I’m curious what the difference is between that and having pinned windows in Ubuntu
For example I’ll have sublime text, one terminal, a notepad, and a firefox window with multiple tabs, pinned to the four corners of my desktop
Everything You describe is possible without emacs so I’m genuinely trying to understand the appeal
The advantage is that it’s a single program with multiple windows with a shared list of buffers. So you can save transform your layout and save it, manipulate the content without any real boundary, unified bookmarks,…
Your WM can’t access sublime and firefox internals. And the last two can’t talk with each other.
But why would I want them to as the default interaction?
I can ctrl-c > ctrl-v from one to the other trivially and ffx has console and inspector internals visible with one Function key press
Some find it useful. One can write code in vim or sublime, the only differences are the specific actions to get it done and the user experience. People will lean towards one another, not for the capability to write code, but for how it’s done.
So what I’m trying to showcase is not the capability, but the Emacs ways of getting it done. You can also implement something with a WM setup or tmux (with w3m). But it’s a different feel and effort, so people like ones and dislike the others.
Although I don't browse the web with Emacs, I see the utility.
Tell me: How would you select some text, and with a keystroke, send the selected text to yourself via email, with the URL included and the title of the page in the subject line?
I think the takeaway for me is that emacs is for people who get enjoyment out of building their own custom computing environment
The example you describe is something I would rarely do so building a process around it would be inefficient for me but if that’s where people want to spend time then great
For long time users, that's the whole point of Emacs.
> The example you describe is something I would rarely do
I used to do it at work all the time (not with Emacs, but with AutoHotKey). Find a long article I'd really like to read, but don't want to waste time at work. With a keystroke, I'd quickly send the email to me with the URL etc.
There is some tacit knowledge in here that is nearly impossible to convey in words. Your "rarely do" is the thought that subconsciously ignores the principle of least effort; there's an affordability gap, e.g., I am more likely to use a thesaurus simply because I have it built into my text editor, versus someone who has to reach for it in the browser/separate app.
> building a process around it would be inefficient for me
That's what you're missing. Emacs makes it incredibly cheap. It's a Lisp REPL. When you understand what it means in practice, it opens up incredible opportunities for exploratory computing. You're not "building processes around", you're essentially living in one mega-process, a simulation, a digital universe where you are a god, commanding change at will. Right now, as I'm typing this very comment, I noticed a gap in the way how I'm using LLMs, so I'm changing this stuff on the fly, with the agent running things in the adjacent buffer while I'm finishing this comment. The crazy thing is that it doesn't have to compile, restart, ask me to pause typing, etc. It's applying changes to my editor live. By the time I finish typing the comment, it will probably have already built this feature, tested it out, and it will be ready for me to use, instantly.
Indeed, this describes it better than anything else possibly could.
I've been carrying a Leatherman around for almost a decade. My friends make fun of me for it, and when I tell them what I do with it on a weekly basis, they always say they wouldn't need to do those things.
Yeah, I didn't either. Until it was trivial to do it because I happened to have it on me all the time.
It's like explaining to someone over 20 years ago why you "need" a camera or GPS in your phone. They wouldn't get it. Especially not if you tell them it costs $700+ "I already have a digital camera and who needs a GPS outside the car?"
The difference is the text consumption. Can you easily retrieve any text you're seeing on the screen into your editor?
Inserting any piece of text I see on the screen into my notes? A couple of keystrokes. I can grab text in any form, even if it's a frame in a YT video - I'd just select the area, snapshot it - Emacs OCRs the text out of it, which takes milliseconds. I fully control the playback of said video directly from my editor - rewind, pause, mute, speed up, extract transcript, etc. It is enormously liberating and empowering.
I have complete and utter control of text - I can send a query to an LLM in the middle of typing this very comment, without ever missing a beat. I can ask for anything - fact-check it, spell check it, reword it, explain it, give me different variations of the same sentence. I can replace part of it with diffing in-place. I can even start changing the behavior of Emacs itself, right there on the spot.
I can directly grab any part of the log of a process running in the terminal. Grep through it, semantically search through it with an LLM, whatever. Similarly, I can directly grab a piece of any text in Emacs and send it over to another app, e.g., as a command to run in the terminal.
> Everything You describe is possible without emacs
Everything is possible with Turing-complete languages, but Turing-completeness is a trap. It's all about affordability. What Emacs grants as a Lisp-REPL loop with a built-in text editor is lowering that affordability bar to the lowest imaginable minimum.
