There's one thing I'd like to mention to everyone who says it is impossible to build a new browser: while that may be true today, I would just like to point out that the ancestor of the dominant browser engines of today, Blink and WebKit, is KHTML from the Konqueror web browser.
While KDE has had some contributions from large corporations, it would be a gross exaggeration to call KDE (or any Linux desktop) a popular success. It more closely resembles a hobbyist project than a commercial juggernaut.
While it may be true that KHTML/Konqueror was not as polished as WebKit/Safari or Blink/Chrome before those forks poured massive amounts of corporate development effort into those browsers, the fact is that it was a solid enough foundation for those later corporate contributors to leverage into world domination.
In other words: Today's small scrappy upstart browser that only hobbyists use may be the embryo of tomorrow's near-monopoly. Please don't abandon your ideas for new browsers+engines, give them a shot. Even if you do not directly benefit from its future success, at least you may give the web a gift that everyone will someday enjoy. The new boss Chrome is an unqualified improvement for the web over Internet Explorer, if only in standards support and open source code. I only hope that someday we can replace it with something better than a new corporate overlord.
I see these sorts of appeals a lot on HN but never really understood the point. Sure, software had a good run of small devs making big splashes, but it's pretty obvious to anyone watching that the garage days are over as far as big projects are concerned. The only real exception is the European software devs who, from a market perspective, exist mainly to sell the most promising projects to China or the US since they can deploy those projects in a way that drives growth. Why shame people for making smart moves?
Is asking people to keep hope alive the same thing as shaming them for feeling hopeless? I hope my words don't come out that way, since it's certainly not what I intend.
Also, one point I'm making here is that a small dev may not be able to make a world-dominating project by themselves, but it may serve as the foundation of something larger later. Is that really surprising or controversial?
Perhaps a more important point I'd like to make is that neither Apple nor Google were quixotic enough to start a browser engine from scratch, for all of their expertise and cash. Without KHTML, we might still be suffering under IE or something even worse. It may take someone willing to tilt at windmills to dethrone Blink/Chrome, I don't think the suits will save us.
There are amazing almost turn-key FOSS solutions just an "npm install" away. (Plus 3D printing, plus dropshipping from Shenzhen, plus all the other stuff I'm not even aware of, but startups can and will use.)
Sure, garage scale nowadays means just a few millions of USD angel/seed, compared to the 100K USD a decade (?) ago, but it doesn't mean that there are no niches waiting to be occupied, no new markets emerging ever in the future, etc.
The thing is, back when KHTML was made, it supported pretty much everything you needed to browse the web. Even the version in KDE1 (which is only a handful of KLOCs) supported almost everything you'd see online.
KDE actually looks halfway decent though. Unity (if that's what Ubuntu's still calling it's UI?), though, most definitely has "hobby project" written all over it. I tried Ubuntu for the first time recently and all kept hitting little catches and speedbumps - but they weren't "oh, that isn't implemented" or "oh, this doesn't work that way", they were all bugs. The alt-tab subwindow thumbnailer lets you select multiple application windows at once. The screen locking system not only gets extremely confused by fullscreen windows, once said fullscreen apps exit the system belatedly picks that exact moment to go to sleep. Notification toasts cover the type-to-search box, which is fun when said toasts decide to get stuck on the screen. The hotkey system eats ctrl+alt+(left|right) even though the system is configured to use vertically-aligned desktops, so I can't switch between desktops in my openbox VNC session unless it's fullscreen. The system doesn't feel holistically tested. I'm sure there is actual testing done, but it feels like <10% (<5%?) of what Windows/Office/etc gets.
Sure, perhaps things will be in a different place in 5 years. (Especially if The Dispatchers Of Wisdom™ don't decide to give up and start again with a new UI and applications. Cynically not giving the current effort a high chance of surviving...)
OK, my point/question: I think one of the prerequisites for hobby projects to make it big is playing in a space that's largely uninteresting or undiscovered to the majority of people that will notice (and understand, and feel threatened by, and respond to) the new thing, before it has reached critical mass. Perceived risk being somewhat irrational and all that.
KDE and Unity feel like hobby projects because "desktop Linux" is forever the thing that never really existed, and the only people who pursue the idea are people that don't get that. But their work gives enterprises some free source code with APIs that let them make application windows with buttons in them! Nice! But there's no point for those enterprises to contribute functionality back to the UI engines, because that doesn't further the enterprises' application-specific focuses.
Web browsers definitely aren't an undiscovered thing by this point. I'm sure more people than might reasonably be expected probably check out where NetSurf is up to every 6-12 months or so, for example. (I don't mean hobby project users here, I mean would-be commercial users with adequate domain savvy to be aware of the entire ecosystem. Total theorization on my part but I do wonder if I'm right.)
"A new kind of web browser" is a fun idea, but impractical in practice as it's preeeety much the polar opposite of an unknown idea now. A new web browser would be subject to an atlas-holding-up-the-entire-world level of passive and active social scrutiny if it looked like it was taking itself seriously. (Servo is kind of a good reference example there. It wouldn't survive if it didn't have Firefox and the indefatigable Rust team behind it; and even as it is things are still struggling).
By the way, serious question - what does source-level access to Chrome mean, at the end of the day, for the vast majority of internet users? The only real quantitative (not qualitative) answer I have are the small high-reward collection of bug bounties that have been rewarded over the years to external contributors. The fact that you can build Chromium and ship it [as part of Linux distributions] does feel a bit like an implementation detail that fulfills Chrome's goal as a technology moat (and competitor to Firefox).
While KDE has had some contributions from large corporations, it would be a gross exaggeration to call KDE (or any Linux desktop) a popular success. It more closely resembles a hobbyist project than a commercial juggernaut.
While it may be true that KHTML/Konqueror was not as polished as WebKit/Safari or Blink/Chrome before those forks poured massive amounts of corporate development effort into those browsers, the fact is that it was a solid enough foundation for those later corporate contributors to leverage into world domination.
In other words: Today's small scrappy upstart browser that only hobbyists use may be the embryo of tomorrow's near-monopoly. Please don't abandon your ideas for new browsers+engines, give them a shot. Even if you do not directly benefit from its future success, at least you may give the web a gift that everyone will someday enjoy. The new boss Chrome is an unqualified improvement for the web over Internet Explorer, if only in standards support and open source code. I only hope that someday we can replace it with something better than a new corporate overlord.