Google, Apple and Mozilla Team Up To Build a Better Browser Benchmark (engadget.com) 26
Speedometer 3 will be a "cross-industry collaborative effort" from the Chrome, Safari and Firefox makers to create a new model that balances the companies' visions for measuring responsiveness. Engadget reports: Three companies making a tool that will rate the effectiveness of their competing products sounds like a recipe for disaster. However, Speedometer's governance policy includes a consent system that differs based on potential ramifications. For example, significant changes will require approval from the other two companies, while "non-trivial changes" will need consent from one of the other two parties. Meanwhile, "trivial changes" can be green-lit by a reviewer from any of the three browser makers. The policy's aim is that "the working team should be able to move quickly for most changes, with a higher level of process and consensus expected based on the impact of the change."
The project will follow Speedometer 2, the current de facto benchmark developed by Apple's WebKit team. The Speedometer 3 project is still in its infancy, and its GitHub page warns that it is "in active development and is unstable." The groups recommend using Speedometer 2.1 until development is further along, though we don't yet know when Speedometer 3 will be ready.
The project will follow Speedometer 2, the current de facto benchmark developed by Apple's WebKit team. The Speedometer 3 project is still in its infancy, and its GitHub page warns that it is "in active development and is unstable." The groups recommend using Speedometer 2.1 until development is further along, though we don't yet know when Speedometer 3 will be ready.
Power (Score:3)
We need a way to measure how much energy a browser uses, especially on mobile. I'll take 100ms longer loading for 20% less power consumption.
Re: (Score:2)
That's easy to do. Cut out all the ads and you'll find that you cut down on processing power considerably if you have less popping up, over, under, and with less animation to deal with you also reduce power consumption considerably.
Re: (Score:2)
I tried that in Firefox and it hammered the battery.
Maybe if it's built in to the browser instead of an extension. Chrome's much hated manifest V3 basically does that.
Re: (Score:2)
One has to wonder why. Maybe the site tried desperately hard to load that ad? Or was the blocker the culprit?
V3 is yet another thing you'd have to trust, and given its source, I have no reason to consider it trustworthy. If anything, it will be Google's version of extorting advertisers and those that cough up the dough will be allowed to pester you with ads.
Re: (Score:2)
I think the Firefox for Android JavaScript engine, and therefore all add ons, is really inefficient. Fast, but not light on the CPU.
"Better" (Score:2)
In other words, "one where our browsers test better than anyone else's".
Re: "Better" (Score:2)
Re: "Better" (Score:2)
Re: (Score:2)
Re: (Score:2)
Soon websites will just be webassembly drawing on canvas, problem solved.
Re: (Score:2)
In other words, "one where our browsers test better than anyone else's".
There are almost no other browsers. They are almost all based on one of these. There is Pale Moon, but it is already well-known to be much slower and significantly incompatible with modern HTML.
About apples and oranges... (Score:2)
No idea if they'll get anywhere (Score:2)
But it's impressive that the owners of all three existing (remaining) web rendering engines have signed onto this.
Oh god no. (Score:2)
Three companies making a tool that will rate the effectiveness of their competing products sounds like a recipe for disaster.
Yes, for everyone.
However, Speedometer's governance policy includes a consent system that differs based on potential ramifications.
Oh good.
For example, significant changes will require approval from the other two companies,
So Apple can slow down browser development unilaterally
while "non-trivial changes" will need consent from one of the other two parties.
So anything that can somehow be described as "non-trivial" and yet not "significant" (anything which is significant is non-trivial, and vice versa, by definition) can be done with only one other interested party?
Meanwhile, "trivial changes" can be green-lit by a reviewer from any of the three browser makers.
Oh good, and nobody will ever argue over what is or isn't trivial.
This is not only not better than the chaos we have now, it is actually worse, because the same exact fuckery will occur plus some new stuff. Now they wil
What exactly do you benchmark, though? (Score:3)
Without a standards-compliant test site produced by a third party, there's no way to test without biasing results somehow. Most "real world" sites are currently built and tested for Chrome compatibility/functionality. Google is screwing up web standards just the same way MS did years ago (IE 6). The only way anyone else can compete is to out-Chrome Chrome. Or you cut corners somehow (e.g. cheat the test software).
Add a test for bad page code (Score:2)
Re: (Score:3)
Most autoscrolling sites have this problem. When they scroll they leave handlers active in stuff that's way up the page. They need to remove the click and hover handlers from the invisible elements. But arguably, the browser should do that for them, since you can't possible click or hover over stuff that's far, far away from the current view...
Re: (Score:2)
Chrome effectively does that. JavaScript gets throttled to the point where it's unloaded. Works for event triggered and background processes.
One of the reasons why Chrome uses so little energy on mobile.
Re: (Score:2)
Chrome effectively does that. JavaScript gets throttled to the point where it's unloaded
All that means is that the site slows down while you're using it, it just doesn't slow your PC down in the background.
Re: (Score:2)
It doesn't work like that. Stuff that is actively updating the page can run so it feels responsive. It's the background BS like telemetry, mouse tracking, ad loading and so on that gets hammered. You might wait a few extra milliseconds for that JS driven page to pull in new data.
The only time it is really noticeable is on mobile with background tabs not updating until you look at them.
without the benchmark (Score:1)
I first read "Google, Apple and Mozilla Team Up To Build a Better Browser "
Oh interesting !
Then the disappointment kicked in when reading that last word of the title.
Sounds pretty good! (Score:2)
Lots of things need to be measured so that we can choose the best browser. Things such as:
1. How quickly a consent-to-spy form pops up instead of the web page.
2. How many hundreds of things need to be explicitly deselected on the consent-to-spy form.
3. How many clicks can be cached while the page is loading to then be handed to the advert link placed under the click.
4. How much, if any, of the information exists on the loaded page which a link claimed it would take you to.
etc
No-ads benchmark, please? (Score:2)
The only thing I need from a browser is how effective it is in stopping ads, pop-ups and personal information syphoning. Really, without AdblockPlus I would stop using WWW alltogether.
Sometimes I happen to "enjoy" a browser without an adblocker and each time it was a harsh reminder of how ad bloated WWW has become. So for me browser speed is worthless if it is only used to speed up the amount of ads I'm seeing, how fast I get pop-ups or cookie consent questions and how fast the websites are tracking me.