exactly. it's like money laudering, only for bad anecdotal information, or conspiracy theories. I'd sooner invoke Hanlon's razor than a single div tag that Edge's hardware acceleration should have been robust enough to handle anyway.
A lot of statements for and against this or that company having stewardship of the web, but the more important point is we should, IMO, strive for a web which cannot be controlled by any one entity. In this view, whether Google is a 'better evil' than MS or Exxon controlling the web is beside the point. The web has a good chance to approach the ideal of direct democracy rather than a representative one. Sure it might be a long road, but let's not give up and concede to the established paradigms on the web too, though some will say it is already too late.
Maybe Google should split Chrome and related assets (Chromium, Blink, are there others?) into a separate foundation that has control over the source with a board that has membership by major players such as Google themselves, Opera, and Microsoft. That way Google can put a wall between their direct influence in such a powerful and relied-upon piece of software so as not to have the undue influence we’re all worried about.
Android's governance is entirely under Google/Alphabet for all practical purposes today, even as the OS is open source. Reasons for this are numeorus (licensing, Play Services, etc.) Chromium is in a similar boat. Parent is proposing that Google effectively spin off Chrome and related assets as a separate company or nonprofit entity, much like the Linux Foundation, to avoid potentially monopolistic behavior that such vertical integration may lead to. I think this is one possible way to a solution.
Though please not exactly like the Linux foundation, which is doing some pretty shitty things. Companies can become members of the Linux foundation and violate the GPL of Linux no problem, they are paying their membership fee after all. The Linux foundation is effectively the "Megacorps Linux user group". Its executive director Jim Zemlin famously doesn't use Linux himself.
His arguments are really weak. He uses both SPDY and QUIC as examples of Google behaving badly. From my perception, Google basically said "this major web standard is inefficient. If we make these changes, the network is faster (especially for people with poor connectivity." They validated those hypotheses with real traffic on real devices over real networks, and then proposed standards based on that work. The standards bodies started there, and made adaptations until there was consensus, and new standards were minted.
As another example, YouTube uses a feature called HTML imports to load scripts. HTML imports haven't been widely adopted, either by developers or browsers alike, and ECMAScript modules are expected to serve the same role. But they're available in Chrome and used by YouTube.
At least in this particular case my understanding was that Google was pushing for the adoption of HTML imports as a web standard, which would have been better for all browsers, but other browser vendors balked and Google was forced to use JS module imports instead. AFAIK I know they are now discouraging any further use of HTML imports. This is a bummer because it's not possible to leverage web components without JS now.
I'm concerned about a web monoculture too but so far Google has mostly done a good job IMO. Even Mozilla has done some pretty dumb things when it comes to web standards (unilaterally killing WebSQL, for example). I certainly can't think of any other big tech company I'd prefer to see calling the shots on the web.
Yeah this was the argument at the time. I still think it would have been better to formalize that spec instead of just pull indexdb out of some unmentionable orifice.
The problem is that maintaining a browser engine is enormously expensive. There are so many features you need to maintain, security issues to deal with, etc. In the past Microsoft made money off of their operating system, and they made Internet explorer then Edge off of that, but now they make money off of the cloud and clients expect people to support Chrome. Microsoft was spending too much money they didn't need to for return. I think because of the expensiveness of maintaining a browser engine used by so many people, they were destined to be basically a natural monopoly operated by a few concerned actors. Of course, google as one of the two major internet giants, is the most concerned with supporting the development of a web browser.
> even if Google takes Chromium in a direction that Microsoft disagrees with or opposes, Microsoft will have little option but to follow along regardless.
I keep seeing arguments like this and it doesn't make sense to me. Help me understand. My impression is that since Chromium is open-source Google cannot have a monopolistic wield over it. And now the fact that a bunch of Microsoft engineers will be contributing to Chromium seems to imply that Google and Microsoft will have the same amount of muscle mass to struggle for control. For example if Google tries to make a change in Chromium that will break Microsoft's products Microsoft can quickly respond to this. So why is the argument that Microsoft is ceding the web to Google, rather than trying to wrest back? To me it seems like the latter is at least as plausible as the former.
The problem is this, who is seen as controlling the standard?
Perception is reality here.
Right now people get updates to Chrome and ChromeOS from Google. Microsoft is free to fork it. But if Microsoft goes one way and Google another, Google's version is what gets used. And everyone knows it. So anyone who wants to influence the direction of Chromium will talk to Google. Anyone who wants an idea of what is coming, will ask Google. Microsoft is a distinctly second class citizen.
