Chromium and the Browser Monoculture Problem
dev.to
dev.to
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.
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.
Sentiments such as "Google controls everything, there is nothing we can do" are inaccurate and also counterproductive.
Nowadays for most consumers it is whatever Safari or Chrome support.
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.
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.
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)?
We very well may be assuming different actors and motivations where there are none: there's just messaging and market differentiation.
When people still believed in the slogan 'Don't be evil' [1]. Google even removed the slogan. I guess they don't believe in it themselves anymore. The replacement 'Do the right thing' seems like a variation of Nike's.
You could have some low level thing like web assembly (or even just we assembly itself and get all of the current optimization and standardization work for free) that other languages compile too. And a low level GUI library like Vulkan that you can build libraries on top of.
That's what Google did with the shadow DOM they used on YouTube: since Firefox had no support for it they wrote a polyfill (which is a JS library to emulate a feature) and the resulting performance was terrible.
No. When Web devs test in multiple browsers and find that something doesn't work in one of them, they will usually pick an alternative solution that does work in all of them, avoiding the bug in the buggy browser, thus avoiding a situation where their site depends on that bug continuing to exist. It doesn't always work out as well as that, but it often does.
You can wish for something better than the Web platform, but it's a tall order. You want a platform that's not controlled by a single vendor, that has multiple open-source non-patent-encumbered implementations with significant market share, that is available across just about every device, that has a vibrant content/application ecosystem and developer community, that supports everything from simple static content to extremely complex applications like office suites and 3D games. Good luck with that.
Who really needs that in the browser? Native apps are consistently better in every single aspect - more powerful, more privacy-respecting, more eco-friendly, etc. Except native apps are vanishing - a month ago I wanted to fire-up SketchUp on my old Mac, only to find out that the casual 3D scene is now using 100% webapps. The reason we have these abominations is that they're one of the only avenues left for financing app development (either ad-driven, or in a fremium model) because nobody buys software anymore. And that is indeed a contradictory effect of making software a commodity through F/OSS when the original motive was to empower users. Now I know folks chase browsers and wasm and whatnot as the ultimate cross-platform app delivery platform (coming from a 1990s MS and 2010s iOS app dominance experience) and make browsers even more complex. However I whish they could stop and think about if their work is actually helping the situation.
What I think would be helpful is to define a base web profile for content delivery, and then additional profiles for webapps. The "march of progress" argument against versioning and profiling HTML, CSS, devices, and JavaScript we've heard from web standardization efforts isn't convincing at all, and only helps the big guys (basically, only Chrome at this point).
Enterprise IT departments, for one. IT people hate having to install, maintain, update and troubleshoot desktop applications, for obvious reasons. It’s orders of magnitude easier to simply install the app on a web server and have users run it in their web browsers.
I question how many developers do that. Especially for smaller sites if it works in Chrome it's generally good enough.
• Office Suites
• Search engines
• link sharing sites
• internet runtimes
that all do roughly the same thing?
Maybe we do. Maybe Google should be broken up into a bunch of smaller Google's and be forced to interoperate with smaller search companies like DuckDuckGo, Qwant and TinEye. Like MVNOs but for search engine infrastructure.
The standard we should strive for is to not lock any given 3rd party in to facilitate data exchanges. If I want to email someone, I shouldn't need to coordinate with Google. If I want to email a document I shouldn't need to coordinate with Microsoft and if I want to browse the internet I shouldn't have to coordinate with the Mozilla foundation.
We've seen this play out a few times; data lock-in is not properly appreciated even within the tech industry, let alone outside it. The risk of making Google a mandatory part of exchanging data between two parties on the internet is serious - if they ever manage to insert themselves in the same way IE got its claws into South Korea back in the day that will be decades of foolish pain for no gain. Great idea to keep people aware of the threat.
What happens if Google shut down Gmail outside the US with no warning and no big export of data? Maybe because the US government decides foreigners need a bit of prodding because of a trade war? That would be a Nasty Shock.
If instead of Blink we had Blink AND Gecko, instead of Chromium bugs being the standard, Chromium AND Firefox bugs would be the standard. How is having to implement workarounds for TWO engines worse than having to implement workarounds for ONE engine?
Your last line doesn't follow from the rest of the paragraph. One platform is easier to fix than 4. And it makes it much easier to develop for. Either your site works on Chromium or it doesn't work. You don't get into a situation where it works for 80% of your users.
