Firefox desktop market share now below 9% • Hacker News Board

Re: Firefox desktop market share now below 9%

#201

Post by kbd » Tue, Dec 04, 2018, 4:15 AM UTC

atombender wrote: post #195

kbd wrote: post #185

Earlier quoted context omitted.

> I also tried out Firefox on iOS some time ago... I was in the process of switching to Firefox from Chrome, and a few days after installing Firefox on iOS it interrupted me with a notification advertising Pocket. I immediately uninstalled it and haven't looked back.

I use Pocket, though not with Firefox. It is a very good and useful (and free!) service. I wouldn't disregard a browser just because it comes with built-in integration.

It wasn't about Pocket itself but that Firefox was rudely using notifications to advertise to me. So I went back to Chrome, which has never done anything like that.

Re: Firefox desktop market share now below 9%

#202

Post by beizhia » Tue, Dec 04, 2018, 4:16 AM UTC

I use Firefox as my main browser (but I'm a Linux user, so I'm already away from mainstream). (edit: tree-style-tabs is the killer feature/extension, and I can never go back)

I do still pop open chrome for the devtools when I'm working on web stuff. I wish I didn't have to, but there's one thing that FF doesnt have that I need at work: goto line in the sources browser.

Lots of our files at work are several thousand lines (I know...), and its a huge missing feature to me.

I might see if I can code it up myself and submit a patch, but that's not exactly trivial.

Re: Firefox desktop market share now below 9%

#203

Post by broodbucket » Tue, Dec 04, 2018, 4:16 AM UTC

godelski wrote: post #60

muxator wrote: post #15

Earlier quoted context omitted.

Historically, that crown goes to Firefox. Its 2004 1.0 release was also publicized in the New York Times. Way before Chrome came out.

To be fair, FF shot themselves in the foot. They didn't keep pace with Chrome. I use FF and think it is better NOW, but that wasn't always the case. I think that's where they lost their market share. I'm glad they put the work into Quantum. But it is going to be hard to regain that market share.

Easier said than done when comparing a traditional open source project with a strategically important project for a massive corporation...

Re: Firefox desktop market share now below 9%

#204

Post by pcwalton » Tue, Dec 04, 2018, 4:18 AM UTC

Brakenshire wrote: post #79

Quantum I think was too incremental and conservative, trying to integrate Servo into Gecko. They’ve ended up making a desktop browser which is as good as Chrome, but not good enough to demand attention. Mobile, where parallelism should be most beneficial, has lagged behind, and parallel layout, which is where the real prize is, has disappeared over the horizon. Servo has been relegated to WebVR experiments which are…

If you've been following #servo lately, you'll see that we're starting a reboot of parallel layout right now (which I'm leading, more or less).

It just so happens that I'm also juggling a million other things simultaneously. For example, I'm working on power usage improvements on macOS that should help address some of the complaints upthread. Also there's the entirely new GPU font and SVG renderer (which is a novel technique). Also there's making sure WebRender ships. Etc.

Re: Firefox desktop market share now below 9%

#205

Post by flukus » Tue, Dec 04, 2018, 4:19 AM UTC

pcwalton wrote: post #198

flukus wrote: post #187

Earlier quoted context omitted.

> Firefox is still pretty ugly, too. On macOS, it feels chunkier and less natively integrated. Mozilla have put a tremendous amount of effort into the "ugly on every platform" cross platform tools over the years, from XUL to whatever the current one is. Developing a native UI for each platform would have been less effort and given much better results than continuing with this tried and failed approach. Their latest a…

Every single rewrite of a Gecko component in Rust has been accompanied with architectural improvements, most notably Stylo. Nobody is rewriting things just because Rust.

And those architectural improvements broke things that worked for me previously, from my point of view they aren't improvements at all.

I'm a user, we should I care that they are making architectural improvements while they rewrite things in a different language when there are so many obvious issues preventing me from using it?

At this rate they will be rewriting it for no one.

Re: Firefox desktop market share now below 9%

#206

Post by Groxx » Tue, Dec 04, 2018, 4:20 AM UTC

kristopolous wrote: post #93

saagarjha wrote: post #70

Earlier quoted context omitted.

> Transforming itself to be visually nearly indistinguishable from chrome hasn't really worked in its favor. Chrome changed its style more recently and ended up being more similar to Firefox.

Is that the timeline here? In that case it was Chrome using its market position cleverly and Firefox not responding adequately. I can only speak for linux interfaces btw. Small "bullshit" interface changes can have dramatic user and marketshare impact. We can all sit around and wish the world didn't work this way but it's not going to change.

Firefox has alternatively had closer-to-curved-rects or boxy tabs for practically its whole lifetime. Chrome went heavy on diagonal-edged tabs early on, and relatively recently they've been undoing that.

they're even mimicking Firefox's scrollable tab-bar since it's such an obvious win over a million super-narrow tabs that you can't possibly identify.

