Hacker News Reader

Comment by rubymamis | Hacker News Reader

The problem is OSS software not even trying to compete with the market. People using OSS software taking it for granted that the UX is going to be subpar, and it really is. Regular propriety software faces the risk of their users not paying, therefore adapting to make end user experience great. OSS usually doesn't have that risk. Open source needs to be exposed to risk from end users.

I tried to change that with Notes[1] but I find it hard to live on solely on ads. I tried to incorporate paying for some premium features (like Kanban) but the app is still fully FOSS therefore everyone can compile it from source easily.

I think I'll close-source my next app[2] before it launches. I just can't risk not being paid for my hard work. I also believe the Linux community will benefit from that, since getting paid will allow me to invest more in making UX-focused apps on Linux. I might open source some parts of it tho (or maybe all of it in the far future).

[1] https://github.com/nuttyartist/notes

[2] https://www.get-plume.com

View on HN

There is a large intersection of Linux users and people who cherish their freedoms. For many people, using non-open source note taking software is a non starter because of the vendor lock-in and privacy concerns. Thus, you'll probably lose out more by not open sourcing the program than by open sourcing it and risk having some people not pay. I wager a lot more people is willing to pay than use proprietary software. Also note that open sourcing doesn't require you to distribute the code to everyone, you can give source code only to paying users if you are charging them money, though they will have the right to modify and redistribute. If they don't care about open source, they might as well just use Notion or Evernote. If someone is looking for an open source state of the art note taking program, I recommend Logseq.

There are some programs that sell binaries but also give away the code.

The premier pixel art editor Aseprite is free (source available, not open source) if you're willing to build it from source yourself, but most users buy the prebuilt binaries. I think this way you satisfy both the privacy conscious / cheapskates, and also the developer's economic needs. RMS doesn't approve I'm sure, but you can't please them all.

if the license is GPL you can give away the source and sell binaries. you can even restrict giving the source to those that buy your binaries.

red hat is doing that.

the only limitation is that you can't prevent anyone who has the source from building it themselves and giving away those binaries. but most people who do pay for the binaries won't do that. so whether this strategy works depends on who is your target audience.

if you are targeting developers you probably won't sell much. but other audiences may be fine. i don't know.

RMS totally approves, and selling binaries (as long as they are accompanied by source code) is explicitly approved by the FSF.

Open-source has never been about hating money or working for the glory.

Approves selling software, but doesn't approve making it non-free - and Aseprite is non-free (it used to be free in the past).

I'm a big believer in open source software. I've had wonderful contributors and I'm using qutie a few open source libraries/components in my own apps. That's why I open sourced my own app.

I've tried many different options so I could make a living with it (ads, donations, premium features). There are definitely more ways to explore. But for now, I need to ensure that my financial situation is stable before anything else. Then, maybe in the future I'll be able to afford tinkering with a sustainable way of open sourcing.

I've also started to notice many VCs starting to invest in open source software (just today saw Godot featured on HN). So that can also be a potential route.

It's great that you generally believe in FOSS. You deserve to profit from your work. I wouldn't be opposed to paying (the app looks interesting). But I'd definitely want the security that, should you abandon the project, someone will be able to perform some maintenance. This is why I'm fairly okay with releasing the code late or licensed as "FOSS if unmaintained". Ardour's model of charging for binaries doesn't seem too bad either.

Well, your data will always be just plaintext. So you'll be safe whatever happens.

> This is why I'm fairly okay with releasing the code late or licensed as "FOSS if unmaintained"

That seems reasonable.

I think you're making the right decision. Open-source is extremely difficult to make even a living wage off of - the number of people who do is orders of magnitude smaller than those who make a living off of commercial software.

>Open-source is extremely difficult to make even a living wage off of…

So are you saying it’s even possible? I wish to learn at least one option to do it. Few is even better. How at all you can realistically make living this way?

Ok, not living , just partial income. Ok not partial. Some income?

