[nijave]

I've been using Firefox for nearly 15 years and there's technical reasons, too. Mainly memory leaks (which I haven't noticed for a few years anymore) and loading just "feels" slower. Webrender with GPU acceleration helped but it was much more complicated poking at about:config than it "just working"

Recently, I've noticed a "this SaaS only works in Chrome" trend again. Usually that means "only works in Blink" but it still sucks if you need to use SaaS (for work) that has this silly limitation. I've noticed cheaper restaurant web ordering software tends to have problems in Firefox.

For technical or even semi-technical folks those aren't an issue. For non-technical folks, those issues above can easily lead to a poorer web experience without them understanding why. On the other hand, the author didn't actually call out technical issues

[madeofpalk]

You'll be happy to know that recently I shipped a bug that only crashes (in) Chrome, but not Firefox, because I only really develop in Firefox.

[paulryanrogers]

Wow, crashes all of Chrome and not only the tab process?

[madeofpalk]

Ah sorry no. Just a naive error throwing an exception that's easier to hit in Chrome than Firefox.

Pro tip - don't try and spread 200k arguments into a function!

[tucnak]

The reason Firefox feels slower is because it uses a subpar caching strategy that's minimizing RAM usage. Otherwise, they could no longer claim Chromium is "wasting memory" which is Firefox-speak for "we refuse to implement comprehensive caching"

[nijave]

iirc more about:config tweaks from FasterFox project helped a bit. Anecdotally, I've never seen Chromium be any more or less memory efficient (well maybe 10 years ago but not recently). For the longest time, if you left Firefox sitting open > than 1-2 weeks it'd start leaking memory to the tune of 10s of GB

Chromium is pretty good about suspending pages when it's left sitting for long periods of time

[tucnak]

Chromium is amazing about suspending pages, & evictions in general. They maintain multiple levels throughout the render pipelines, and because the tabs are composed of individual processes, it can trade-off somewhat against kernel. This is what a comprehensive caching strategy is: a layered, machine-aware strategy that is both aggressive and predictable under real workloads. It wasn't always perfect, but it's pretty near-perfect these days. And this is not even considering V8 and the talent pool it has acquired due to quality engineering and the network effect that came from it.

Firefox engineers are trying, and you cannot blame them (like you would blame Mozilla) because it has not seen the kind of commitment from industry talent that Chromium. There have been multiple re-writes of major components in Chromium, and they tried really hard to keep the codebase somewhat up-to-date. Firefox is comparatively dated, it's not at all sexy; had people understood this, they wouldn't be so quick with jumping to conclusions, & all. There are many objective, engineering reasons why Firefox is lagging behind the gigantic Chromium+V8 ecosystem of browsers and JS backend industry. For some champagne-socialist reason, it's really vogue to call for Firefox adoption, donating to Mozilla, etc. However, it's rarely an argument made on merit. I think, they just hate to admit that Google has largely succeeded in elevating Chromium to the likes of Linux. If you're making a browser in 2025, unless you're making an artistic point a-la privacy, or contra Chromium specifically, you would rather pick it up at upstream like you would pick up Linux, instead of re-inventing the kernel because Linux bad. That is not to say that OpenBSD, or whatever, Plan9—is without merit, just that it wouldn't surprise anybody had you picked Linux for the job. If you hate Google's SWE people, or its leadership that ostensibly could sway the course of Chromium development, then this is probably really painful! If you don't hate them, you just carry on about your business. And so everyone complies. Because why the hell wouldn't you? There's battles to fight, and there's the other kind.

It doesn't help that Mozilla is cutting the branch from underneath itself.

[nijave]

Thanks for the details on cache & agree with the rest

LadyBird and Servo are both interesting projects however I imagine still many years out of their succeed