Have you ever thought: "What if I could change the 'Open File' dialog in Sublime so I can browse GitHub repos directly instead of folders and files?" This kind of thinking has probably never occurred to you - for an Emacs user, that is some prosaic, just-another-Wednesday kind of thought.¹
---
¹ https://www.youtube.com/watch?v=BDn8ak-8yIs
By the way fishing that url out of my browser history took a keystroke - of course Emacs can search through my browser history too - https://github.com/agzam/browser-hist.el
The example you describe is something I would rarely do so building a process around it would be inefficient for me but if that’s where people want to spend time then great
It's just a drop in a vast ocean of things. Your close-minded thinking just sees one thing and makes generalized assumptions.
> where people want to spend time
You're reading it completely backwards. This is the thing that I never understood about the majority of programmers. For whatever reason they just readily accept the status quo. Here's a typical example:
So often do I feel like rolling my eyes whenever I pair with my colleagues - I'd do something trivial, like fetching a list of PRs related to a specific ticket number at the cursor. They'd be like "whoa, that's cool", and then never do anything about it - they stick to their "learned helplessness" paradigm - copy the ticket number, switch to browser, navigate to Jira, pass through SSO, look around for the phone just to push the button, paste the ticket number, find the linked PRs, etc. The other teammate watching all that may say something like: "you could do it faster directly on GitHub"... And here I am - pressing a key, voila - the list. Why the heck don't they do anything about it, I just don't get it. Trained engineers, they spend years dealing with cranky software, why in the world are they unwilling to do anything about these seemingly small annoyances? Why do so many programmers treat software as if it's magic? And I think I know the answer: because the software you use shapes your affordances. If there's no downloadable get-me-list-of-prs-for-a-jira-ticket VSCode extension and no simple way to build it, it would never occur to them to be even annoyed. In Emacs, I can write a picker in a scratch buffer by hand and it would take me minutes. Today with LLMs, I can just spend fifteen seconds, not even typing, speaking my thoughts and it goes to build it, run a sequence of e2e tests and hand it to me on the spot.
Sure, go ahead, talk about time. Make me chuckle. The irony is that the time you spend using a Lisp-based system never feels wasted; it only gets better, it has a compounding effect. It's the time you waste convincing skeptics that it is in fact enormously pragmatic and time-saving that feels like tilting at windmills. Why am I even still talking to you here, if it was clear from the beginning that you didn't in fact have good-will curiosity, only simply wanted to find evidence for your already settled and stiffened mind.
> Inserting any piece of text I see on the screen into my notes? A couple of keystrokes. I can grab text in any form, even if it's a frame in a YT video - I'd just select the area, snapshot it - Emacs OCRs the text out of it, which takes milliseconds.
Yes, Mac does this natively/automatically with screenshots. I'd be surprised if I've ever used/needed OCR more than a half-dozen times. Far more often it just gets minorly in the way when I want to copy-paste the image itself.
> I have complete and utter control over text - I can send a query to an LLM in the middle of typing this very comment, without ever missing a beat. I can ask for anything - fact-check it, spell check it, reword it, explain it, give me different variations of the same sentence.
Mac natively offers AI writing tools context menu options in every app for text selections. Right click -> search effectively does the same thing (every search engine has AI functionality featured prominently by default). I believe most browsers also offer straight-to-AI equivalents as well.
> Similarly, I can directly grab a piece of any text in Emacs and send it over to another app, e.g., as a command to run in the terminal.
I can do this natively and trivially on my Mac in every single app without having to learn a bizarre alternate universe of awkward, inconvenient hotkeys.
> What if I could change the 'Open File' dialog in Sublime so I can browse GitHub repos directly instead of folders and files?"
My IDE has featured this prominently as a native option for approximately forever.
> What Emacs grants as a Lisp-REPL loop with a built-in text editor is lowering that affordability bar to the lowest imaginable minimum.
It's literally not affordable to spend my time inefficiently re-inventing the wheel on every common UI/UX task in a usually fundamentally limited and broken way. I need an IDE where it's not a major source of controversy as to whether it's possible to configure basic auto-complete.
> Emacs can search through my browser history too - https://github.com/agzam/browser-hist.el
Having to install a random third-party extension to use a basic browser feature is a bug, not a feature.
...I'm not sure what I'm trying to accomplish by poking these holes. By all means, enjoy using emacs. It's just...strange to me that an emacs user would believe these are selling points when in fact having to configure them myself is a major downgrade from just using the native features in commonly used software.
You will never see the appeal until you give it a heartfelt try yourself; there's too much tacit knowledge that is not easily describable in a few short paragraphs.
You're making it sound as if all the things you're talking about are unknown to me. I wasn't born into Emacs, and I didn't find it in my college years and just decided never to leave. I've been staring at computer screens for decades and have been using Emacs for only a fraction of that time. The only difference between you and me is that I have now seen both sides, I'm intimately familiar with your worldview, and I know the reasons why that side sucks. I'm not saying the Emacs-side is perfect and that there are zero tradeoffs to consider, no.