I am not making fun of I honestly look some way .

I understand that if you maintain 10 years some essential part of essential software project then people eventually come to you for consulting or will pay you to shift this part toward their requirements.

But other then that? Like can you write useful free software program and really have some money from it ?

The main options that I'm aware of are: selling support (RedHat), donations (I can't recall examples but they do exist), open-core (GitLab), and paid hosting (WordPress). In the past, there was also the option of selling physical media (which the FSF did for GNU projects, for instance) but that's been killed by the Internet.

All of these models are very hard to get working (e.g. the donations model requires that you have a very large number of people interested in your work who are generous with their money, kind of like Twitch streaming), but they can work. They're not an option for the majority of programmers, though.

Joey Hess (https://joeyh.name/) did, at least partially. I remember him raising money to develop git annex.

It's definitely possible to make a decent living with FOSS software. One of the main contributors to my note-taking app is @bjorn[1] who developed the awesome Tiled editor[2] and managed to figure out a sustainable way to earn a living from donations/sponsorships. There are probably more instances of that. I tried to go the same route, but other than a $5 monthly payment on Patreaon (thank you awesome contributor!) I couldn't figure it out.

I hate serving shitty Google ads. They are also not a stable source of income and the revenue is low in my case. It's time for me to move on.

[1] https://github.com/bjorn

[2] https://github.com/mapeditor/tiled

Oh, yeah, it's totally possible - just do a HN search and you'll find people who are doing it (e.g. Andrew Kelley, the Zig designer - and he's making a programming language, which nobody will pay for in this day and age!).

However, like being a social media "influencer" (e.g. Twitch streamer), it requires a lot of both hard work and luck, has a limited amount of "capacity", and isn't a viable source of income for most people.

I understand your frustration, but I have a couple of counter-examples, and I don't offer them in any derogatory way.

1) Notes looks a lot like Standard Notes. If you built Notes using copy-left, OSS it's hard to complain unless you pay "royalties" to all the programmers whose work was incorporated into yours.

2) I am probably atypical. I have used Linux for over 20 years, but I can't write so much as a Bash script or a Perl one-liner. Yet I have spent a whole lot more money on software than I would have (or most individuals do) on other platforms. Because I can't code and appreciate the work of others and the freedom it gives me I support the following (not exclusive):

* Kaisen (my distro) and Debian (the distro Kaisen is built on) both

* tFSF for both core utilities and Emacs

* My DE (KDE), and even if I prefer the UI of other applications I mostly try to use those provided by KDE,tFSF, Debian, or Kaisen (e.g Akregator over RSSGuard) unless I donate, like:

* Mozilla (for Firefox), but

* Betterbird for my email client

* Syncthing

* LaGrange (gopher/gemini)

* Joplin

* ClipTo

* and more, including services (e.g. SoulSeek, envs.net)

Some annually, some monthly, some one-time.

In fact, when Standard Notes first came out I paid for a seven year subscription, and then ended up using it for about 2 months before becoming dissatisfied. There are some, like Alacritty, that I use extensively but which I have found no way to donate to, but I try to keep it to a minimum.

There are some amazing people that valued what I'm doing. One person even donated $1000 years ago. But that doesn't make it sustainable.

I think it's time for me to compete on the market. I believe Plume is going to be a good addition to the pool of productivity apps today, but I'll let people judge that when it's released. I hope you will sign up to get an update[2] and let me know what you think of it when it is released.

BTW, are you suggesting I copied Standard Notes? Not really, I don't like their design. Notes is in active development since 2014 (first commit on GitHub in 2015), back then I didn't even know about Standard Notes and the core design barely changed.

[1] https://github.com/nuttyartist/notes/releases/tag/v0.8.0-bet...

[2] https://www.get-plume.com/

"BTW, are you suggesting I copied Standard Notes?"

No. But it reads like I was. Apologies.

Personally I think KDE is light years ahead of macOS and Windows.