Doubly so since Microsoft's last two attempts at a browser (IE and Edge) are now failed also-rans. If it attempts to fork Chromium, nobody is going to believe that their third try will win.
In its younger days, Microsoft fought hard to avoid being in such positions. For a flavor, read http://gunkies.org/wiki/Gordon_Letwin_OS/2_usenet_post to see how they thought about ownership of standards. But Microsoft has decided to cede this particular battle. At least for now.
Interesting bit of history. Thanks for this article. I'm curious, though, how the fact that Chromium is open-source will change this dynamic, because the ethos of open source seems directly against the idea of Google abusing their control over Chromium.
Firefox is Open source. Does Mozilla control what standards get adopted on the most used platforms? Or which ones influence your search ranking and traffic?
No, because Mozilla don't own them, Google does.
The problem is not Google having a browser, open or closed. The problem is all the influence Google has elsewhere.
Do you mean rolling a new protocol in Chrome without warning that magically makes Google faster because they rolled it out on their end as well? This already happened with https://en.wikipedia.org/wiki/SPDY.
As long as what they do is technically good and not too heavy-handed, developers won't resist. In fact it is likely to be welcomed. But the effect will constantly leave Google a half-step ahead of everyone else on the upgrade train. After all it isn't hurting developers, it is just leaving Google with a slightly better competitive edge.
Google effectively has the right to veto patches into Chromium. In the worst case, Google could fork Chromium, and since Google controls the build artifacts, in a dispute between Google and Microsoft, it's Google who wields the power.
How is it Google? Anyone could just fork Chromium and do anything about it, it's open source after all. Microsoft could buy people from Google if they wanted to, they probably could buy an entire Chrome team. Google has some control over Chromium engine right now, but this control is fragile. It's the same control that Oracle has over MySQL. Abuse it and you'll have MariaDB instantly. Abuse Java control and you'll have plenty of companies forked Java and ready to provide real LTS for Java 8. It'll be the same for Chromium.
The best outcome would be to have some Non-commercial organization funded by Google, Microsoft, Yandex and other big players who benefit from Chromium which would have some kind of independence. But current situation is not that bad.
There are other things than just source code license. May be Google has patents for Chromium sources and could use them to forbid fork? Then it's a bad situation but I didn't heard about that point of view yet.
Your examples of successful forks are due to branding amongst technologists, who understand the concept of a fork and who were following the most-trusted of the descendents (another example being Hudson and Jenkins).
Consumer brands are not as easy to fork. If Coca-Cola opensourced their recipe and then faced a fork by Examplesucre, I don't think anyone will really be betting against Coke.
Similarly, the longevity of mainline integrity for Linux relies heavily on the fact that it is a trademark and that the trademark holder (Linus) is trusted with it. Fork Linux as much as you please, basically nobody will care.
Google's power over Chromium comes from its power over Chrome. If you don't think that's true, fork it yourself and prove us all wrong.
> Fork Linux as much as you please, basically nobody will care.
Last I heard, Android was doing OK…
Similarly, I should think that Microsoft still has considerable clout in what browser Windows users will use by default (although I'm not sure what exact legal restrictions they currently have to observe re: bundling of browsers).
Because a majority of users are using Google Chrome.
You're right, if Microsoft could convince chrome users to switch to Edge, they would have more power. Things can change, but I don't see how this would happen.
But, isn't it reasonable to assume that microsoft will get hold of the Whole codebase if they embrace this as their new default browser?
I mean its Microsoft we're talking about..
Well, it's likely they'll only use the core rendering engine (Blink, formerly Webkit) and wrap their own application around it. Hell, with all the applications they've released or are working on with electron (Teams, VS Code, SQL Operations Studio), I'm far less surprised.
I've been hoping they'd come up with for a more shared model for applications in the node/npm ecosystem similar to Adobe's Air. If MS did a good job of such a platform from open-source with tooling to build applications that run everywhere, more power to them.
Maybe I'm naive but it seems that if a fork happens, a team or two of good Microsoft engineers could keep the fork up to date with features from the original branch.
Chromium is an incredibly complex codebase, and they change their internal API's quite rapidly, so upgrading is a lot more difficult than you would guess. For example, this is what it took for Electron to upgrade from Chromium 69 to 70: https://github.com/electron/electron/pull/15405/commits, and Electron doesn't even make that many modifications to Chromium. If Microsoft were to make a decent number of changes, they would need a whole team of people just to import Google's changes, and if they start adding major new features I doubt they'd be able to keep the two in sync at all.
