How and why to use Lynx – the faster web browser
shkspr.mobi
shkspr.mobi
But the killer app for me with web browsers is link hints, like in several browser extensions (e.g. vrome, tridactyl) and in browsers like qutebrowser. It takes a little getting used to at first to use efficiently, but I hate not having it when I use Lynx or w3m.
Slightly apropos, and for what it's worth, I've noticed that 99% of my browsing seems to be or start from maybe 20 or so web sites. I've started creating shell aliases to open those sites and there is no reason not to use the best browser for the site for each one. It works well with a philosophy like seems to be the one behind the surf browser, where you open each thing in its own window and use your window manager to deal with it, instead of using tabs.
and those shell aliases sound very similar to surfraw: https://en.wikipedia.org/wiki/Surfraw
Quite interesting to read Assange wrote Surfraw in 2000.
[0] https://en.wikipedia.org/wiki/ELinks#cite_note-2 [1] https://en.wikipedia.org/wiki/Links_(web_browser)#cite_note-...
But I used to use it more often. Some years ago I worked mostly on a terminal window and by using lynx, I could browse the internet while everybody thought I was doing some serious work on the server (folks over there respected people working on black text-based screens)
Slow and perhaps also unusable --- such as the popular anti-pattern of requiring JS to render static text, a great example of that being Twitter.
That's my main use of Lynx these days - for ad-hoc extracting a bunch of links from a page, there's not much easier.
text/html; lynx -width 72 -assume_charset=%{charset} -display_charset=utf-8 -dump %s | sed 's|^ ||'; nametemplate=%s.html; copiousoutput
Trying out the alternatives after reading the other comments here, perhaps I should give `w3m` a shot some day as it renders white space rather pleasantly?My blog's homepage is dominated by that construction, so my preferred alternative is w3m.
Lynx's keybindings never make sense to me.
I've set up various aliases in my dotfile repo that make it easy for me to do search through Lynx to google, ddg, wikipedia etc as well as instantly bring up my most used and read news (and other) sites.
It is really rather handy when you get used to it. But it simply could never replace a modern browser due to the JS dependencies needed.
I've used both and links is vastly preferable, IMO.
Although I tend to use the name "elinks" since it's much more google-able.
elinks started out as an experimental fork of links but these days they’ve diverged a little. For example, links supports a whole plethora of operating systems, including DOS! While elinks is POSIX only.
One of the projects I intent to one day do is a modern HTML-only browser with tabs, video support and other useful amenities.
I've been messing with Gemini and some of its browsers (I'm a big fan of Lagrange) and I have to say it is spectacular. The markup language for documents is great, the extremely limited functionality is great (although I'd like to see italics and bold text) and I really like the mostly thoughtful content. That latter one would probably change if it picks up a lot of steam, but so far I wish all websites who's audience is just reading were Gemini documents.
I have at various times found myself unable to use a GUI, and these browsers saved my butt finding information to restore my system.
This is why I test all functionality in all three of them and ensure that they are just as able to use the site (or better) as Chrome Canary.
Some other terminal user-agents to try:
rdrview, go-readability: these are CLI utilities that use Mozilla's Readability algorithm to send HTML of just the article to stdout. Pipe that into w3m/lynx/links to read articles in the terminal. Testing in these ensures that your website will work in Firefox's Reader Mode. I also like to diff my website's HTML against the HTML output of these tools, just to see how much extraneous HTML I have.
Mozilla's Readability is written in JS; these tools are written in C and Go, respectively. With the exception of a few whitespace characters, their output of both seems to be identical.
=> https://github.com/eafer/rdrview rdrview
=> https://github.com/go-shiori/go-readability go-readability
Edbrowse is a line-mode browser that supports JavaScript via Duktape. Originally written for blind users, it's great for sighted users as well.
=> https://edbrowse.org/ edbrowse
If you want to go the extra mile and ensure that your website works with the most basic HTML parsers, try sending it through an HTML-to-Markdown converter that doesn't include inline html in the Markdown output. I test with pandoc's strict markdown output and make sure the resulting markdown is readable:
pandoc -f html -t markdown_strict https://example.com
None of this is a replacement for real a11y testing; most a11y technology leverages mainstream browsers (Chromium, FF, Safari, et cetera).Yes, you can do some sort of accessibility testing with it (and I'm sure it is an useful tool for a lot more corner cases) but I was expecting a lot more after reading the title
I often open a news page or something and it's completely blank at first because so many scripts and assets got blocked by uMatrix. Before I actually unblock any domains, I try toggling reader mode on - a lot of times it's enough to let me read the plain HTML text which I actually wanted.
Browsing stackoverflow and forums with lynx is possible and saves you in situations when your graphics card drivers fail for some reason.
I occasionally use eww, though it a bit of unerogonomic pain. Esp without url autocomplete based on browser history.
Though I'm sure I could yak shave that with some elisp if I was motivated enough
(Firefox seems to eat all 8GBs or my RAM lately so it's getting more tempting)
How many tabs are you opening? I have right here, right now 20 tabs opened in Firefox and Task Manager says my RAM is at 3.7 GB usage. That includes everything, from OS to other background apps that I'm running.
Like you say, it doesn't release memory so I need to restart periodically. If I forget then I need to do a hard restart when the system locks up (sometimes loosing work :/)
It seems strange.. it used to handle hundred of open tabs just fine and now it's completely crippled. My guess is that in making the browser faster they cache more and more - and Firefox devs prolly run on fancy machines with tons of RAM.
I can record a video if you don't believe it.
I've started using Lagrange probably more than I ought to lately.
PAGER='lynx stdin' perl -d my/dev/server.pl
... hit a breapoint
| p $mech->content # nicely rendered html in the pager
No thanks.
'void main() {}'
I can play full games like krunker.io right now in browser with minimal lag.
I hope all the OS developers are taking note. OS might be just a browser plugin in the future instead of the other way round.
Humans have a brain that soaks new information like a sponge. Internet is the new crack. Having a limited browser is sad.
Edit: ok I tried it. I have to allow 2-3 cookies to run google.com. Facebook doesnot run. I think its a niche thing. Like among peers it would be fun to learn to run and run it but not a stable thing.
lol?
> I hope all the OS developers are taking note. OS might be just a browser plugin in the future instead of the other way round.
Maybe the promise of "thin clients" might be realized. Anything is possible.
> Humans have a brain that soaks new information like a sponge.
Where did "Can't teach old dogs new tricks" come from then? Maybe fpr young children, but older people have trouble picking up new technology.
> I think its a niche thing.
I guess you are right - like a sponge.