In fact I find macOS to have been be of the worst window managers of all the popular platforms (sure it’s pretty and easy to use, but trying to do anything beyond the basics requires magical incantations that are impossible to discover organically).

Each to their own though.

Well, I'm a Qt programmer (and love it!), so I know a bit about this ecosystem. Unfortunately in terms of UX and aesthetics, KDE apps don't come even close to the macOS ecosystem.

But that's alright. I'm here to change that.

Can you share some of this ‘love’ to me? Because my experience with QT so far was : it does ‘who knows what’, ‘who knows when’ and ‘who knows why’ And if you wish it to work the best tactic is “do not change anything, it will surprise you and there will be no way to go back” :)

May be I am missing something important in understanding? I was trying to tweak some projects for my needs. I successfully did it but it was a nightmare.

If you learn how to separate your logic in C++ and your views (GUI) in QML you can achieve the best of both worlds. C++ is fast and I love programming with it. QML is easy and powerful you can create slick looking apps with it with beautiful (and easy!) animations.

I bought the Udemy course of Bryan[1] and learned QML in one day. The next day I already had a prototype for a Kanban[2] that is based on Markdown.

[1] https://www.udemy.com/course/qml-for-beginners/

[2] https://rubymamistvalove.com/notes/kanban.mp4

That second link is the the sort of thing I wish I saw more being pushed as an example of QT. Looks great!

Thanks! I know, it's totally possible to create beautiful apps using Qt. The problems are:

1. Most people in OSS don't care about UX and aesthetics (most of the people using Qt) 2. The Qt Company's examples are absolutely ugly (with the exception of one). If that's what you show developers they would either not believe it's possible to create beautiful apps with it or that it's too hard.

Can you elaborate more? What features you are talking about?

I can criticise Mac OS endlessly and was even actively downvoted here when reported facts of my personal experience with bugs on newly purchased M1 (you can dig the history if you wish). I was shocked how quality degraded in my opinion.

People here were even claiming that it’s my specific faulty machine was to blame. This ‘faulty’ machine by the way works to this very day without any hardware problems except flickering of the monitor which is idiotic in my opinion design decision (to use PWM).

So as you can see I am very critical of macOS and consider using Windows after macOS like completely unbearable. And yet your comment have puzzled me.

When I was learning macOS after years with Windows I remember that certain things were like you say “impossible to discover”. Yet with all that being said I find other GUI implementations in GNU/Linux not any better.

Perhaps I’ve missed something in KDE? What did I miss? What “doing anything beyond basic” means in your usage? What kind of tasks/workflows are those ? How KDE better in those?

KDE _functions_ well, but my god does it need someone with an actual design education to help out.

People with actual design education created macOS and Windows 11. Both of which strip back features in the name of design purity.

I see KDE more like a professional piece of equipment. And when viewed through that lens, I think it’s designed rather well.

To be fair, the upcoming KDE release cleans up a lot of my UI annoyances (like having frames within frames), so it is definitely getting better.

It is very naive to believe that professional equipment do not benefit from well-thought design and UX.

Design does not have to be fancy colours and flashy animations. Design does not mean Fisher-Price UI and massive amounts of whitespace. Ask Dieter Rams.

> It is very naive to believe that professional equipment do not benefit from well-thought design and UX.

I know. I never claimed otherwise.

> Design does not have to be fancy colours and flashy animations. Design does not mean Fisher-Price UI and massive amounts of whitespace. Ask Dieter Rams

I never said it did. You’ve completely misinterpreted my comment.

My point was that commercial operating systems have design specialists and the end result is they suck for a great many professionals. Whereas KDE, despite “not having a design specialist” (I don’t actually know if that’s true, just quoting the GP) feels like it has the same well thought out design that specialist technical hardware does.

I’ve worked closely with designers (because I have zero design skills myself) and seen first hand how an arguably less beautiful design can actually be more intuitive for users

So at no point am I claiming that design is just about fancy graphics. If anything, I’m making the same point you were.