>Say you want to rewrite the Chromium style system for parallelism with Rust, like Mozilla did with Firefox, and Google says no?
I'd be concerned about that if Chromium hasn't been eating everyone's lunch for nearly a decade now. In terms of rendering/js engine they're clearly ahead of everyone else. For all of Mozillas efforts they still haven't matched chromium let alone surpass it.
However, the only reason this is possible is because we have web standards that are independent of a single browser and vendor. If we end up abandoning the standards entirely in favor of just using Chromium, then it’s game over for open web.
Throwing in the towel and making everything Chromium based would be the worst possible scenario because it gives Google an insane amount of control over the future of the web. And the fact that Chromium is open source doesn’t really make a bit of difference here.
The reality is that it takes a huge amount of effort to maintain it. It would require an organization like Mozilla to fork it from Google and take it in a new direction. Since we already have Mozilla and Firefox now, it’s much better to make sure they continue to exist.
- IE was a Windows-only product (after Microsoft dropped Mac support), so as Windows lost its dominance (thanks to the internet, and mobile devices, and Linux, and Apple's rebirth, and...), so did IE. It helped sow the seeds of its own destruction.
- Chrome has an open-source core, so they're going after the embedding market much more than IE ever did. That means that a lot of "Chrome" use is in places where users feel the consequences, but have no control over being able to switch engines (e.g., Electron apps).
Google's web browser group feels like the classic "big, smart" quadrant competitor that nobody will engage. If your only goal were "take over the world usage share of browser engines", it's hard to imagine doing a better job than they're doing now.
The client-side web was not traditionally Microsoft's home. They were a little late to the game, they didn't quite understand it, and they needed a good browser on their platform but it wasn't fatal if it wasn't theirs. (They're doing quite well in services and games and many other areas.) Google was always web-first, and they will never cede that.
Crazy tech prediction: Chrome will be the last dominant web browser engine. What defeats it will be the successor to the web, not another "web browser". Its successor will probably even embed Chrome for backwards compatibility, like Edge embeds IE, or Chrome embeds FTP. We'll look at legacy Chrome-compat internet sites the way we look at Electron apps, or non-TLS websites.
Then, along came Firebird, and it was light and fresh. Unfortunately, a lot of sites didn't work with it. But users demanded that it should work, and eventually it did.
Eventually Firefox itself became bloated and Chrome was the faster alternative. It's the guys at Mozilla that dropped the ball this time.
Fast forward to now: Chrome is bloated, everything is bloated, everything is terrible, but we now rely on a tech stack that has grown exponentially just to get really basic stuff working. The difference is, this time we can't reasonably start fresh anymore without breaking backward compatibility.
Perhaps that's what needs to happen though: Break everything. Delete HTML, CSS, Javascript and all that crap. Admit that most non-trivial web sites are not documents, they're applications. Applications can display documents and do all kinds of other things. Figure out the minimal API that can efficiently output graphics, sound and video. Add in some accessibility hooks. Provide some optional interfaces to geolocation, storage, camera/mic etc. Done.
Fuck the semantic web. Just let us write applications.
This is an important point. One thing that led to IE's downfall was its lack of standard compatibility. At the time, there were real web standards. Today, the "standard" seems to be whatever Chrome does. (Indeed, despite good standard compatibility, the pre-Chrome version of Edge suffered in terms of real-world compatibility, and that was an animating reason for the switch to Chromium: https://blogs.windows.com/windowsexperience/2018/12/06/micro...) So we are very differently positioned than we were when Firefox/Chrome overtook IE.
I would agree, but not for the reason you state. IE was the standard back then, too. The difference is that once it was "good enough", Microsoft reduced investment in it and allowed it to stagnate. Others caught up in terms of compatibility and passed it in terms of usability.
Google, on the other hand, invests so much that no on else can keep up, and knows that this is the means by which they maintain control over the web (and web standards).
Nowadays, we're asking people to choose, for all appearances, a slower browser that's not as technically advanced in some ways, with no clear merits and a lot of changes to workflow. The choice today is between freedom and convenience. The reason people flocked to Firefox in the first place was that it was both freedom and convenience.
It's unclear to me Chromium is still significantly faster in any meaningful way. They certainly feel roughly the same.
The reason we switched from IE was because IE just straight up did not work. Sites were broken. Now it doesn't really matter but for these invisible things that normal people don't see or care about. Most people are like my dad.