Lisp and Lisp-based systems have been an impossible sell for a long time, and mostly for emotional rather than practical reasons. "Emacs user" sounds like a lifestyle, and some people outright avoid it just so they don't have to join the tribe.
Those who genuinely try to evaluate the practical side out of curiosity pick some random features or one demo and typically remain unconvinced. I have never heard anyone say "I watched this video/read an article and I decided to try Emacs", because the value of Emacs does not live in any single feature. It lives in how the features compose, and composition is invisible in a demo.
Even using Emacs for a few weeks may not help; the value builds up over time. Each piece on its own looks, sounds and feels unimpressive. Nobody is ever persuaded by a sum they have to imagine. There's no frame that I can use for you to even listen to me; you don't have one. It's very much aligned with Graham's Blub paradox - from inside a less flexible tool, a more flexible one looks like the same thing with weird extras. You only see the difference by looking back from the more flexible side. Most people have never experienced an environment they could change while it runs, so they have nothing to compare it to. My enthusiasm reads to you as sunk-cost bias, even though it clearly isn't. I would happily switch away from Emacs whenever a clearly better alternative comes around. Alas, it just doesn't exist.
> in fact having to configure them myself
Yesterday, I would've said: "Oh sweet summer child, do you really believe you don't configure the 'commonly used software'? You're configuring your own brain and habits to match that software instead..."
But today, that's not even an argument. You can just install Emacs and ask a good model to build you a solution to a problem and then observe and evaluate the workflow differences, easily. Then it opens up a whole world of possibilities where the only limit is your imagination.
"Right-click -> search effectively" is one thing, but when you have a system where you can bind it to a key, to something more complex than a simple right-click, to a foot pedal, to a gesture, to a voice command, to a "Fus Ro Dah" Skyrim call - you just laugh at your own younger, naive self for being so dismissive without even knowing what the hell you were talking about back then.
Hahaha... So you think thousands of people using it are stuck in it because they hate themselves and I just love inconvenience and inefficiency? What odd masochists they are, these weirdos with zero skill of bashscripting. Total losers.
“batshit insane list of hotkeys”
..Is exactly the part that the emacs user likes working on, they get exited and look forward to adding those configs because it’s like world building that they get to live inside and configure to be their full interaction layer with computers
You have to think about it more like, they like being cypher in their own personally built text based matrix
People write these things, and use these methods, specifically because they don't conceive of usefulness the same way you do.
In this case, the justification offered doesn't mention all the various keybinds that have been pushed down to the "how" section. But clearly the simplicity is a feature:
> D will download images locally, b goes back a page, a will save bookmarks. All of this makes browsing in eww truly enjoyable.
The point is the same as usual for TUIs: you don't have to switch over to the mouse and hunt out a button, as the necessary key is always literally at hand.
People who find this sort of thing "truly enjoyable" (what clearer statement is there of featurefulness!) commonly take it for granted and don't explain the virtues of it very well. That sort of explanation is found in think pieces about TUI design more generally.
TFA describes a goal of avoiding distraction. That is not at all the same thing as "inconvenience and impracticality". Clearly the author finds these keybinds quite convenient and practical. It's just that they don't cover a bunch of web interactions that you perhaps value, but the author doesn't.
You’ll have people who demonstrate their whole set up and interactions and hot keys and shortcuts and everything and then when you say OK well why would you do it like that it all boils down to:
“This extremely precise implementation works for only me and my specific workflow”
I found that it’s almost more useful to think about emacs users showing off their org mode bindings are basically like Warhammer players showing off their painted figurines
That's their value.
But also, consider this: maybe they just have very specific, unusual needs that you're unable to relate to because they're very different from your lived experience. I'm not saying this as a criticism. I feel like most people struggle to comprehend just how different someone else's lived experience can be from their own. I certainly do. Usually, when I read these emacs setup blog posts, I'll dislike most of their setup, but often there will be some nugget in there I'll absolutely love, and then I might add that nugget to my init.org.
Let me try to explain my singular point of view. I have very severe ADHD-C, and GUIs always felt horrible for me to use. Especially big complex GUIs. Things like MS Office, Photoshop, big IDEs etc.
Why? Multiple things:
1. Visual processing:
ADHD can include strange visual processing or spatial reasoning problems. It's very hard for me to describe this one in a detailed manner, but suffice it to say, Icons don't make sense to me. Text does. I am capable of learning some icons, certainly, but certain ones never stick. For example, the protonmail web client has the reply and forward button displayed as arrows pointing in opposite directions, I always click the wrong one. Every time. And everytime I'm positive it's the right one and click it immediately. As I'm sitting here now, looking at an HN textbox, I can reason my way into understanding that to the western mind things flow from left to right, and so left should be reply and right should be forward. But when I'm actually there wanting to reply to an email, I hit forward every time without thinking. If on the other hand the buttons just said "reply" and "forward", this wouldn't happen. I can't explain it even to myself really. That's just how my brain brains.
