Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Chromium and the Browser Monoculture Problem

dev.to

Post reply on HN
Print view View original

11–20 of 128 posts

Re: Chromium and the Browser Monoculture Problem

#11

Post by roca »

Some major problems with the "let's all use Chromium" approach:

Google controls Chromium because they do the heavy lifting. All the companies that use Chromium are using it because they want Google to do the heavy lifting, so that isn't going to change, even if Google decides to put an open governance face on it, and there's no sign they're going to do even that. Should Google really be ceded complete power over the Web?

A single browser codebase hurts the maintainability of the Web platform. To the extent Chromium has a monopoly, Chromium bugs are the standard and sites will depend on them. Yes, this is already true to some extent, but it could get much worse. So fixing bugs would get harder.

A single codebase makes radical implementation experiments more difficult. Companies can patch Chromium but architectural divergence gets expensive fast. Say you want to rewrite the Chromium style system for parallelism with Rust, like Mozilla did with Firefox, and Google says no? Good luck maintaining that long-term. (A good example was how JSC vs V8 caused endless friction between Google and Apple in Webkit, leading up to the Blink fork.)

Another problem is that it would be very hard to block Web features that Google implements in Chromium. (P)NaCl/Pepper is a good example. "It's a ton of work to implement (especially if there's no spec)" is a helpful argument for not implementing a feature that's bad for the Web, both in public and to one's own product management. Hard to hold the line on something like PNaCl/Pepper or WebSQL if it's just a matter of flipping a configuration flag to increase Web compatibility. If the world was all-Chromium a few years ago there's no doubt we'd have PNaCl/Pepper instead of WebAssembly --- and we'd be stuck with the insane duplication of Pepper vs standard Web APIs, the bloat of LLVM everywhere, etc.

Re: Chromium and the Browser Monoculture Problem

#12

Post by philjohn »

potatocannon wrote: post #8
danShumway wrote: post #6

Ultimately, Google will still decide what goes into Chromium. As it stands, Google decides what patches get merged, and Google decides how core features are implemented. And they can be implemented in ways that are very hard to toggle and very specific to Google's business model. We need to get a version of Chromium that's not owned by Google -- but, I don't see any motivation for Google to give up control of Chromiu…

Good thoughts, except for Firefox 'not doing great right now'. I've switched over for almost a year now, both for regular and development use, haven't looked back.

Same - I moved to Safari when I got sick of Chrome making my MBP sound like a jet engine (and be almost as hot to the touch).

Then I got tired with certain sites not working properly and went back to Firefox - and it's plain to see they've done great things to the browser engine, it's fast, pretty lightweight on CPU and power usage and all of the problematic sites in Safari work fine.

Re: Chromium and the Browser Monoculture Problem

#13

Post by kevsim »

danShumway wrote: post #6

Ultimately, Google will still decide what goes into Chromium. As it stands, Google decides what patches get merged, and Google decides how core features are implemented. And they can be implemented in ways that are very hard to toggle and very specific to Google's business model. We need to get a version of Chromium that's not owned by Google -- but, I don't see any motivation for Google to give up control of Chromiu…

We've at least seen some recent examples around browser vendors like Brave and Vivaldi deciding not to take some of Google's Chromium changes around browser extension handling [0].

So while this is not as compelling as a community controlled fork, it at least shows the alternative vendors are not willing to blindly swallow Google's changes.

[0] https://www.zdnet.com/article/opera-brave-vivaldi-to-ignore-...

Re: Chromium and the Browser Monoculture Problem

#14

Post by adrianN »

kevsim wrote: post #13
danShumway wrote: post #6

Ultimately, Google will still decide what goes into Chromium. As it stands, Google decides what patches get merged, and Google decides how core features are implemented. And they can be implemented in ways that are very hard to toggle and very specific to Google's business model. We need to get a version of Chromium that's not owned by Google -- but, I don't see any motivation for Google to give up control of Chromiu…

We've at least seen some recent examples around browser vendors like Brave and Vivaldi deciding not to take some of Google's Chromium changes around browser extension handling [0]. So while this is not as compelling as a community controlled fork, it at least shows the alternative vendors are not willing to blindly swallow Google's changes. [0] https://www.zdnet.com/article/opera-brave-vivaldi-to-ignore-...

Not merging upstream changes eventually kills smaller vendors with maintenance burden as the diffs grow larger and larger.

Re: Chromium and the Browser Monoculture Problem

#15

Post by ximm »

roca wrote: post #11

Some major problems with the "let's all use Chromium" approach: Google controls Chromium because they do the heavy lifting. All the companies that use Chromium are using it because they want Google to do the heavy lifting, so that isn't going to change, even if Google decides to put an open governance face on it, and there's no sign they're going to do even that. Should Google really be ceded complete power over the…

> Google controls Chromium because they do the heavy lifting

Google already controls the web standardization process with its manpower. So standardizing on chromium would not change that issue, it would just move it to another level.

