Largely it's an issue that no one sees redevelopment as a likely path to get rich. Cost wouldn't be outrageous. But in this age no one's building browsers and no one's building OSes because it's already a well served market. It's unclear why we would reimplement. Especially when there's perfectly good ways to completely redo the user experience while using the existing open source implementations!
Expecting to build a world class software platform for less than, oh, I don't know, maybe $25m invested in a good capable engineering team seems silly. What does the web take to redo? More? $100m? Maybe, but I expect significantly less.
Folks use this as such a scare tactic: oh, it's be so hard to do! There's some incidental complexity sure but it's not the bulk of the design, and most of the core & a huge amount of the capabilities are purposeful and solid in nature. Folks also try to work people up about performance: oh, it's so big and bulky. But there's plenty of data heavy very fast sites doing enormous things out there, for quite reasonable footprint, and we're only just getting into the webworker & wasm age where we can start using the existing platform much better. The web's stance continues to improve drastically, and that's not just by being harder to implement, it's by ever refining & improving. It's not gloom I see as we iterate: it's strengths adding together & reinforcing.
Last it just seems so hideously developer-centrix to imagine doing anything else. Nothing else has any legs to stand on for giving users agency. SVG, HTML, and CSS are completely unique in software development in that they leave the door open, are alive computing systems that users can modify & extend with the browsers and plugins that they want. There's no proposals remotely on the table anywhere to any degree that do 1/1000th as much for the user, to respect them & maintain their agency.
So, I don't think the web's scope is a problem. I think it's very doable to build your own world's best platform, for a pretty reasonable sum, if you can put together a small or medium business to do so. That's an ok bar. We can cobble together demoscene grade cool flashy things for less, sure, and that scratches a nerd itch. And maybe some of those will be good high and low level platforms both, be really good maybe even enduring particular ways of doing software. But will they be able to be adapted by users? But will they get it right & never change & never bloat themselves? But will perhaps they too need to evolve & change over time, allow for different styles? The Extensible Web Manifesto ideas have allowed the web to be an enduring low level platform across great architectural explorations & expressions, while still maintaining greatly similar base truths; that empowers developers to find & speak their own truths, while keeping an incredibly capable deepening hyper-multi-media lingua franca that right now the whole world speaks. The web is amazing and building a new browser is a bargain, especially given how excellent the ready-made test suites are.
But no, most engineers won't be able to diy one in their garage; that doesn't mean it's too big & it's ridiculous to insist on. Programming languages/llvm, kernels, none of these are simple or things folks could do de-novo, cheaply, with the needed breadth scope. The popular versions of all of these have had enormous effort poured in. But for some reason we've invented and applied a bar, saying sorry, the web and the web alone has to be doable as a hobby project. It's ridiculous to insist the platforms for the world should necessarily be primitive works. The web keeps winning precisely because it has generous platform for all.
Forking chromium or Firefox or using Webkit plus adding the missing components and fixing broken ones is also not cheap. Again, maintenance is expensive. That is after you find (or train) the people who can reason about those codebases.
Where the Web is heading, very soon it's end users won't be well served with Google being the only one willing and capable of footing the bill for it.
Expecting to build a world class software platform for less than, oh, I don't know, maybe $25m invested in a good capable engineering team seems silly. What does the web take to redo? More? $100m? Maybe, but I expect significantly less.
Folks use this as such a scare tactic: oh, it's be so hard to do! There's some incidental complexity sure but it's not the bulk of the design, and most of the core & a huge amount of the capabilities are purposeful and solid in nature. Folks also try to work people up about performance: oh, it's so big and bulky. But there's plenty of data heavy very fast sites doing enormous things out there, for quite reasonable footprint, and we're only just getting into the webworker & wasm age where we can start using the existing platform much better. The web's stance continues to improve drastically, and that's not just by being harder to implement, it's by ever refining & improving. It's not gloom I see as we iterate: it's strengths adding together & reinforcing.
Last it just seems so hideously developer-centrix to imagine doing anything else. Nothing else has any legs to stand on for giving users agency. SVG, HTML, and CSS are completely unique in software development in that they leave the door open, are alive computing systems that users can modify & extend with the browsers and plugins that they want. There's no proposals remotely on the table anywhere to any degree that do 1/1000th as much for the user, to respect them & maintain their agency.
So, I don't think the web's scope is a problem. I think it's very doable to build your own world's best platform, for a pretty reasonable sum, if you can put together a small or medium business to do so. That's an ok bar. We can cobble together demoscene grade cool flashy things for less, sure, and that scratches a nerd itch. And maybe some of those will be good high and low level platforms both, be really good maybe even enduring particular ways of doing software. But will they be able to be adapted by users? But will they get it right & never change & never bloat themselves? But will perhaps they too need to evolve & change over time, allow for different styles? The Extensible Web Manifesto ideas have allowed the web to be an enduring low level platform across great architectural explorations & expressions, while still maintaining greatly similar base truths; that empowers developers to find & speak their own truths, while keeping an incredibly capable deepening hyper-multi-media lingua franca that right now the whole world speaks. The web is amazing and building a new browser is a bargain, especially given how excellent the ready-made test suites are.
But no, most engineers won't be able to diy one in their garage; that doesn't mean it's too big & it's ridiculous to insist on. Programming languages/llvm, kernels, none of these are simple or things folks could do de-novo, cheaply, with the needed breadth scope. The popular versions of all of these have had enormous effort poured in. But for some reason we've invented and applied a bar, saying sorry, the web and the web alone has to be doable as a hobby project. It's ridiculous to insist the platforms for the world should necessarily be primitive works. The web keeps winning precisely because it has generous platform for all.