Perhaps sites aren't as broken in Chrome now as they were with IE, but you can bet the web itself will change in ways you and I would consider broken if Google is left to become the sole, unchecked steward and helmsman of the web.
In what way do you still feel Chrome is better/faster?
I was actively using chrome, chromium, and Firefox for a long time. This was probably more because of what I do for work than a personal preference. More recently it’s been 98% Firefox since Chrome has just added too many weird things.
Is this still true with modern Firefox?
Personally I don't see a noticeable performance difference between Firefox and Chrome.
Netflix on (pre-Chromium) Edge, really is insufferably slow. General browsing in Firefox though seems perfectly fine; it's not purely an ideological charity-case.
Anyway the adblocking thing isn't really related. Extension APIs are not industry standards, Chrome already doesn't allow adblocking on mobile, and maybe at some point they'll just write off the users who care and focus on monetizing the users who will put up with ads.
Chrome is business for google and obviously will not want to change the situation, but also people with business on the Internet what want is a friendly and easy environtment which is what google is offering.
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 Chromium or move Chrome over to use a different, community controlled fork.
So if you want a version of Chromium that really is community controlled, you'll need to have a community fork of Chromium that we all agree to work on and use as a browser base. Otherwise, we'll just be giving free labor to Google, and they'll still filter contributions so that primarily their interests are served.
Unfortunately, now that we've forked Chromium for all of our browsers, one of two things is going to happen:
- either the fork will need to be tightly constrained by Chromium, in which case Google will still effectively control it, or
- the fork will diverge from Chromium, in which case we'll be back to the exact same position we are now. You'll have multiple competing browser engines, one of which is used by the most dominant browser on the market.
The TLDR is if you're worried about Google controlling the browser engine, it still seems to me that the only solution is to get a non-google controlled browser engine used by a browser with market share comparative to Chrome.
The relative lack of user adoption for Vivaldi, Opera and Brave suggest to me that standardizing on Chromium won't significantly help with that. Edge is now moving over to Chromium -- if it sees a massive influx of new users, maybe it'll be worth taking that strategy more seriously.
In the meantime, since it appears we are going to need a browser engine to compete with Chromium anyway, what's the harm in having one company (Mozilla) try to compete by building their own browser engine the way they want? Sure, Firefox isn't doing great right now, but is it doing any worse than Brave?
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.
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-...
Maybe in terms of market share as a proxy for success that is true.
As a browser though Firefox has come on leaps and bounds (and has momentum to keep improving), I never switched away from it but used Chrome for dev for the last few years so I have a fair amount of time with both (dozens of hours a week) and Firefox is as fast and reliable as Chrome.
Ironically the one property that gives me issues with Firefox is youtube, they don't fix bugs against Firefox (the white bar header on fullscreen has been in firefox for over a month now and has been reported).
I ended up having to implement a ublock origin rule to remove the header.
EDIT: Mozilla are (to me) so much more than just a browser vendor, MDN documentation is probably my 2nd most visited documentation/help site (behind only stack overflow) because the quality is superb, concise and they quickly mention at the bottom can do this on foo, bar but not fizz which is useful to someone approach 40 because keeping that stuff in my head is well beyond me these days.
Firefox is my daily driver, and it's great.
This might be just me being pedantic, but I thought it might be worth mentioning...
Chromium is basically the same thing as Chrome, except that it's missing a few proprietary blobs that Google is not legally allowed to put into Chromium: some DRM stuff, a licensed H.264 codec, a Flash plugin, etc. But if you're worried about privacy, Chromium still has all the Googley stuff. You can still log into Chromium with your Google account, and it'll sync your bookmarks and history and passwords just like Chrome. Chrome extensions are 100% compatible with Chromium, and you can install them into Chromium via the Chrome Web Store.
Some people might be familiar with "Blink", which is the browser engine (and is a fork of WebKit). There's also V8, which is the JavaScript VM. These two components aren't enough to form a usable browser though. So the Chromium Project develops what is called the "Content module" (see https://www.chromium.org/developers/content-module) which combines Blink and V8, and some other pieces, to form a basic web browser. This is what other browsers (like Opera, and I presume Edge as well) are built on top of, and does not contain all the Google integrations that Chrome/Chromium have.
While Content is missing a lot of the Chrome features that make privacy advocates nervous, it's still a problem because most Chromium/Content developers are Googlers, and will probably not implement things like WebKit's ITP or canvas fingerprinting prevention.
So it's nothing special.
I’d like to see the privacy conscious “user agency” browser builders pivot back to WebKit.