The problem is that most design and UX is not "well-thought" but just cargo-culting and bad intuition. Almost never is there any real user study to show the design improves things in the real world.

To do what? Add more unremovable whitespace everywhere? No thanks!

KDE is the definition of all UI and no UX. Like there's no one actually planning out how a user would/should actually use the thing. That's why I generally prefer Gnome despite it's faults. Even if I don't agree with every decision and lack of customization at least it feels like there's some vision on how it should be used.

A few thoughts:

1. I've never heard of your app, and I follow this space closely. It's extremely competitive.

2. I clicked to check how much you're asking for premium. It's reasonable on an annual basis, but I only saw an option for a monthly subscription. That's not happening.

3. I'd be more likely to give you a donation than to pay for a premium version for an app like this. That model has worked well for Obsidian. I don't think getting early access to features is worth much, but a lot of people have paid them to support their work.

4. My observation is that a closed core product with a large open source ecosystem around it gives folks a reason to pay.

Hi! Thanks for your comment.

> but I only saw an option for a monthly subscription.

What do you mean? The annual subscription doesn't work? Or something else?

> I'd be more likely to give you a donation than to pay for a premium version for an app like this.

Well, for my Notes app, I can understand. Hopefully you will find my new app as a breath of fresh air in this space. Sign up to get an update please! I would love to hear what you think when the app is out.

> My observation is that a closed core product with a large open source ecosystem around it gives folks a reason to pay.

Overall, I think people will pay if an app gives them enough value.

There are definitely some challenges with getting people to pay for OSS, but I wouldn't be so quick to blame that all on what happened in this situation. I think you'd be experiencing the same (or worse) if your app was closed source for reasons I'll mention in a minute.

I'm the exact type of person who would be in your target market and likely to pay, but aside from having never heard of your app (which may be your biggest problem), these are the reasons I'm not buying. I'm only sharing this in the hopes that it helps you, not trying to make you feel bad or anything like that. Overall I think you're awesome for making your app open source, and I have a mountain of respect for you for doing that.

1. You're targeting a very crowded and competitive market where top-quality options are 100% free. This would be a huge challenge regardless of whether you're open source or not.

2. Your app is still very young/new. Winning in such a saturated area is going to take some time to even build awareness, let alone get mature/feature complete enough to get people to change.

3. With something like notes, you're looking at people with substantial pre-existing sources. Migration is not trivial either, so there's a big barrier of entry there for people to use your software.

I wish you the best, but I think you'd only be harming yourself by closing your next app. You definitely lose people like me for whom my notes are incredibly important and I won't risk losing them to a proprietary tool. Open is mandatory for me to even consider it.

I'm pretty happy (maybe even in love) with Logseq, but the performance is definitely a downside and the idea of a C++ app is highly appealing to me, so I'm going to give your app a try.

Side note: If your GUI was compatible with logseq's on-disk format, that would make your app quite compelling to me

I appreciate your opinion.

Notes has been quite a success for me. With over 1,300,000 downloads[1], and being featured in awesome tech websites[2]. Just on Ubuntu alone there are 7000 weekly active users (that have downloaded it from the Ubuntu Store).

My new app Plume is going to be a big improvement over Notes. First, your data is and will always be as simple as Markdown/plaintext. My new block editor is based completely on plain text while still offering advance views (with a minimalistic syntax). It's also faster than any other in its category, even faster than native apps on macOS! I'll soon put benchmarks on the website.

All I can say is, sign up to get an email update and check it out when it's ready. Most of the features will be free with only some advance features paid.

BTW, what do you mean by "logseq's on-disk format"? That it's local-first? If so, Plume is also.

[1] https://tooomm.github.io/github-release-stats/?username=nutt...

[2] https://www.omgubuntu.co.uk/2022/09/open-source-qt-notes-app...