I would do it if the money was right but there is also the risk that Google may make the codebase intentionally difficult to port to the fork. I don't think such an effort would be sustainable in the long run.
It's not an easy task indeed, but it's very feasible. We have some examples of that kind. Firefox' fork PaleMoon is doing well, thanks to individual efforts of approximately one man. RedHat used to backport a ton of features to their Linux kernels across the major version numbers.
That doesn't mean anything. They can and will change course depending on the circumstances, and the community could agree with them depending on what the circumstances are, effectively making it a non-issue.
No, it does mean something, in that it affects Microsoft's entire product strategy (i.e. deinvesting in the browser). I don't see Microsoft changing course on this.
People want innovation but ultimately they don't want companies to profit off them. In the beginning, everyone loved Google. Now that their innovation gap has increased so much over their peers for some areas, the consequences of that are starting to sting. I am not sure what people would prefer: for them to be less competent or to not to profit from it?
The truth is Chrome and their web team has been amazing in advancing and supporting web standards. (except AMP which is pretty bad in some aspects)
I think a less cynical way to phrase your (correct) idea is that people love the promise of profit as an incentive to create innovation. It's not a fully self consistent system. But the end goal isn't profit imo, it's innovation, and competition via promise of profit is a way to efficiently allocate resources to innovate. For what it's worth, I think the general arc of antitrust rulings bend toward the ideal over time and it's a fine enough way to deal with the self inconsistency.
Don’t forget where Chromium came from. KHTML lead to Safari lead to Chrome which lead to Chromium. It’s been forked and obounced from open source to corporation to corporation to open source.
The original KHTML base that was picked up by Apple and turned into WebKit was LGPLv2, which is why source was ever available at all. Other stuff has surely been added under other licenses, and I've certainly not audited it. But I'm sure the copyleft portions are still in there.
Also to note on the engines used it went from khtml to WebKit and then google forked WebKit into Blink. Also opera switched from writing their own engine and moved to chromium .And now it seems the next version of edge will also use Blink being derived from chromium. Which puts everyone in a position that on all platforms you are using WebKit derived browsers (also mobile is in the same boat with WebKit powering both Chrome and Safari.
Sure they allstand on the shoulders of giants, however you are mixing technologies a little bit. A browser is a lot more than an html engine, there is almost nothing left from khtml in either Chrone or Safari.
Apple ported KHTML to Mac OS, calling it Webkit. It was released as part of the initial Safari release in 2003. They had a breakup with KHTML a few years later and open sourced Webkit separately. Webkit then became the basis for Chrome when it was launched by Google, until they forked in turn and made Blink.
A browser has many other parts than rendering engine. So it is not fair to say Chrome is derived from webkit or safari. You can say chrome rendering engine was initially based on webkit which was based on khtml. And today that shared parts are quite small in the general picture.
Really cool comic! Thanks for sharing it. Very useful dive into some details I've managed to not know about until this present moment. Any other internal tooling/stack related comics for useful Google products that are worth mentioning in your opinion?
They are fine until you hand it over to them and then they are a problem :)
The pragmatic answer is firewalls between layers (i.e. browser separate from search), and competition in layers, plus decent legislation, plus some degree of decentralization.
> The pragmatic answer is firewalls between layers (i.e. browser separate from search), and competition in layers, plus decent legislation, plus some degree of decentralization.
I’m seeing this in other businesses too. Hell- some airlines now run ads on their inflight tv systems right after their safety videos. You, as a passenger who has paid their fare for a flight, must sit still, strapped in, and watch an ad on a screen straight ahead of you, with sound playing through the cabin, with no way to opt out. Ads are a disease.
> You, as a passenger who has paid their fare for a flight, must sit still, strapped in, and watch an ad on a screen straight ahead of you, with sound playing through the cabin, with no way to opt out.
If someone in the seat next to me decided to opt out by colouring over their screen with a sharpie, I would not stop them.
I’m seeing this in other businesses too. Hell- some airlines now run ads on their inflight tv systems right after their safety videos. You, as a passenger who has paid their fare for a flight, must sit still, strapped in, and watch an ad on a screen straight ahead of yo, with sound playing through the cabin, with no way to opt out. Ads are a disease.
Telemetry is not really a problem except to the paranoid.