> Ideally, this codebase would not be controlled by any single company. I'd love to see a common browser engine controlled by an independent nonprofit foundation
I had a bit of a facepalm moment. This is exactly how Google started out: Chrome(ium) was built based on webkit, so basically there was a mini consortium where Google and Apple would work together. But that wasn't moving fast enough for Google's taste so they forked off blink so they could immediately do whatever they pleased.
It was a relationship which ended on good terms. Your comment makes it sound like it was antagonistic.
If we switch to an open source browser engine monoculture we lose competition among rendering engines, and may never get something much better than what we have now.
Then again they would have to work pretty closely with Google on integrating Servo with Chromium. Otherwise they'll get into the situation that Google did with apple over Webkit and then they be at square minus one with a foreign codebase.
We need concise well-engineered and precisely-defined web standards a small team can implement in reasonable time. Many people would build competing browsers if there were yet we have what we have and no choice but to stick with the Google engine.
That the features aren't proprietary is besides the point: Google realizes that they don't need them to be proprietary, what matters is that others can't easily implement them all.
Now that the major browser everyone's coding against is open source, and competitors can freely borrow their engine, I'd say we don't really need the W3C anymore.
Chrome isn't open source, it's open core.
GCC vs Clang is what comes to mind. The competition from clang improved the overall quality and performance of c / c++ compilers in general
- IE vs Chrome dominance is comparable but not equal. One was closed source and stagnant, the other open and constantly evolving.
- “What if we all contribute to the same code base” ignores that not all priorities are aligned. Apple is a lot more focused on user experience and battery life while Google is all about moving the platform forward, see Web Worker conundrum for example.
- Browser are a lot more complicated now than they used to be, so creating anew from scratch is harder. But the Web is also a lot better documented (edge cases specified) and with more open implementations to get inspiration from.
I doubt the problem is too few developers. Google has over $100 billion, it can hire as many people as it wants. And didn't you already say at the beginning, "Google is huuuuuuuuge. They just have so many people working there."
Rather I wonder if there are just only so many people who can collaborate on a codebase, there are just only so many features you can push through at a time. If all the Firefox developers moved to Google, I doubt we would see features faster. I bet we would see features come out at the exact same pace. It is not a matter of labor but of leadership, and each project will prioritize features differently. Therefore this is an argument for more than one implementation.
Is this not the subject of the famous book The Mythical Man Month, by Fred Brooks? Adding more programmers to a late project makes it later. I know that Google is a massive company, but I wonder how many employees literally write code for Chrome.
Given history it is also personally interesting to consider Microsoft as a company that could have furthered open web standards.
You're wrong and sorry to be blunt. If you think Chromium is pure technical code and no Google hooks, you're wrong.
1. I've been working on it for over a year and it's rife with Google hooks. 2. When we have a thriving project named Ungoogled Chromium, you'd start believing point 1.
for me it isn't...
I hadn't given it much of a thought since edge/ie are such a non starter usually. Would be interesting to look into!
Not happening. It's likely that Apple spends an inordinate amount of engineering on Safari to lock-out W3C web features that pose a competitive threat to their lucrative Monopoly: the app store. Their strategy has always appeared to be to purposefully cripple Safari to drive developers and users to the app store, and monetize there, both on the $99 annual fee for the developer's ability to publish apps, and that 30% cut on revenues from publishers. Then there's the whole arbitrary bans on certain apps that somehow undermine their control over the device and ecosystem.
If they didn't have a plan to cripple Safari to drive people to the store, why else would they ban browser engines from competitors?
When IE first caught on it was extremely fast and pleasant to use compared to Netscape (also netscape wasn't originally free). If you only cared about IE it was actually a complete breeze to develop for using proprietary stuff like ActiveX (ignoring the fact that ActiveX was a bad idea for various reasons).
Lots of websites started showing banners saying that they were optimized IE, because people were primary targeting IE when they made new websites, in part because it was such a breeze to develop for. IE had lots of HTML/javascript features that made things much easier than other browsers at the time.
It's important to realize that IE only became difficult to deal with once 1) Microsoft essentially stopped trying to improve it after they won all the market share and 2) people needed to make web pages that worked on other browsers. At this point people wanted to use newer features that IE didn't support, but needing to support IE 6 necessitated either abandoning more modern features or using all sorts of hacks/polyfills to work around IE's quirks.
I think they're using skia all the time since 2016.
We will see how North Korea turns out in the future?