Well dang! I didn't realize it was that big. That is impressive and definitely changes my opinion a bit. It's quite unfortunate to have such low paid numbers with that many downloads.

By logseq's on-disk format, it is local first but it's a stricter subset of markdown with a handful of extensions. The strictness is because Logseq parses all the markdown into an in-memory graph database, and in order to make that work with markdown the top-level has to be a list and needs some IDs and other identifiers in order to properly link/embed blocks and pages and such.

Thanks! With Plume I use a tree data structure (just with parent child relationships), although in the future I would probably use Boost::ptree.

The things is, the underlying data structure is just plaintext, so having an ID for each block will make the plaintext look out of place (say, if you wanna move your files to another app).

But I'll add some special syntax to allow this feature for specialized blocks.

> so having an ID for each block will make the plaintext look out of place (say, if you wanna move your files to another app).

Indeed, this is definitely a problem for Logseq. For example here's some of my recent Logseq notes. This is one page where you can see an "ID" (represented by `id:: <guid>` is created. Then helpful images/diagrams are displayed. This is then referenced later:

    - Parts of a [[Neuron]]
      id:: 64f4bdc3-29de-4b6c-bc41-20a5f052badf
      - Dendrite: https://en.m.wikipedia.org/wiki/Dendrite
        - ![Anatomy_of_neuron.jpg](../assets/Anatomy_of_neuron_1693760694470_0.jpg){:height 411, :width 718}
      - Axon: https://en.m.wikipedia.org/wiki/Axon
        - ![Blausen_0657_MultipolarNeuron.jpg](../assets/Blausen_0657_MultipolarNeuron_1693760712899_0.jpg){:height 471, :width 718}

Then in a different page, this is referenced by ID. In this case it is "embedded" but you can also reference and other possibilitie:

    - #vocabulary [[Brain Myths Exploded]] #Notes
      - {{embed ((64f4bdc3-29de-4b6c-bc41-20a5f052badf))}}
      - Encephalization Quotient (not a good measurement)

That's actually not too bad, and similar to what I plan to do. They only create an ID for a block/bullet-point when necessary.

I just noticed Logseq doesn't support selecting and editing text across multiple blocks. It's indeed not easy to do, but Plume does support this, I believe it's an essential feature for a text editor. Notion took a long time until they supported this[1]. I implemented it in less than a month using C++ and QML, a testament more to the strengths of Qt than to my own skills as a developer.

[1] https://www.notion.so/releases/2022-01-19

> People using OSS software taking it for granted that the UX is going to be subpar...

After using the abomination known as Windows 11, I'd argue that the UX of Linux is well beyond par, especially when par for Windows 11 is showing Microsoft ads at you while you're using almost every application.

I am in an interim CTO position and the year before the title of a role I acted as had "Principal" in its title. I still write Commonmark files to render them to HTML and PDF with Pandoc.

There are dozens of Markdown note taking apps out there. There are half a dozen Notion clones out there. I know which one will overwhelmingly win mindshare due to a special feature but I won't give away why (the feature itself is more valuable than a Notion clone or even Notion itself).

You might be able to get paid for providing something valuable to others. Sometimes hard work is involved in providing value to others, but performing hardwork doesn't necessarily mean you have provided value to others.

> but performing hardwork doesn't necessarily mean you have provided value to others.

Very much agree. I guess only time will tell (when I launch).

Notes is a beautiful app! Wow.

Regarding Plume: just from that landing page, it looks 100% like Notes. Anyway, I'm interested to learn more about it when it launches!

> Open source needs to be exposed to risk from end users.

Why do open source applications need that? You said yourself that users of open source software accept the subpar UX, so what problem are you trying to solve? Making open source development somehow 'riskier' for programmers without simultaneously providing more reward would only drive programmers away.

Users of OSS software accepting the subpar experience is the problem. You don't think they look away to their macOS counterparts and wish these polished apps were available on their Linux machines? I used Linux as my primary device for 2.5 years, then switched to macOS and never looked back. Now I want to bring the same polished user experience to other operating systems using Qt. I believe my latest attempt [1] is the closest to get there without the awful sacrifices of Electron apps.

