As far as I remember, when I wrote my email reader/writer in C, ncurses was not even a thing, so I made my own windows and dialogs in text mode with VT-something codes or something like that. That was not on Linux, but on DOS at the time, but anyway. I did not wait for graphical stuff to be common.
I used my software for months to read and write mails, interact in mailing lists, newsgroups, etc.
Funny thing is when a sysop sent me a mail once: basically, he said to me that one of the machines he had the responsibility of back then warned him about a non-conform header. He asked me the name of the email software I was using to report the bug to the author. I thanked him for my first report ever, for his diligency, etc. and of course, I fixed it promptly ;)
The article has it backwards: GUIs are the third way of using Linux. TUIs are the 2nd way. I mean, stuff like pico, nano, vi, Emacs - these are all TUIs.
Perhaps it's because I grew up on DOS that I think that way, but one of the reasons I switched from Windows to Linux decades ago was to use TUIs.
All the TUI stuff you can leave running in tmux and then have a consistent workspace remotely accessible from your PC or phone is indeed really nice. IRC, mail, torrents, notetaking, etc.
I stick to CLI's and TUI's and use zellij as my window manager. This way the configs and muscle memory I develop on my work macbook are transferrable to the linux box I use for personal projects, and visa versa.
It seemed a bit limiting at first, but after adapting to it I feel like it was the right move. GUI's represent a sort of explosion in the set of valid states for software to be in, and I find that software which opts out of trying to support that explosion tends to work better, likely because the maintainer just has less to worry about.
I'm exploring aspects of these "other ways". I've been pushing Kitty terminal graphics compositing and then layering libraries over that. I've also been working on remoting Golang programs over the network as well as compiling them to WASM and embedding that in browsers. This WebGPU browser shader is an example of that mashed together [1]. Same code for native and browser.
Shout out to TUIOS [2] that is in the article; author has a lot of neat libraries and his `sip` protocol is the basis of that remoting work.
I was hoping the third way would be somehow related to a type of hands-free neuralink so I can code at the speed of thought with my eyes closed laying in bed. Someday the tech will exist but, by then I'm not sure coding will be a thing to do anymore. I'll probably still do it anyway though.
another comment on a different HN thread pointed out a similar issue. Now I'm wondering if backdrop-filter can be used correctly at all or should just be avoided entirely
which, wierdly enough I am useing, and the site scrolls just fine,my browswr has NO java script, No Dom storage, no images/video displayed
though there are a few websites that have some kind of alternate way to display images in short video loops that auto play
What do you mean by "pretty bad"? I'm on Zen (Firefox based) and the scrolling seems to work fine. But maybe it's getting resource constrained on some hardware.
I don't see any suspicious JS on the page, so I suspect it's a styling issue triggering some inefficiency in Firefox. There are animations and text-shadows.
As far as I remember, when I wrote my email reader/writer in C, ncurses was not even a thing, so I made my own windows and dialogs in text mode with VT-something codes or something like that. That was not on Linux, but on DOS at the time, but anyway. I did not wait for graphical stuff to be common.
I used my software for months to read and write mails, interact in mailing lists, newsgroups, etc.
Funny thing is when a sysop sent me a mail once: basically, he said to me that one of the machines he had the responsibility of back then warned him about a non-conform header. He asked me the name of the email software I was using to report the bug to the author. I thanked him for my first report ever, for his diligency, etc. and of course, I fixed it promptly ;)
The article has it backwards: GUIs are the third way of using Linux. TUIs are the 2nd way. I mean, stuff like pico, nano, vi, Emacs - these are all TUIs.
Perhaps it's because I grew up on DOS that I think that way, but one of the reasons I switched from Windows to Linux decades ago was to use TUIs.
All the TUI stuff you can leave running in tmux and then have a consistent workspace remotely accessible from your PC or phone is indeed really nice. IRC, mail, torrents, notetaking, etc.
I stick to CLI's and TUI's and use zellij as my window manager. This way the configs and muscle memory I develop on my work macbook are transferrable to the linux box I use for personal projects, and visa versa.
It seemed a bit limiting at first, but after adapting to it I feel like it was the right move. GUI's represent a sort of explosion in the set of valid states for software to be in, and I find that software which opts out of trying to support that explosion tends to work better, likely because the maintainer just has less to worry about.
Try herdr. I ama tmux then zellij switcher and love it once you minimize the ai crap
Any particular reason to try herdr?
I'm exploring aspects of these "other ways". I've been pushing Kitty terminal graphics compositing and then layering libraries over that. I've also been working on remoting Golang programs over the network as well as compiling them to WASM and embedding that in browsers. This WebGPU browser shader is an example of that mashed together [1]. Same code for native and browser.
Shout out to TUIOS [2] that is in the article; author has a lot of neat libraries and his `sip` protocol is the basis of that remoting work.
[1] https://nimblemarkets.github.io/ntcharts/demos/shaders/
[2] https://github.com/Gaurav-Gosain/tuios
Emacs text interface, Emacs GUI interface, and ... sorry what's the third one? :-)
The only one that matters, Emacs in evil mode.
Emacs Window management via exwm.
ghostty running ssh w/ emacsclient connected to emacs daemon running ghostel
Emacs in ed mode, but you list them in the opposite order from the article.
why would one use the madness of adhoc tuis instead of emacs, I'll never know.
Other ways I often use:
- Using a web browser to connect to a web app running in a Linux VM. I write all my custom tools that way.
- Using VS Code to connect remotely, using ssh
I was hoping the third way would be somehow related to a type of hands-free neuralink so I can code at the speed of thought with my eyes closed laying in bed. Someday the tech will exist but, by then I'm not sure coding will be a thing to do anymore. I'll probably still do it anyway though.
Like the days of Norton Commander
Or Z-19 terminals... green screen, 80 char lines of text. Lightning fast 1200 baud modem!! Those were the days :-)
> What gives you the benefit of a visual interface, without the bloat that comes with Wayland or X11?
fbdev?
Early Reuters and Bloomberg terminals had a text-centric aesthetic.
AFAIK on a PC, when booted into text-only mode, you are still using a bit-mapped framebuffer. From my experiments you can mix graphics and text.
PC is the weird exception in having a character mode at all for compatibility reasons. every other platform does the character mapping in software.
Well, if we count claude as tui.
Ah yes, website talks about "the bloat of x11 and Wayland", and doesn't even scroll correctly on Firefox mobile likely due to some JavaScript fuckery.
Practice what you preach. And practice first.
Not JavaScript in this case, just CSS bloat (backdrop-filter).
another comment on a different HN thread pointed out a similar issue. Now I'm wondering if backdrop-filter can be used correctly at all or should just be avoided entirely
I guess the site scrolls smoothly in a TUI browser?
which, wierdly enough I am useing, and the site scrolls just fine,my browswr has NO java script, No Dom storage, no images/video displayed though there are a few websites that have some kind of alternate way to display images in short video loops that auto play
edit:videos are mp4's
Scrolling is pretty bad on Firefox desktop as well. So bad that I didn't want to finish the article.
Why do people do this?
What do you mean by "pretty bad"? I'm on Zen (Firefox based) and the scrolling seems to work fine. But maybe it's getting resource constrained on some hardware.
Weird. I can't reproduce the slowness you guys experience with fennec nor ironfox.
They only tested in Chrome.
I don't see any suspicious JS on the page, so I suspect it's a styling issue triggering some inefficiency in Firefox. There are animations and text-shadows.
Isn’t that what resolution aware ttys and screen accomplish?