Hacker News new | ask | show | jobs
by fartcannon 1599 days ago
I would use Brave if they didn't use chrome as a backend. We need more diversity in rendering engines otherwise things like FLoC happen.
3 comments

If you want to ship a browser without writing a browser engine, there are three realistic options. Chromium is the best of these options.
> [...]there are three realistic options.

Is the Librewolf option included in those three? Because that (a patch on top of newest Firefox) is something I have been wanting for a long time and I just recently realized someone had done it.

> Chromium is the best of these options.

Easiest is maybe the word you are looking for here?

Because best depends on your goals. If your goal is to reduce dependency on Google (as it is for many here) it could be close to directly counterproductive.

FLoC has nothing to do with rendering and can easily be removed from the codebase. People make this argument a lot but I don't think it makes logical sense. We all use libicu too, what does that ultimately matter?
Rendering engine diversity is a different problem. In order to actually have a web standard, then we need at least two frequently-used implementations. If there is only one implementation, or one dominant implementation, then people stop coding to the spec and start coding to whatever works in browser.[0]

Right now, if Mozilla goes bust, then the last hope for web standards is the fact that Apple's mobile OS rules preclude shipping Blink on iOS. If everything is Chrome, then we're right back to the days of Adobe Flash Player, where bugs in one implementation effectively become part of the spec.

If we didn't have or didn't care about this problem, then yes - we could just clip FLoC out of the browser and run that. However, that also assumes Google won't mandate FLoC or ban alternative browsers. Their recent attempts to cut down on credential stuffing through headless CEF browsers[1] have already made it harder to actually ship an alternative browser that can log into Google. So hypothetical future Evil Google could totally say "no FLoC, no service" and we would be screwed.

[0] This is the same reason why WebSQL died on the vine. SQL is a standard with a lot of wiggle room, enough that most applications either only work on one particular flavor of DB, or ship separate drivers for each DB they need to work with. In practice everyone shipped SQLite, so WebSQL was just "here's a sandboxed SQLite instance".

[1] Which, BTW, is not a bad thing in and of itself.

> where bugs in one implementation effectively become part of the spec.

Isn't this basically how HTML5's parsing rules came out, except instead of bugs in one implementation it is based on bugs on all implementations? :-P

Whats wrong with websql being sqlite?
Well that is great news.

It ultimately matters because if we didn't all use Chrome, Google would have to compete more on good things, instead of building bad things that we have to remove. Basically, it's the position of the privacy whack-a-mole oveton window.

Diversity of code bases limits the impact of bugs and implementation quirks.

For example I used to be part of a team that ran an important DNS service. We purpously used several different server softwares on different nodes to increase resiliency.

I believe Brave was initially (supposed to be?) based on Firefox but needed support for a feature that only Chromium supported and had go with that.
feature was DRM
IMHO it was a big mistake of Firefox fighting the DRM for so long. It pushed many users and projects towards Chromium.
I wish they'd kept it up myself. It's not as if folks can't have more than one browser for different things. Now that even the W3C has sold us all out we're pretty much doomed to get screwed over by DRM at some point in our lives online.