Hacker News Reader

Comment by Touche | Hacker News Reader

Am I the only one that is confused by Firefox's "fast again" narrative? I remember that they launched the https://arewefastyet.com/ site a few years ago to fight the idea that FF was slow. It feels like there's a new blog post that essentially says the same thing "we're fast now" every couple of months.

So I find this all confusing. Am I just remembering incorrectly, or haven't they already declared "we're as fast as Chrome now" several times in the past. Were those lies/incorrect? Or again, am I remembering incorrectly. Can someone explain this?

View on HN

> Am I just remembering incorrectly, or haven't they already declared "we're as fast as Chrome now" several times in the past.

https://arewefastyet.com/ refers specifically and only to Firefox's javascript engine. It runs js benchmarks only, not whole websites, and doesn't test responsiveness or rendering performance.

Javascript performance in firefox has been on par with competing browsers for a while now. The current initiative, called Quantum, is complementary to this effort. It is supposed to make the rest of the browser components fast, using among other things a new renderer, a CSS styling engine, and multiprocessing support.

Not entirely true anymore. If you select eg "Quantum Reference (Windows, browser, x64)" and look at the Breakdown list, you'll see a couple of browser benchmarks that look at responsiveness and rendering. Speedometer and Unity WebGL, for example. But yeah, it's mainly all about JS.

Since the Quantum effort started, the JS team has prioritized perf work higher, so quite a bit of perf-related improvements have been made in Spidermonkey for 57 as well. (And Quantum DOM needed some rearchitecting to re-multithread the engine in a limited way.)

(Source: I work on Spidermonkey.)

Performance is a Red Queen problem. [1] People keep expecting more from computers. On the web, pages keep getting bigger and more complicated. Firefox's competition isn't going to stop trying to improve performance, and people's baseline for "fast" will mainly be other things they use. If Mozilla isn't always trying to be faster, they're going to be getting slower.

[1] https://en.wikipedia.org/wiki/Red_Queen_hypothesis

Jevon's Paradox [0] is probably more appropriate.

As resources become more plentiful (available performance-units, whether delivered through faster software or faster hardware), demand counter-intuitively increase, in part because the cheaper per-unit price makes people feel they can use the resource more freely.

This phenomenon describes why no matter how fast our computers get, it's unlikely that they will ever be "fast enough".

[0] https://en.wikipedia.org/wiki/Jevons_paradox

Both are definitely relevant, but since you insist on ranking them, the Red Queen phenomenon is more relevant to his question, which is about how Firefox is portraying themselves in the marketplace, which is a competitive environment.

There have been multiple separate initiatives over the years to speed up various aspects of Firefox. The largest one is culminating in the release of Firefox Quantum, which is launching as Firefox 57 tomorrow. You've heard this more lately as the Quantum version of Firefox was released to Nightly and then Beta/Developer channels over the past couple months.

I'm confused by it, too. The main reason I've preferred Firefox to Chrome for years has been performance. It's pretty much the same in responsiveness and uses far less memory.

Chrome being the fastest browser is ancient history.

The main reason I originally left Firefox (for Edge, mind you, I don't touch Google software anymore) was the single-process issues they fixed with electrolysis which came out in Firefox 48.

Bloated websites (usually Google ones) would actually lock up the entire Firefox UI, and Windows would recognize the browser application as "Not responding" until the page finished loading. In the case of a super bloated site like the old Google+, that could be as much as twelve seconds non-responsive while all the cruft loaded up.

I am back on Firefox release channel now, and pretty satisfied with current performance, but looking forward to Quantum.

I'm a Firefox or native-platform browser guy. Some sites are laggy on anything except Chrome. Browsing around Groupon causes periodic freezes for me in Firefox. I haven't noticed this particular case in Edge but it does lag over time on other sites.

The whole issue of sites freezing and lagging without using Chrome is concerning, and it's not pushing me to use Chrome. I think that's Google's aim by creating that situation but I'm not budging.

It's pushing me to more native apps in both the iOS AppStore and Microsoft Store.

Of course, the sad state here is that a lot of apps are now Electron-based, so you end up with Chrome anyways. >.<

I've been wondering about that, and how for the "app" side of the webstack if wasm will wipe out the traditional webstack "apps" and put things back in a more orderly fashion. Webstack for web documents, wasm for applications. If wasm becomes the univeral app platform for many appstores and the browser, I'd welcome that over Electron.

It's probably inevitable that shoving Node based apps into the market falls flat in favor of more efficient native code solutions due to mobile. I've been viewing Electron as a temporary stopgap with that in mind at least.

I had major problems using LiveReload and BrowserSync during development (two tools that automatically refresh a local website when you make changes to your code). After a few saves, it was too slow to use. Chrome and Safari would refresh about twice as quickly, and remain fast.

57, once it got closer to Beta, was the first version I could develop with.

That wasn’t a marketing site, it was a tracker to measure the increase in performance over time due to the ongoing work the newly-spun-up performance team was doing.

Yeah, sorry, I didn't mean to imply it was a marketing site. But there was marketing surrounded it, and I distinctly remember it being used to show that they were, indeed, as fast as the other browsers. I remember a narrative being that the problem wasn't Firefox being slow, but FF being slow in certain perceived-speed areas (like new tab speed).

That's why I find these sorts of announcements confusing, I can remember them saying that there are already as fast as Chrome several times in the past. But maybe my mind is playing tricks on me.

No, you're right.

They'd declare that the JS engine was just as fast (or faster) - but that didn't cover the whole of the UI, and laggy bits of the UI could slow things down no matter how fast the JS engine was.

Thankfully they now seem to have improved the whole UI.

I recently (a couple of weeks ago) finished giving Firefox another try for about two weeks. Sorry to say, but it just felt laggy compared with Chrome. I'll give it another shot when the new rendering engine is fully integrated into the release channel.

Did you try the nightly or Beta version? If not, you weren't experiencing the new changes. These will drop on the stable channel within the next day or so.

All the stuff they are talking about is being released with 57 tomorrow.

So maybe give it another spin :)

Yep, will do.

Reply on news.ycombinator.com