I am not saying that this is great. I just think the argument is invalid.

Re: Chromium and the Browser Monoculture Problem

#16

Post by srikz »

jpochtar wrote: post #2

This is a good take. Standards were a weapon to fight Microsoft's closed source IE— emphasis on closed source. Why? Competing browsers needed to be IE-compatible to matter, and without standards spec-ing out what IE's doing, it's hard for other vendors to know if they were 100% compliant. Now that the major browser everyone's coding against is open source, and competitors can freely borrow their engine, I'd say we do…

I used to feel the same way but I think there should be competing implementation of standards to make faster progress.

GCC vs Clang is what comes to mind. The competition from clang improved the overall quality and performance of c / c++ compilers in general

Re: Chromium and the Browser Monoculture Problem

#17

Post by lazyjones »

roca wrote: post #11

Some major problems with the "let's all use Chromium" approach: Google controls Chromium because they do the heavy lifting. All the companies that use Chromium are using it because they want Google to do the heavy lifting, so that isn't going to change, even if Google decides to put an open governance face on it, and there's no sign they're going to do even that. Should Google really be ceded complete power over the…

> A single browser codebase hurts the maintainability of the Web platform. To the extent Chromium has a monopoly, Chromium bugs are the standard and sites will depend on them.

It's not the single browser codebase that hurts the web, any browser with noticeable (maybe 3-5%) market share causes these problems. The main issue is that today, the web as a platform is garbage. It's full of cruft, mostly useless features that cause browsers to be bloated, unmaintainable codebases nobody except 9-figure-budget entities can hope to reimplement. It's an unstable, far too complex platform that wastes billions in unnecessary maintenance work for web developers.

Much of this could have been avoided by knocking some sense into the entities that constantly propose new features and those who implement them as native browser elements instead of JS libraries where possible. Then there's the insanity of preserving unsafe behaviour as default and adding response headers and extra HTML tag attributes to avoid it, so broken sites could stay broken and everyone else had to do maintenance work. I'm not saying Google is the driving force behind this, but it reeks of the same stuff that Microsoft was doing to prevent Word, Excel file formats from becoming readable and writable for the competition.

It's too late to freeze the web platform, it needs to be replaced with something sane and stable, controlled by people with no ulterior motives.

Re: Chromium and the Browser Monoculture Problem

#18

Post by saulrh »

lazyjones wrote: post #17
roca wrote: post #11

Some major problems with the "let's all use Chromium" approach: Google controls Chromium because they do the heavy lifting. All the companies that use Chromium are using it because they want Google to do the heavy lifting, so that isn't going to change, even if Google decides to put an open governance face on it, and there's no sign they're going to do even that. Should Google really be ceded complete power over the…

> A single browser codebase hurts the maintainability of the Web platform. To the extent Chromium has a monopoly, Chromium bugs are the standard and sites will depend on them. It's not the single browser codebase that hurts the web, any browser with noticeable (maybe 3-5%) market share causes these problems. The main issue is that today, the web as a platform is garbage. It's full of cruft, mostly useless features th…

> controlled by people with no ulterior motives.

Everyone used to think that was Google.

Anyway, if by some miracle you do find anyone that matches that description but knows where the lever is to kick everything over to a new paradigm, don't tell anyone about them, because as soon as you do they'll have more ulterior motives then you can count.

Re: Chromium and the Browser Monoculture Problem

#19

Post by Tinfoilhat666 »

jpochtar wrote: post #2

This is a good take. Standards were a weapon to fight Microsoft's closed source IE— emphasis on closed source. Why? Competing browsers needed to be IE-compatible to matter, and without standards spec-ing out what IE's doing, it's hard for other vendors to know if they were 100% compliant. Now that the major browser everyone's coding against is open source, and competitors can freely borrow their engine, I'd say we do…

W3C failed to create an open platform. Browser vendors use W3C as a rubberstamp for their implementations. There's many W3C Notes and Editor Drafts with only one vendor. Just look at the latest Google firehose of new features.

Re: Chromium and the Browser Monoculture Problem

#20

Post by lazyjones »

saulrh wrote: post #18
lazyjones wrote: post #17

Earlier quoted context omitted.

> A single browser codebase hurts the maintainability of the Web platform. To the extent Chromium has a monopoly, Chromium bugs are the standard and sites will depend on them. It's not the single browser codebase that hurts the web, any browser with noticeable (maybe 3-5%) market share causes these problems. The main issue is that today, the web as a platform is garbage. It's full of cruft, mostly useless features th…

> controlled by people with no ulterior motives. Everyone used to think that was Google. Anyway, if by some miracle you do find anyone that matches that description but knows where the lever is to kick everything over to a new paradigm, don't tell anyone about them, because as soon as you do they'll have more ulterior motives then you can count.

> Everyone used to think that was Google.

When was that? Around the time they decided to liberate the web from the evil clutches of Microsoft/IE and ... Mozilla/FF (20%+ and growing at that time)?

Post reply on HN