browsers and b) abandoning CSS entirely.

Of course, I agree with the rest of your points.

Re: Chrome Won

#573

Post by roca »

bokan wrote: post #558
pcwalton wrote: post #518

Earlier quoted context omitted.

> Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from dozens of domains on any given page load. We need to fix the web platform so that apps don't need to…

> which means that layout performance affects what looks like script time Chrome's profiler actually shows synchronous layouts and style recals forced from JS as layout/style so it's not misleading in that way. The traces I've looked at (admittedly, not something I do terribly often) showed those to be typically ~20% or less of the time spent while content is loading. > And, due to ads, a lot of the performance cost…

> I think this is a great example of how having multiple rendering engines is healthy for the web.

Yeah, it's too bad that Google's Chrome-first/Chrome-only approach to Web development these days is making that increasingly difficult to sustain.

Re: Chrome Won

#574

Post by jhasse »

asfdsfggtfd wrote: post #375
jhasse wrote: post #177

Earlier quoted context omitted.

Not in the same way as IE.

Closed source is a boolean. Chrome is distributed only as a binary. I.e. it is totally closed source. It currently has a very large relation to the open source Chromium project. But Google could change that tomorrow if they wanted to - they could also gradually move more and more to their closed source Chrome builds (as they have done with Android).

> Closed source is a boolean.

I disagree. There's a difference between having 99% of the source available and 0%.

Re: Chrome Won

#575

Post by scholia »

AnimalMuppet wrote: post #497
scholia wrote: post #486

Earlier quoted context omitted.

> as part of a campaign where Microsoft was paying web sites to put in stuff that was incompatible with Netscape/Mozilla I'd be interested to see the evidence for that....

Sorry, a quick search doesn't turn it up. This is from memory. Microsoft's strategy - from their own memos (maybe the Halloween memos?) - was to "try to make using Netscape a jarring experience". They were paying (maybe in equipment or some kind of freebies) websites that included at least three IE-only elements in their pages. This was not just competition. This was a deliberate campaign to break the web in a way wh…

My memory is different, obviously. The IE team took great pains to implement Netscape additions even when they were not standards so that the IE rendering was as good. I never heard of any attempt to make Netscape look worse. That doesn't mean it never happened, but I followed the story very closely at the time.

Microsoft was much more co-operative than Netscape in the early days. It was one of Microsoft's advantages when Netscape was winning and running on pure arrogance. See How the Web Was Won, High Stakes No Prisoners and a few other books for details.

Microsoft did introduce ActiveX, which Mozilla considered supporting, and then decided not to.

Re: Chrome Won

#576

Post by scholia »

willhslade wrote: post #480
hooluupog wrote: post #353

Earlier quoted context omitted.

Except china.

Can you elaborate for people not in the loop as much?

Android is pretend-open. Technically, you have to use Google Play to use the Android name. If you use AOSP then you lose the store and Google's proprietary apps, so you have to build an alternative store and plead for third-arty app support.

That works in China because Google is relatively weak there. It also works for Amazon, which has its own store for Fire products.

Re: Chrome Won

#577

Post by pcwalton »

roca wrote: post #572
pcwalton wrote: post #518

Earlier quoted context omitted.

> Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from dozens of domains on any given page load. We need to fix the web platform so that apps don't need to…

I think that's a bit unfair to Houdini. As others noted here, one of the main goals for Houdini is to make CSS cleanly extensible, so every time some cool new feature comes along you don't have to choose between a) waiting for it to be standardized and implemented in browsers and b) abandoning CSS entirely. Of course, I agree with the rest of your points.

That's fair. I don't have an issue with Houdini if it's actually used to create custom layouts that you can't reasonably do in CSS.

Re: Chrome Won

#578

Post by pcwalton »

stupidcar wrote: post #571
pcwalton wrote: post #518

Earlier quoted context omitted.

> Servo improves the speed of layout, style, etc. but is that really what's making the web experience slow? If you do some performance tracing, you'll see your browser already spends a paltry amount of time doing those things. It's spending most of its time running ridiculous amounts of JS that's being pulled in from dozens of domains on any given page load. We need to fix the web platform so that apps don't need to…

I have never implemented a custom layout in JS because the browser's layout is slow. My JS-based layouts are always slower than using CSS. I do layout in JS because the existing layout capabilities don't cover my use cases. I create complicated visualisations in SVG, and SVG's layout capabilities are so pathetically anaemic that I have no choice but to take over and use a mixture of D3 and custom JS to arrange things…

If you're actually using Houdini for custom layouts (including SVG—SVG doesn't really have any layout at all), then I don't have a problem with it. My issue is more with the performance-oriented Houdini specs, for example Animation Worklet.

Re: Chrome Won

#579

Post by aerovistae »

madeofpalk wrote: post #531
aerovistae wrote: post #261

Earlier quoted context omitted.

Well, personally I read books. I think I learn more and remember more from the books than I or they do from social media feeds, which I think are a lot like putting your mind on hibernate to get from point A to point B in a temporal sense. I don't ask that everyone read books, god forbid, but it would be nice if people weren't using their phones to put themselves into hibernate nearly all of the time. The phrasing of…

> Well, personally I read books > I just wish people would engage more. Some people loose themselves in their phone, others loose themselves into a book. They're both the same.

Do you really believe that?

If you do, then you must agree that you could replace every moment in your life that had been spent reading a book with you perusing your phone, and you would be the same person today.

Do you believe that to be true?

Re: Chrome Won

#580

Post by ocdtrekkie »

kenoph wrote: post #570
ocdtrekkie wrote: post #359

Earlier quoted context omitted.

Almost every Chrome install I see on laypeople's PCs contains malicious extensions (officially hosted on the Chrome Web Store!) which have permission to see and modify all page content and collect data and/or inject ads. Google may focus on specific types of exploits, but they let everyone bad just walk right in the front door.

That's true. But if you want to look at it from the perspective of how laypeople use PCs, the discussion about security is totally different from a purely technical one. Assuming you know better than installing adware or installing random chrome extensions, Chrome is the most secure browser out there. Btw, regarding extensions, before Mozilla switched to WebExtensions, you could do nasty stuff like extension-reuse at…

I understand that people at Google tend to think in the "purely technical", but that's why I feel they're so poorly equipped to protect people. They fail to account for the real world, where people do dumb things. It's not okay to build a platform you intend non-technical users to use, and then blame the user for getting exploited when you do nothing to prevent it.
Post reply on HN