[1] https://www.get-plume.com

I really like the functionality of your notes app, but I don't think emulating Apple UI design is the solution to subpar UI on Linux.

Linux users expect basic desktop conventions like toolbars, dropdown menus, and not mobile design concepts like cramming everything into a hamburger menu.

Have you even tried using my app? It doesn't have any of the "mobile design concepts" that you're talking about.

> cramming everything into a hamburger menu

- Import/Export

- Check for updates

- Start automatically

- Hide to tray

- Change database path

- About Notes

- Quit

Is that EVERYTHING?

No. It's not. I don't like using hamburger menus for everything myself. This is why the options are minimalistic. Look at other Gnome apps using hamburger menus, no need to go as far as "Apple UI" (which actually discourages hamburger menus for desktops apps).

What's the difference between "notes" and "plume"?

Hows it different from Obsidian or any other note taking app that has a larger community. I'm sure OP put in lots of hard work and it looks good, but it's not apparent to me why I should abandon my existing platform (stickiness) for another one. There are a lot of note apps out there already which probably makes it hard to sell.

Plume's editor is a block-editor I wrote completely from scratch using C++ (model) and QML (view). This allows me to create an advance editor that can have complex elements (like Kanban, Columns, advance image components, etc) within the text. The performance using Qt C++ compared to Electron is a mssive improvement over current web apps (like Notion and the likes), and actually, i's even faster than native apps on macOS (Craft, Bike) and apps made in Flutter (AppFlowy).

Example: https://imgur.com/wajSjbn (I'll implement the Kanban feature next week (:)

Okay that's actually cool, that you switched to something else than electron, but it looks nearly identical from the screenshots, that's why I asked.

Thanks for sharing. Appreciate your fighting the good fight, and sorry for the results so far. Maybe apply for a Futo scholarship for one? But definitely close source for Plume.

No need to be sorry, really. I'm just learning how to navigate this world. I've been pretty lucky so far. Notes is one of the top results in Google (for the keyword "notes"). So I get a lot of traffic and a nice passive income. But it's not something I can fully live on (plus it's unstable and I hate serving ads).

Hosting?

What do you mean by hosting?

on your ‘com’ site mentioned in github page download button does not offer (near by) payed options. Why? Why it is only in the separated Pricing link? Why not to present the same choice in download ?

Thanks for your remark! I never thought about it. I suspect it will drive down the number of downloads? How do you think it will be beneficial?

I think it could be beneficial because it would be easier for those who wish to pay. They would not have to do extra step looking for that opportunity.

I personally do not like (in soft words) when button says one thing and delivers something else. So I think renaming Download into ‘Download options’ were you give choice download ‘ free’, ‘donate’, ‘pay for pro’ etc as you like is acceptable and feels good enough. I do not see why in that case it should reduce number of downloads but every prediction is always a prediction and nothing is better then the real life test.

What does this have that Joplin doesn't have?

For one thing, it’s a lot prettier than Joplin

Personally I think Joplin looks better and this looks like a poor macOS clone that lacks many of the features like markdown support or synchronizing across devices including mobile. IMO if Joplin’s problem is the UI, or even just something OP is passionate about, why not contribute and open a PR? Why is the pattern “I care about this one thing so I reimplemented it in a new app from scratch”?

- Plume supports Markdown (plus additional plaintext syntax).

- Plume is a block editor, unlike Joplin. Therefore plume can have advanced views inside the editor itself (like Kanban, image handling, columns, etc)

- Plume is faster than Joplin's non-native editor (the one that renders your Markdown)

- IMO Plume is way prettier with far more attention to detail.

The CLI is peak UX for anyone technically competent.

I agree but CLI just can't cover all use cases - just on top of my head are graphics, animation software, video editing, circuit making...

