This how I used to read HN at work, nobody bats an eye at text in the terminal. I'm sure I could have gotten away with sipping my coffee in the morning with the orange website plastered all over my screen, but I felt like it was important to keep up appearances--call it productivity theater.
That's why I play dungeon crawl stone soup in the terminal. I'm defending myself against fire dragons and dodging iron shots, but it looks like I'm aggressively debugging something.
Only problem is having to type https:// in front of news.ycombinator.com because lynx assumes you’re trying to make an nntp connection to anything with a “news” subdomain.
I just tried that on mine and using `lynx news.ycombinator.com` went to the website fine. I don't see anything obvious in a quick scan of the manpage nor config that might affect that. Weird!
Lynx (in tmux) is still my default browser. I only go for Firefox if I really have to. Its darn FAST, has no remote code execution, and is typically all you need to read text on the web. Its also a good test if the web dev is full of shit :-)
I have not used Lynx interactively in ages, but when I want to save some pages of text from somewhere (typically an old web forum thread or blog) I often call lynx -dump (followed by gzip) from a script. It is an easy way to get nicely formatted text with all the HTML nonsense removed, and it also includes a list of all links from the page. Not exactly what a real archivist would do, but it works for hoarding text.
I did not know that it had that feature, and having just tried it out it does work really well.
Its output kind of reminds me of the gemini protocol and its markup format[0],
> It doesn't need a blocker for third-party scripting, service workers ... because the browser's technology simply doesn't recognise these page components
I'd never thought about it like that, but it's a nice side-effect. You don't have to worry about obtrusive JS, if your browser doesn't know what JS is. :D
It's worse than that[0] - Lynx statically renders the page at load time which means any "dynamic" updates just can't happen. I did have a bash at integrating the Mozilla JS engine but gave up after realising that anything more than just `document.write("flange")` handled at load time would need the entire inner engine rewriting and that was/is far beyond me[1].
[0] This is based off my memories of the Lynx source back in the Olden Times, ~1995-2000ish.
[1] But I did do the original cookie jar and colour styles (as seen in the screenshots) work (`lynx.lss` is all my fault, sorry)
Chawan also renders CSS + images, but unlike the above two, it comes
with its own browser engine (from scratch). It also does JS with
QuickJS - if you're lucky :P (Lately I got Transmission's web UI to
load, quite happy about that.)
Do you remember the time you could buy on amazon.com with lynx? (or links2).
And all the other dominant online services which broke noscript/basic (x)html to the benefit of the absurdely and grotesquely massive and complex (including their compiler) whatng web engines? This did shut the door definitely in the face of all citizen/state sponsored/small commercial alternatives which [could] have been super stable in time (including their SDK).
This is not "evolution" or "innovation", this is a scam and planned obsolescence at accute levels.
You could perfectly stream a video with a dash/HLS URL transfered to a media player (ads can be text on the web page and they can include some directly into the video stream).
IMO, a most useful thing is Lynx in DIRED mode. A great read-only directory browser - just enter a local url when you start it. For example, <lynx .> or <lynx /> .
Readit News
Bad YC!
I mean, I’m sure it has its vulnerabilities, but nobody is actively looking for them.
In .mailcap include the line
in .muttrc include the line[0]: https://geminiprotocol.net/
I'd never thought about it like that, but it's a nice side-effect. You don't have to worry about obtrusive JS, if your browser doesn't know what JS is. :D
It's worse than that[0] - Lynx statically renders the page at load time which means any "dynamic" updates just can't happen. I did have a bash at integrating the Mozilla JS engine but gave up after realising that anything more than just `document.write("flange")` handled at load time would need the entire inner engine rewriting and that was/is far beyond me[1].
[0] This is based off my memories of the Lynx source back in the Olden Times, ~1995-2000ish.
[1] But I did do the original cookie jar and colour styles (as seen in the screenshots) work (`lynx.lss` is all my fault, sorry)
1: https://github.com/fathyb/carbonyl
2: https://github.com/browsh-org/browsh
Chawan also renders CSS + images, but unlike the above two, it comes with its own browser engine (from scratch). It also does JS with QuickJS - if you're lucky :P (Lately I got Transmission's web UI to load, quite happy about that.)
And all the other dominant online services which broke noscript/basic (x)html to the benefit of the absurdely and grotesquely massive and complex (including their compiler) whatng web engines? This did shut the door definitely in the face of all citizen/state sponsored/small commercial alternatives which [could] have been super stable in time (including their SDK).
This is not "evolution" or "innovation", this is a scam and planned obsolescence at accute levels.
You could perfectly stream a video with a dash/HLS URL transfered to a media player (ads can be text on the web page and they can include some directly into the video stream).
In fact w3m was originally intended as a filesystem browser and file pager, with HTML functionality added to it. See: <https://w3m.sourceforge.net/>
I have elinks as a default `tmux` pseudoterminal. I don't use it often, but it's always there.
My emacs setup uses w3m as the parser for HTML mail in VM, my email client. It works really well. <https://news.ycombinator.com/item?id=39793342>