2. I lose my train of thought extremely easily:
You know how sometimes when you walk into a room, you suddenly forget why you're in there? This supposedly happens to most people occasionally. For me it's more like 15 times a day, not just with rooms, but any sudden sensory transition. Like a pop up window. Ah yes. I'm in some stupid baroque GUI program. I realise I need to change a setting. I locate the settings button from the dropdowns, which every program seems to have in a slightly different place, big settings window pops up covering most of the main window. I immediately forget why I'm in the settings. I try to remember why for half a minute, and fail. I close the settings window and try to remember what I was doing before. 2 minutes later I remember which setting it was. I open the settings. I forget why I opened the settings again. Ok, annoying, but not that big a deal right? Oh but you see opening the settings is just the first roll of the dice. The settings window has 15 different tabs. And some of the tabs have buttons that spawn new sub windows, and even some of the subwindows have tabs(looking at you, Microsoft. What is wrong with you people?!). I call it menu purgatory.
3. Extraneous information distracts me very easily.
This one should be self-explanatory
4. My hand eye coordination is not fantastic, so I often click the wrong thing, which then feeds back into the wrong thing, which then feeds into 2 and 3 again. Keybindings don't require coordination, only muscle memory.
On a bad day, with a shitty piece of software like that, I can spend literally half the day locked in a cognitive battle of attrition with the software instead of getting work done. That's not an exaggeration.
Because I'm so hampered by what to other people is just minor annoyances, I need software that I can configure completely arbitrarily. No change can be off limits. That is almost never the case with GUI software, at least without cloning the source and patching it, figuring out the build system, possibly learning the implementation language. With Lisp software, the config file is really just a hotpatch. And Lisp is my favourite language, so yay. I would prefer Common Lisp, but beggars can't be choosers and Elisp isn't that bad once you learn to avoid the footguns. Certainly better than vimscript, which I've always found incomprehensible.
So that's why my setup is roughly this:
Never use GUIs with the exception of Firefox(I want to use Nyxt but I could never quite get past the learning curve, and there are still key features missing) and occasionally mpv(which is barely a GUI anyway). Everything else, find a good TUI. If you can't find one, find a shitty one for emacs and bend it to your will. This keeps (auto-) defenestration at a minimum.
Yes clearly.
I’m glad you’re enjoying your set up.
The keybindings aren't mentioned in the "Why?" section of this blog post, and clearly aren't the point of the post. You can get extensions to add Emacs bindings to full-featured browsers, and there's even the Nyxt browser which has Emacs keybinds and Lisp configuration.
The core reason listed in this blogpost for browsing this way is the very fact that it is less featureful and less convenient. As someone who uses Emacs, I know that eww isn't as convenient as a fully-fledged web browser and I know that sometimes the hard way of doing things can be beneficial. I found this blog post disappointing because it doesn't afford you any new information, nor tell anyone about any actually useful features about browsing in eww.
It wouldn't be unconventional if most found it useful. That doesn't mean the author themself doesn't find it useful:
>> I have seen in my own use of the modern browser (chromium/firefox) that it is far too easy to get distracted
They find EWW useful to cut distractions.
>> In addition to this, my belief is that we should make strides toward leaving behind Layer 8 of the internet - the limiting frontends of social platforms and locked away corners of the net that limit actual discourse (Discord, I’m looking at you). [...] The internet is just a delivery mechanism.
They also find it a useful way to stick to their ideals which I understand to be that they should be in control of the front-end and the internet is only to be treated as a delivery mechanism of information.
So yes, it's a focus and mindfulness thing for me, but not because it's obscure and difficult. Rather, it allows me to keep one set of commands and one text manipulation paradigm cached in my head—that of Emacs which I'd rather use anyway—and do everything with it. For browsing tasks that rely on JavaScript and are not amenable to eww, I use a regular browser.
I once worked with a guy whose brain was so ADHD-cooked that he literally had five screens going on at once: his Mac's built-in display and four external displays. I was like, dude, how do you work this way? He was like "I have to have this many screens due to all the documentation I read." And he was constantly switching back and forth between all of them. Madness. How did anyone write any halfway decent code with just a single, 14" CRT monitory that could do 80x25 characters?
I think the idea that you do something in a more difficult or less convenient manner for the sake of "mindfulness" or "focus" is well-established, and what I hope for (and am always disappointed by the lack of) when reading such blog posts is an actually useful feature of the unconventional method in question.
The title of this blog post promises some usefulness from using Emacs as your browser, and upon reading, you discover that the "usefulness" is the inconvenience and impracticality of using it that way.