This was done long ago. I forget the name, but there was a browser that only rendered the text parts. It was useful on slow modems or text-only screens.
I have this project of moving most of my browsing to EWW (the native web browser in Emacs) and using a full-fledged web browser only when I need to use some web app. But many content-only websites are going with completely dynamically rendered pages with React or what not, so all you have is void at worst if you disable JS or view with a text-only browser, and garbled up unreadable stuff if you're lucky. Also e.g. the comment hierarchy of HN does not render nicely with text browsers.
I kinda wish there was an alternative hypertext to html focused on terminal rendering.
I would love to be able to write cui the same way i wrote web pages. right now the best solution seem to use a text-based web browser, but the are kinda limited, (i didn't find a good way to render menus for example)
This seem closer to what i want
Yes, but this is the UI of a browser. Arguably it could be more efficient to do it this way as you can strip out some of the libraries used for the native rendering.
They would have problems rendering text though. I'm working on a text editor and made some research on the fastest way to render text. It's really hard to beat the OS API.
Big fan of custom rendering approaches, but wouldn't such a Design system sidestep any and accessibility tools? Nothing for screenreader to hook into. Text and interfaces would be neither native nor DOM based.
In theory you can build[1] the text renderer on top of the Servo web engine. Something like elinks 2.0. It will allow to browse modern Web while being lean.
This is a very cool project. It uses headless firefox and then renders to the terminal. Text remains text, and then videos and images get projected to super low res bitmap-via-characters in real time.
The closest thing currently is probably https://github.com/trimental/inlyne which differs in two ways: it only support markdown not arbitrary HTML, and it renders to screen rather to an image. But it's a good start.
IMO the main blocker for web rendering in Rust right now is better text layout, and in particular support for embedding non-text content within text ala `display: inline-block`. If/when that is implemented I think we'll be able to do a decent job of rendering basic web pages.
I want to use this, but I'd really like text only. When I say that, I don't mean "render images in text". I want text only. Text I say!
Created a GitHub issue regarding this on the project, but it was promptly closed with "Press CTRL-M", which is not what I'm after nor works by default.
But this wouldn't even have, like, font rendering. Input methods. The amount of stuff every single site would have to implement just to show something on the screen is staggering, unless the API in question is no longer small and simple and provably safe.
Not GP, but there are browsers that can render and dump HTML to plain text, like Lynx. You can of course also run remote X programs, but that seems a bit clunky.
This looks interesting. It's probably also worth investigating ways for website owners to make pages that are easy for this engine to render as text while preserving as much of the page's structure as possible.
No, I meant a text mode browser that does not use HTML. Lynx, and all the existing text mode browsers are intended to render HTML as text.
This would be a browser that targets only text mode user interfaces. So pages would need to be written specifically to be rendered in such an environment.
Why? Because text mode browsers are really trying to do the impossible - to render modern web pages as text is essentially not doable in all cases. Instead make a browser for the text mode use case specifically designed for the medium.
A mashup of a web browser and ANSI/VT100/teletext. It would need a sandbox and a JavaScript interpreter.
reply