The Gemini Protocol does a good job at being a minimal alternative, but it’s quite limited in what is capable. It’ll never be adequate for banking for video games.
Social media and reddit alternatives? No problem.
Comment on I tried Servo, the undercover web browser engine made with Rust
pr06lefs@lemmy.ml 1 day ago
yeah I’m glad servo exists, but its far from ready for prime time.
What’s crazy is what a monumental undertaking implementing a web browser is now. Would love to see someone propose a new standard to replace html-css-js but distilled down to a radically simplified essence, dropping the accumulated cruft of decades, applying lessons learned. Ideally would be something that could be implemented by a few devs over months rather than requiring a team of hundreds over a decade.
The Gemini Protocol does a good job at being a minimal alternative, but it’s quite limited in what is capable. It’ll never be adequate for banking for video games.
Social media and reddit alternatives? No problem.
It is crazy. And maybe it could be distilled down, but maybe because of what it’s become and how it’s used, that’s just not an option anymore. The context is that the “web” that browsers are browsing has grown from mere rich text and links into basically a fully distributed operating system. There are entire software suites that exist only through web protocols now. Literally anything you used to be able to do on a Desktop OS you can now do directly on the web, often at very close to bare metal performance levels. And over the years and decades the standards have evolved to not just enable that anymore but to actually require that level of functionality. It has become completely expected to have javascript APIs allowing extensive and instantaneous DOM manipulation, HTML5 canvas and storage and WebGL available, they’re not just “optional addons” you can pull in with an extension or that a text based browser might not bother to implement, they’re a core part of the web and little will be functional without them.
So when you’re building a “modern” web browser what you’re effectively really doing is implementing an entire cross-platform OS, sandboxed and virtualized for security within any host OS you choose to support.
Of course technically “the web” is still backwards compatible with the old pure HTML, no javascript, no CSS, web 1.0. There’s nothing stopping anyone from writing such a simple site today, and those websites are still out there. And that’s still sort of where you have to start with projects like Servo, because that’s just the basic level of absolute minimum functionality. But it’s taken a long time to build all the features of the modern web and so of course it’s going to take a long time for a new browser engine to implement all of them or even enough of them to actually start supporting the most commonly used websites.
While there are definitely a lot of quirks related to handling old sites and the various inconsistencies and incompatibilities that developed over the years, I don’t think that’s the real sticking point on developing a new web engine at this point. I think the issue is simply the fact that the web does so much and is such a comprehensive technology platform, and if you tried to simplify it, to make it easier to develop browsers, you would lose a lot of actually important functionality for developing websites that allow them to do the things they are doing today. Granted some of those things I wouldn’t mind losing either, but a lot of them are legitimately required for what we do with the web now and what we expect it to be able to do.
There are gopher browsers that allow you to visit an alternative, more simple “web”
m_f@discuss.online 1 day ago
That’s pretty much en.wikipedia.org/wiki/Gemini_(protocol). You can explore it through an HTTP proxy like this: portal.mozz.us