Not only is it anonymized, it's very useful. It helps provide essential data for bug fixes, especially for an OS and application software that rely so much on not breaking backwards compatibility.
> Telemetry is not really a problem except to the paranoid.
How can you call anyone paranoid after all that we learned in the past decade?
> Not only is it anonymized
Oh so that's why Microsoft uses a hash of your mac address in multiple products. To make it more anonymous. I'm sure your unique advertisement id is also very important and efivars are only used to make sure your activation is legit.
Such a crummy headline. In all honesty, Google has been an extremely competent organization. Don't hate them just because they are at the top. "Rich get richer, poor get poorer" rings true for a reason, they had the first mover advantage in a lot of things. Monopolies are always a danger, like a once competent ruler turning into a tyrant. Having structures in place to deal with them is a must.
Is it possible the way to fix this issue is to actually have all vendors do the same thing.
Should all the vendors including Mozilla switch to Chromium and then fork it, couldn't we look at this as a "rebasing" of the Web?
If the forks are true hard forks, then we shouldn't necessarily end up in a situation where one vendor has more power than the other.
On the contrary, all vendors would now be on an even footing with the same strong, highly-compatible core engine.
All the issues people have with Firefox and Edge would no longer be relevant, and the only thing left to judge one company from another would be their value-add services or their ethics.
The problem isn't really who has the engine itself-- I'd say for most purposes, most browsers are about as good for 99% of workflows.
It's more of the forces and plays each company has around their browser strategy.
IE wasn't the best browser and didn't have the best engine, but it came bundled with Windows and was free. Firefox was a better browser by some metrics back then, in spite of having a company with less resources backing it.
Today, Chrome isn't the best and doesn't come bundled with devices, but Chrome-only sites are starting to appear (YouTube TV notoriously was Chrome-only until just this August [1]. This article lists Google Meet and Google Earth as other Chrome-only sites, at least at launch [2].
Moreover, you hear stories like these [3]:
> "I very recently worked on the Edge team, and one of the reasons we decided to end EdgeHTML was because Google kept making changes to its sites that broke other browsers, and we couldn't keep up. For example, they recently added a hidden empty div over YouTube videos that causes our hardware acceleration fast-path to bail (should now be fixed in Win10 Oct update). Prior to that, our fairly state-of-the-art video acceleration put us well ahead of Chrome on video playback time on battery, but almost the instant they broke things on YouTube, they started advertising Chrome's dominance over Edge on video-watching battery life."
So even if all browser engines were to be hard-forks of the same, each one would be developed different, and each company would play their cards differently to try to gain or maintain dominance. Google being such a big web company makes it difficult to compete against.
> IE wasn't the best browser and didn't have the best engine.
When IE 6 was released in 2001, Firefox didn’t exist - it was still Netscape Navigator. At the time, it was highly regarded, much like the previous releases. Tridend (it’s rendering engine) was leaps and bounds ahead of the competition. Microsoft are in this position now due to massive complacency - the vert same reason that they are not in the mobile market.
Forks diverge over time and then we'll have more or less the same incompatibilities start cropping up. Besides forks prevent fundamentally new ways of doing thinks.
Given that the major engines all implement specs pretty closely, it should be up to web devs to do minimal cross browser testing...
What did they do different since then? Or are you saying you've changed not them? Since one can find fault with any company, is there any company you would trust even if given a rebuttal for why you shouldn't?
The main difference is that I now see them striving for profit instead of excellence. They go together to some extend but the focus has shifted.
It used to be that they made a ton of money off advertising and put that money towards building all kinds of crazy (free!) shit.
It’s possible it was always their goal to just lock in their userbase, but I think it’s just that the company has transformed into something different.
The specific point I’ll always refer back to is the switch from the playful ‘Don’t be evil’ to ‘Do the right thing’, as cementing the difference.
Yet I'd choose Google over Facebook, Apple, Microsoft or Amazon every time if I have to. Even though monopoly is always worse than competition, Google has managed to advance the web experience using Chrome/Chromium (e.g. V8, pushing HTTPS, supporting the latest CSS and JS features, GQUIC, WebP)
It seems people keep complaining that Google is "evil" but they forgot that until only a couple of decades ago, the business world was ruled by oil, banking and finance ultra selfish assholes who had no shame to bring their own countries and even the world if possible to its knees to make profits.
People also forget that when those companies were young they were hailed as tech innovators who brought economic prosperity, not evil monopolists destroying the planet with their unintended consequences. That part comes after