Well, they think it is. Most of the time, it's not. Command-line tools have terrible ergonomics for many workflows and use-cases.

For a subset of tasks, maybe. But I'd rather use Firefox than lynx. And I'd rather use Thunderbird than mutt or pine.

I get Firefox (with keyboard navigation) over Lynx but Mutt has been fantastic for my productivity.

Utter nonsense. The CLI lets you do complex custom things that you sometimes can't do with a GUI. But it is pretty awful from a UX point of view. Terrible discoverability, terrible UI, very unfriendly, no contextual help, etc. etc. I can't think of a worse interface from a UX perspective.

Terrible discoverability? A shell with tab completion is (in my opinion) a lot easier to navigate than for example Photoshops menus.

Really? Do you communicate with ChatGPT by pointing at pictures?

That is a REPL, not a CLI.

A shell is a REPL. Though I do I agree there's something to using one with good features (ex. I would completely understand someone using exclusively zsh just for the better tab-completion).

A traditional UNIX shell is more of a poor man's REPL, hence the difference in wording.

Conversation mode, where you talk with it, is quite an improvement over having to type at it, especially for slow, hunt and peck typers.

What has ChatGPT got to do with it? We're talking about Unix style CLIs.

The manual exists for a reason. I find it much easier to type `man program` then `/whatever` until I find what I'm looking for rather than scanning the screen for the correct icon. There are also some conventions and after a while you get to a point where it's intuitive to use any new program. Just like with a GUI and icons, what does the wand do?

"You just have to read the manual" and "you get used to it" are basically synonyms for "the UX is bad".

1 second of googling: https://fstoppers.com/opinion/stop-telling-people-read-manua...

I'm sure there are better articles if you search more. If you have time for a book, I highly recommend "The design of everyday things". If you don't agree with me that book will change your mind.

You get used to it as in "used to how CLI programs are usually designed". Just as you get used to what the select icon in a GUI does. Without any prior experience in any both seem like alien technology, but coming from GUI to CLI you have the bias of already knowing one when comparing.

People claim macOS is intuitive and everything works the same, but I'm handicapped in it. I just call it inexperience not being bad UX.

Regarding the manual, does GUI programs have a unified way to access guides and documentation? Can I scroll to the bottom with G and find the configuration files and example uses?

The difference is, you can teach a user who's never seen a computer before to work with a GUI in an afternoon, maybe a weekend -- and if the GUI is well-designed, the skills they learn on one program will transfer well to others.

Meanwhile, there are XKCDs about how even experts can't remember how to work tar: https://xkcd.com/1168/

Before 1984, computer UX was, universally, sucky. As Alan Kay put it, the Mac was the first computer worth criticizing.

Ease of use isn't the only metric for "peak UX" though, composability and how the programs interact should also be taken into consideration. Selecting a filter on instagram is very easy, with the caveat that you're limited to those filters alone. The subject is more complex than clicking buttons vs typing commands.

That assumes that you can already articulate the action you are looking for in the program's vocabulary. For people who think more pictorially, they may have a sense of what they want but benefit from visual cues.

I don't target specific options, but different words of what I'm trying to achieve, crop/cut/extract as a pretty bad example. It takes me to the description of the option I'm looking for where it will also have useful extra options to combine.

I can't say the same for GUI icons, many times hidden behind dropdowns. If I hover it will say select, then I have to find how to crop that selection, behind another dropdown, probably a few levels deep. And none of them searchable. If I need help there's no one place to find a guide, my best bet is google or the website.

For both an intuition has to be learned from use, but GUI people act like they were born with computer literacy. I used to think the CLI was stupid as well, but now that I've seen the light I can't imagine going back.

In principle maybe, but all present implementations fall a bit short. Why can't I use my mouse to position the cursor in any shell? Placing the cursor with the mouse is something people commonly do in terminal editors like vim or emacs; the efficiency arguments against it fall flat because this is entirely optional and besides, many people use laptops with either a track-point directly in the middle of the home row keys or with a touchpad directly below the keyboard.

