We aren't going to see browser competition unless and until we fundamentally rethink what the internet should be.
The specifications are simply too complex to foster any meaningful competition. There's too much ground to cover playing catchup and Google has such a large market share that, combined with their data and ad monopoly, allows them to create new specs and features that focus entirely on growing their business model.
Web sepcifications should be simple enough that it's pheasible for small teams to actually maintain a quality open source browser with modest funding. We've ended up falling into the same trap as many other big systems, we lost sight of what the goals and limitations for the web were and instead chased convenience and shiny new features above all else.
I'd like to see an alternative to the Web as well. The truth is that we don't need web browsers as much as we need say Wikipedia, YouTube, Reddit, Amazon, Facebook, GitHub, StackOverflow etc. And we don't need those sites so much as we need the data that is on those services.
What is a better source of how to tutorials YouTube? What's a better venue of asking questions than Reddit and StackOverflow? What use is the Internet without those?
The AI wave is doing this, by which I mean it is finally separating the content and presentation of information and allowing the collation and reformatting of substance across multiple entities which today each have their own interfaces.
For example, the entire idea that we visit specific websites and manually compare prices when buying things is ridiculous and misses huge parts of the benefits of using computers in the first place.
That's because the pendulum hasn't swung back in that direction yet. By this time next year, it's likely that many of the popular LLM sites will slip in advertisements, pushing the agenda of whoever trained the models.
I think this is actually a modern misunderstanding of the original benefits of computers. Computers were originally designed as a tool, they allowed us to work more efficiently but weren't directly taking over doing the actual work. A spreadsheet doid the calculations for me, but I had to provide both the data and the actual formulas to run.
AI, or more accurately machine learning, attempts to take us out of the process. It isn't separating content and presentation, it's separating content and understanding. The algorithm will read in data, compress it, and save only the patterns it learned. By the time the content is presented there is little connection to the original source and what I'm seeing isn't just a separated presentation layer, it's something entirely different like playing a game of telephone.
Google's browser engine is open source. You can reuse Google's browser engine meaning that you can focus on the features that differentiate your browser instead of implementing the spec like you suggest.
Yes. I see chromium as the linux kernel. Piece of fundamental infrastructure that does the the heavy lifting and one can reconfigure, tweak, add features or eventually fork if needed. Never been easier to build an independent browser.
The open web is dead if the only option is to base every browser on a Google core. How do specs fit into this if every browser is just a UI shell around chrome? Where is the competition driving innovation of the web platform when the only competitive landscape is how pretty you can make your toolbars and bookmarks?
We need independent groups of passionate people making sure that the web platform is focused on the best end user experience. We don't always have that today, but that vision is completely dead if the web functionality is decided entirely by an advertising platform.
Keep in mind that when using chromium you can innovate on any part of the stack not just the UI. You can focus on what matters to you: You can reimplement CSS animations, the network stack, add any new APIs, define any privacy or security policies…
A robust and reusable browser engine lowers the barrier of entry for anyone to innovate and differentiate. Browser compatibility is close to solved since everyone shares most of the implementation. Better user experience across the board at lower cost.
Specific examples are Brave, Opera, or more recently the Meta Browser shipping in the Meta Quest: very small team now leading the WebXR standard. it wouldn’t be possible without Chromium.
>How do specs fit into this if every browser is just a UI shell around chrome?
The browser engine is open source. You can make changes and proposals with your own fork. If you are making a browser for musicians you could try and advance the WebMIDI spec and add new functionality to it and then try to get it standardized. But as a browser developer you don't want to me spending your time implenting specs that other people have already implemented, but rather spending time working on things that improve your browser for your userbase.
What you're describing is the opposite of the whole point of specifications. The aim wasn't to circle around a single implementation, it was to define feature specifications that would allow multiple vendors to build cross-compatable web browsers.
What would be the point of open specifications if there was only a single implementation?
>What would be the point of open specifications if there was only a single implementation?
I never said there was? When you write the first implementation then there will only be a single implementation, but that's how counting works. 1 comes after 0. Anyone is free to write another implementation.
Exactly, in that model the only way to build a browser is to depend on someone else to implement the specs. It's not feasible to build a new browser, at best you can start with what Google implemented, including proprietary features, and go from there. At that point you likely don't understand the codebase fully and are going to be dependant on staying in sync with chromium to get new features.
This model of starting with chromium really has nothing to do with building a web browser, it is just building a UI around Google's browser. There isn't much daylight between that and the often complained about limitation of Apple only allowing the use of WebKit on iOS.
I think we're just thinking about very stiffener goals for the web. I'm not interested in a version of the web that is driven by one browser or company, even if it can technically be modified or tweaked as long as you don't diverge see enough to break upstream compatibility.
Your original comment was about competition between browsers. For browsers to compete they don't all need to have their own browser engine written from scratch.
Web standards are not driven by one browser or one company. That would still be true if everyone based their browsers off of Blink.
There isn't any real browser competition when the only difference is in what the UI around the actual web browser is. The browser experience is what actually renders the web page, not the menu system around it.
We wouldn't need standards at all if everyone used the same browser engine. Whoever owns the engine, likely Google, would make the engine however they choose and could change core fundamentals of how the web itself works.
Many devs were up in arms when a GitHub issue raised the idea of effectively DRMing the internet. Google can bake that right into chromium if they want to, and if they were the only browser engine they could easily get away with closed sourcing chromium and only shipping binaries so the feature can't be disabled or removed.
Web standards in that world would be an internal discussion at Google, likely driven entirely by their business needs. There's no browser competition there, the open web would be completely dead.
>There isn't any real browser competition when the only difference is in what the UI around the actual web browser is.
I am not suggesting that people only make UI changes. I am suggesting that there is value in reusing work that already exists.
>The browser experience is what actually renders the web page, not the menu system around it.
Chromiums renderer is going to be fine for most people looking to build a browser. People building a browser should focus onan building what makes their browser unique. Not every browser is setting off to create a new rendel engine.
>We wouldn't need standards at all if everyone used the same browser engine.
There still would be a need for cross industry collaboration on how to move the web forward. If Clang was the only C++ compiler there still is a benefit on having the standards process and working groups figuring out how to evolve the language.
>Whoever owns the engine, likely Google, would make the engine however they choose and could change core fundamentals of how the web itself works.
This is a problem with browser marketshare and not browser engine marketshare.
>and if they were the only browser engine they could easily get away with closed sourcing chromium and only shipping binaries so the feature can't be disabled or removed.
Other browsers can work together on developing on top of the last public release of chromium. This is still less work than all of these browser developers independently creating their own browser engines.
The specifications are simply too complex to foster any meaningful competition. There's too much ground to cover playing catchup and Google has such a large market share that, combined with their data and ad monopoly, allows them to create new specs and features that focus entirely on growing their business model.
Web sepcifications should be simple enough that it's pheasible for small teams to actually maintain a quality open source browser with modest funding. We've ended up falling into the same trap as many other big systems, we lost sight of what the goals and limitations for the web were and instead chased convenience and shiny new features above all else.