It's pretty sad that the only way to ship a usable browser these days is by fork...
https://lunduke.substack.com/p/ladybird-a-truly-new-web-brow...
A specialized one? Not so. You don't need to fork chromium to render some HTML on the screen of a specialised hardware. Nor do you need to fork chromium for rendering the "about.html/css" in your mobile app. And so on. For some reason people do this anyway: good chance your smart fridge is running full chromium to show its temp and some images. But it could use a simple rendering engine optimized for embedded hardware just fine.
Modern browsers support a lot of most definitely not modern features. I believe it would be feasible to create an entirely new browser from scratch if those older features were ignored. Take for instance Mozilla's experimental Servo rendering engine and Deno for ECMAScript - these were created in mere years with (relatively) small teams. I would be quite comfortable using a web browser that supported, say, CSS Flexbox and Grid quite well but didn't implement every historical way of centering text on a page.
With HTML <5 (and excluding XHTML from the discussion) the browser has no clue what all those <div> and <table> elements are supposed to be, so CSS and JS are necessary to reproduce basic things like screenreader flow for all but the most basic page layouts.
Ironically, the opportunity for users' browser to apply stylesheets on generic, semantically meaningful sites was one of the original concepts behind CSS (hence 'cascading' in the acronym - the browser takes precedence), which is now usually employed to overwrite the browsers' defaults!
Even Blink, the rendering engine of Chromium, is a fork of the WebCore component of WebKit, which was originally a fork of the KHTML and KJS libraries from KDE.
EDIT: TIL about Ladybird, not sure how accurate its rendering is.
https://awesomekling.github.io/assets/ladybird-proggit.png
It's impressive, but it also shows how hard and everchanging this sort of task is. I don't see Youtube or Google Docs as always being reliable.
LibWeb/Ladybird was made by a handful of people working part time on it over a couple of years. During that time, they managed to implement absolutely everything from image decoding to font rendering to javascript parsing and execution, all FROM SCRATCH.
Stop trying to twist it as if they applied tons of resources and came up short. Before last month they had nobody working on it full time.
They don’t have anywhere near the corporate support real browsers do, it’s just chasing cars.
They don’t have the HR to scale, some day they’ll get Google Docs to work “hurray we did it, looks like it is possible” then they’ll try to get Spotify working. Next thing you know Google Docs will break again. It’s gonna be that times 100+ other apps that people use.
You can call me pessimistic, hell even antagonist, I don’t care. I’m being honest in order to outline how unrealistic it’s become to write a browser from scratch for the modern web.
Andreas or whatever his name will get the apps he wants working and maybe some more, but it will NEVER be something people can just jump on and expect to use as a functional browser. It’ll be some Reddit Discord HTML4.5 viewer.
I’d say if they wanted to make something usable, they’d fork WebKit and bolt on SpiderMonkey, QuickJS, JerryScript or whatever and build up from there. But they care too much about their stupid pride “look what we did”. And that “roll your own” nonsense isn’t attractive to corporate adoption unless they’re some project like OpenBSD that does it with a real purpose, not HN front page gratitude. SerenityOS is advertised, not as a microkernel, not architecture portability , but as a “90s operating system”. That is some hipster bs that only depicts the Amish rate of development Ladybird will take unless they really think about the goal they want to achieve. Because right now, they want a usable browser and they want it from scratch and I can only see one of those happening.
Linux might’ve initially been “Just for fun” but now they clearly mean business
The problem is Google, being such a large company is far to powerful to compete with. Not to mention they own a lot of the products on the web that other engines need to be compatible with to be usable.
What should be done is Google should be broken up so it's no longer one company leading this thing but I can't see that.
That's just moving around terms. I think it's widely known that almost everyone knows all their devices are in some magical way connected to each other.