And there's no technical reason it couldn't be done; years ago I once did it as a proof of concept. It just hasn't made it into any major shell yet.. besides eshell anyway. It does work on eshell.

Anyway, it's not really a big deal, not important enough for me to pursue, but I do think it counts as a counterexample to disprove the "peak UX" claim. Command line interfaces in principle are great but in practice we're doing it by emulating hardware that went obsolete decades ago and piling cludges and hacks on top of that to make it feel a bit more modern.

Another example; there are a few ways to inline images in a terminal emulator but they're all kind of shit in their own ways; graphical thumbnails have clear utility when listing some directories; every serious file manager supports it but to do this in a terminal emulator is cludgy hacks.

It’s not done because there’s no use case. No point in writing anywhere and clobbering existing text.

Nonsense. Fixing a typo or amending a command somewhere in the middle are obvious use cases.

That’s only the last line, not quite what they stated. Still a lot slower than emacs or CUA keys, so no it doesn’t fall flat.

The kind of person that would use it doesn’t use a terminal.

> That’s only the last line, not quite what they stated

Then I wasn't clear, I am talking about using the mouse to position the cursor in the current line being edited. If I am writing a long command and realize I made a typo 50 characters earlier in the second argument, I should be able to click in the second argument to reposition my cursor. And yes, I do use vi-mode and know how to do this the vi-way with just the keyboard, but Vim also supports using the mouse for this and there's no reason shells shouldn't as well.

> The kind of person that would use it doesn’t use a terminal.

Absolute nonsense.

Nonsense—a shell is not a full screen ncurses app like the others you've mentioned and probably wouldn't want it to be due to having to carry the scrollback buffer.

You're the only person to implement this slower/problematic/nifty design in five decades, on a FLOSS operating system. Just a guess but maybe your attitude is also part of the problem.

We should have kept using MS-DOS...

That's what people who gained their technical competence in a CL environment always think: we're always most productive in the interfaces we are familiar with. I'll agree that most technically competent people probably prefer using the command line for many tasks, but correlation != causation. It's really easy to mistake our own preferences for being objectively good, generally... which is why designers exist, why most user-facing commercial software doesn't require reading a single line of documentation, and why most FOSS projects-- almost entirely indifferent to or hostile towards deliberate design-- are only useful to people who have a working mental model of what's happening under the hood.

I was strictly a vim guy for a decade and a half... my coworkers occasional snicker merely steeled my resolve, but I knew I was doing the pure and efficient and technically correct thing by sticking to barebones vim. Lots of vim stuff and ex commands are beneficial to my workflow, but after a coworker convinced me to try out some more modern options with vim modes, I realized that my attachment to vim as an editor was driven by my emotional investment in how many times I struggled with vim's clunkiness. I wore it as a badge of honor. I thought it gave me cred. In reality it just showed how utterly inflexible I was. The Jetbrains editors, for example, offered more functionality out of the box than I could practically cram into my vim setup, yet was completely configurable, and the features were generally intuitive and discoverable. I didn't have to look up how to do some relatively obscure thing in ex-- there was probably a menu option for it so I just didn't have to keep it in my head. Sure, vim is and always will be my default editor for light editing, and there are obviously people whose use cases are perfectly satisfied by using it. However, assuming for years that I was more technically competent for doing intensive professional work in complex codebases using that clunky old editor was plain old self-indulgent arrogance, like most other nerd badge-of-honor things are.

I've been using command line environments daily for decades, and usually head there first to do many tasks, but many of these archaic tools only persist because of a reverse eternal-September problem. By the time someone gets technically advanced enough to start influencing how our technical environments work, the curse of expertise obfuscates their downsides and they mistake their own comfort with something for it being objectively good.

The best CLI will never have the discoverability of any competently designed GUI.

Reply on news.ycombinator.com