Hacker News Reader

Comment by cwzwarich | Hacker News Reader

> Why aren't we collectively putting our time/effort etc.. into making Firefox the singular and best FOSS browser ever?

[disclaimer: former Servo team member]

Mozilla had a chance to advance browser tech with Servo and fumbled it.

Not to be negative, but what value would Servo exactly bring to users and the web after all the engineering effort? I'm sure it was technically impressive.

Chrome, when it came, was stupidly faster than any other competitor, including Mozilla Firefox.

What would Servo bring to make web a big jump forward for us users, beyond ideology and fun work for its developers? (Ideology doesn't sell software, as opensource efforts have proven repeatedly over last 3 decades).

Same selling points as Chromium originally: faster and safer. The premise of Servo is that it'd make better use of multicore and would be safer because it was written in a memory-safe language. For modern users the big one is probably battery usage, though, which Safari excels at but would be harder for an open source player that has to support every device.

> What would Servo bring to make web a big jump forward for us users

The biggest advantage would be having a memory-safe browser engine.

As someone with almost no knowledge of Servo, and personally not having a good opinion of the Mozilla foundation, your comment tells me all I need to know why they stopped support of Servo development.

Only social initiatives get feel good money to waste on nice to haves. This "Servo" thing does not sound like it.

I want to draw one's attention to the (at this point impressive) number of CVEs related to memory issues in Chromium's C++ codebase: https://chromereleases.googleblog.com/search/label/Stable%20... and control-f for CVE

Even granting that some of them come from dependent libraries, if the world's arguably best funded software engineering organization can't keep C++ footguns out of an application that runs untrusted code on the regular ... one may want to have a think about whether C++ is the best tool for that purpose

So what's the monetary value of not having those CVEs? How did any of that actually touch users beyond having to wait a bit for Chrome to autoupdate?

https://search.brave.com/news?q=chrome%20exploit%20in%20the%...

I'm delighted to hear that you (and I) were not personally impacted by those but to say "I didn't get in a crash, who cares about seat belts working" is comically misguided

Also your "for Chrome to autoupdate" perspective makes me wish I could search for every single red Update screenshot I've seen in presentations or Zoom meetings

Again, how did that materially affect users to the point they'd switch?

The point is that all this is not seen as valuable by users and market leaders. You don't gain marketshare in b2c by saying "my gizmo is more memory-safe than other gizmos", you have to show features. Chrome's process-separation was a feature: you could easily dispose of non-responsive pages, which other browsers couldn't do as well. Even then, it wasn't really why Chrome took off - it took off because of deep integration with Google's ecosystem, incessant marketing efforts, and monopolistic behaviour (Google "unwittingly" breaking GMail and other G products on other browsers, on a regular basis).

So what feature would full memory-safety actually enable? I suspect that's the bit that the Servo folks couldn't articulate well enough.

> but what value would Servo exactly bring to users and the web after all the engineering effort?

Servo was a much more modular effort.

1. we would get a lightweight (modular) electron alternative. This would have helped rust gui ecosystem a lot, just like wgpu helps rust gamedev scene.

2. Probably have more variety in browsers as everyone can now reuse servo components to make their custom version of firefox. Imagine all the linux distros which reuse gnu/linux, but have their own philosophy of package management, updates, mutability etc.. We could have had the same sort of varieties in browsers for different usecases.

Plus it'd let people use a JavaScript engine other than v8 and JSCore, and one which (in theory) would be memory safe.

> Chrome, when it came, was stupidly faster than any other competitor

Wasn't it just using WebKit initially? IIRC you could actually use Safari on Windows back then.

Are they not using any Rust in the codebase at all? My impression was that they abandoned Servo but still rewrote parts in Rust.

Some big chunks of servo got moved into firefox, so it wasn't a complete loss. But it would have been better if servo kept going, even if it never became something you would daily drive and instead it just kept generating good ideas to be subsumed into firefox.

Servo is active again thanks to Igalia allocating a handful of full time engineers. They are making good progress especially on layout, so keep an eye on it!

Is that sufficient reason to abandon the entire project?

Reply on news.ycombinator.com