Re: Firefox desktop market share now below 9%

#207

Post by tjr225 » Tue, Dec 04, 2018, 4:25 AM UTC

pcwalton wrote: post #204

Brakenshire wrote: post #79

Quantum I think was too incremental and conservative, trying to integrate Servo into Gecko. They’ve ended up making a desktop browser which is as good as Chrome, but not good enough to demand attention. Mobile, where parallelism should be most beneficial, has lagged behind, and parallel layout, which is where the real prize is, has disappeared over the horizon. Servo has been relegated to WebVR experiments which are…

If you've been following #servo lately, you'll see that we're starting a reboot of parallel layout right now (which I'm leading, more or less). It just so happens that I'm also juggling a million other things simultaneously. For example, I'm working on power usage improvements on macOS that should help address some of the complaints upthread. Also there's the entirely new GPU font and SVG renderer (which is a novel t…

Nice! Looking forward to better power management - I'm a daily user in spite of this.

Re: Firefox desktop market share now below 9%

#208

Post by kam » Tue, Dec 04, 2018, 4:25 AM UTC

chipperyman573 wrote: post #174

kgwxd wrote: post #166

If you don't want it on desktop for some crazy reason, at least replace Chrome with Firefox if you're on Android. uBlock Origin support alone is worth leaving behind any benefit you get with Chrome.

Seconded. Plus, you can install HTTPS everywhere. One thing I would mention, though, is that using google services will all downgrade themselves if you use firefox for android. For example, this is the exact same search in chrome: https://i.imgur.com/uK8Tr9I.png and in firefox: https://i.imgur.com/pEBphaQ.png (both using the same android phone). It's even worse if you're trying to use a "near me" search. I don't want…

Or search from the Google widget on the home screen instead of the browser. Results are displayed in the native app but web links open in Firefox.

Re: Firefox desktop market share now below 9%

#209

Post by pcwalton » Tue, Dec 04, 2018, 4:27 AM UTC

flukus wrote: post #205

pcwalton wrote: post #198

Earlier quoted context omitted.

Every single rewrite of a Gecko component in Rust has been accompanied with architectural improvements, most notably Stylo. Nobody is rewriting things just because Rust.

And those architectural improvements broke things that worked for me previously, from my point of view they aren't improvements at all. I'm a user, we should I care that they are making architectural improvements while they rewrite things in a different language when there are so many obvious issues preventing me from using it? At this rate they will be rewriting it for no one.

I'm sorry you had a bad experience. But architectural improvements are done to improve performance and stability. I can't think of a single Rust component that shipped without extensive performance measurement to demonstrate improvements in the important metrics that users care about.

Re: Firefox desktop market share now below 9%

#210

Post by Sylos » Tue, Dec 04, 2018, 4:27 AM UTC

Brakenshire wrote: post #79

Quantum I think was too incremental and conservative, trying to integrate Servo into Gecko. They’ve ended up making a desktop browser which is as good as Chrome, but not good enough to demand attention. Mobile, where parallelism should be most beneficial, has lagged behind, and parallel layout, which is where the real prize is, has disappeared over the horizon. Servo has been relegated to WebVR experiments which are…

Servo is nowhere near ready. It's been a few months since I've seen actual results, but I'd be surprised, if it nowadays completed ACID3 (which was state of the art for browsers in 2009).

The thing is that rewriting a browser engine doesn't just happen over night. No bigger browser engine has been written from scratch since the previous millennium. Also, no browser engine implements all currently specified webstandards. So, Mozilla would have to full-pelt develop Gecko to be able to keep up with Blink in webstandards, and then also develop Servo at more than full-pelt to be able to catch up in finite time.

That just doesn't add up, which is why Servo was specifically started as a research project. It might never be completed. They have been able to integrate various components from Servo directly into Gecko, therefore sharing the development work which they didn't really plan with, so that's one tiny reason to still hope for it to ever catch up, but yeah, just don't be too optimistic.

Having said all that, the integration of components from Servo into Gecko isn't done yet. WebRender is still missing, which should bring another good performance boost. It's sort of been overhyped with artificially complicated CSS animations [0], so the real world effect isn't that big, but it's still very much noticeable.

[0]: https://www.youtube.com/watch?v=u0hYIRQRiws

Aside from that, on Android, Mozilla is building a new framework for building Android browsers in general, called "android-components", which also includes a much cleaner integration of Gecko on Android, called "GeckoView". They've already rebased Firefox Focus onto GeckoView and are actively working on rebasing Firefox for Android (Fennec) onto it, too, which is internally called "Fenix". This should also significantly reduce the lagginess and speed up their development in the long-run.

I guess, what I'm mainly trying to say is that there's definitely still things happening.