Comment by the_lucifer

10 hours ago

> I think it definitely trades user-friendliness for power and flexibility

A tale as old as time with OSS software from what I've seen. So far the only exception that I've seen of a FOSS piece of software with great UX is Blender[1]. With UX not being a first-class citizen on Linux, exposure to a solid universal tooling to create great UX also exacerbates things.

I think the issue of bad UX and OSS are comorbid because good UX needs an extremely opinionated mind at the helm and a culture of prioritizing features and saying "No, this is not important" which OSS generally lacks.

[1]: https://www.blender.org/

I don’t know about that.

I think Libreoffice is easier to navigate than ribbon-era Office.

I think especially Gnome, but also Kde has better settings than Windows, which is spread across a new and a classical view (or more?).

I also think MacOS finder is worse than Windows/KDE/Gnome equivalents. When you strip so much away that people have to remember strange key combinations to jump/navigate to specific directories you have really gone too far.

And mobile UX (android and iphone) seem to me worse than any desktop system. Nothing is consistent (sharing data between apps), nothing ever seems settled (settings move around in the central settings apps every few releases).

We could argue in depth. Let me head it off a bit - taste is absolutely a thing, but consistency across releases is evidently a good thing unless you are constantly in a situation where your existing userbase is dwarfed by the prospective one. There’s a reason we hang on to diskette icons, don’t change from desktop to metro to ios style desktops every release and so on.

Last point: OSS software can be misguided, but it never seeks to bamboozle the user to make some corporate goal succeed. It won’t nefariously stuff files on a cloud or actively make it hard to export data or be incentivized to NOT interoperate with devices or programs shipped by competitors.

  • People say that and I used to agree, but honestly I use neither MS Office nor LibreOffice heavily anymore. Nowadays I can figure out what to do in MS Office very quickly but LibreOffice's menus are a confusing maze (even when switched into the more ribbon-like mode).

> So far the only exception that I've seen of a FOSS piece of software with great UX is Blender.

Most OSS software favors power users instead of total beginners. And even Blender is not a great exercise in discoverability. You need training to use it.

The stock software in Kobo and kindle can be used within minutes because they don't have a lot of features/options. And it's totally fine if that's what you need. Just like Windows 7's Notepad. But if you need more capabilities, your wish is notepad++/vim level (which does require more learning), which is what koreader is.

  • > Most OSS software favors power users instead of total beginners. And even Blender is not a great exercise in discoverability. You need training to use it.

    I think it’s more subtle than that. I am a power user — I was compiling my Linux kernel in the 90s, I wrote my own SSH tunnel manager 20 years ago, I always fiddle with settings, etc. etc.

    However, I am constantly frustrated with a lot of OSS software because it expects you to think like the developer of the tool, instead of having empathy for the user. What feels obvious to the developer of the tool (I am making up a hopefully-representative hypothetical here, but finding the change of font under “Theme” because it’s how “it looks” but finding the font size under “View” because, well, you view it) feels frustrating and nonsensical to a normal human who relates both the font and its size the same way in their mind. Again, generally — I’m making generalizations here and not calling out any specific software, but trying to illustrate the feeling that I get.

    When I switched to the Mac in 2002, it is because these sorts of things made sense. Yes, there were and are frustrations because of this sort of opinionated design mentality, but I judged the benefits to outweigh the positives. Unfortunately, Apple squandered this culture and skill and nowadays I often find the platform as frustrating as OSS ones, to the point where I have switched back to 50/50 time split between platforms.

  • I find this is mostly just a "mindset" difference.

    OSS software aligns mentally with "tool". It only exists because someone had a problem to solve, and they wanted a tool to solve it. Generally, people should expect to have to learn how to use tools. Then they solve their problem and move on. It takes effort, and the result is mastery which gives you new capabilities.

    Consumer software (and especially anything ad supported or social media adjacent, and sadly in many respects, lots of modern SaaS companies) want to be as glitzy and smooth as possible, but they don't actually want you to solve your problem. Instead they'd really prefer that you simply spend as much time looking at their software as possible. So the experience can be quite pleasant, it feels nice, things feel obvious, you get the dopamine hits quick and fast. But after a while you realize you aren't actually all that productive (or even productive at all) when engaging with them. Because the goal of this software isn't to help you solve your problem. The goal is to keep your attention and extract your money. If it happens to solve your problem... sometimes that's fine, sometimes that's actually counter to the primary mission of extracting your time and money.

  • > The stock software in Kobo and kindle can be used within minutes because they don't have a lot of features/options.

    Afaik koreader itself does not support stylus while the default reader in kobo does (somebody made a plugin lately but not sure how stable it is). Not to say that koreader is not amazing, but it does not mean that FOSS always has strictly more features. If koreader had strictly more features and I wanted some of them I would choose that. But I also right now want to use the ereader to read books and take notes, not tweak plugins and configurations. I have been in this "power user" loop where I spend more time with configuring rather than doing the thing, and I am a bit more selective with when this makes sense to me.

  • I love the discoverability of blender in the sense that I can hit space and start typing the name of the thing I want to do and run it from there while it shows me in what menu I could find it and what shortcut to use. And next space it remembers the last thing I ran so to repeat it its even faster.

    This requires however to remember the name of the command (or parts of it). I often remember that but not where to find it or the shortcut.

    It also requires that you chose the spacebar search option, which is the only correct option if you ask me. (Not a pro. Just messing around with blender sometimes.)

  • you will spend much MUCH more time as a power law user than as a new user, such that almost any % gain in productivity once familiar is worth almost any months of unproductivity when new.

> I think the issue of bad UX and OSS are comorbid because

I think UX is simply harder when you're not leading the user by the nose to your most profitable features, and removing or obscuring the others. Commercial UX is more akin to advertising than UX, so there's no wonder that it makes very clear the few affordances that it offers (and bizarre when it doesn't, like the "flat" period.)

The other problem with FOSS UX is that millions of dollars haven't been put into it (literally: 5 developers = 1 million dollars/year.) The wonderful benefit of FOSS UX is that you have plenty of time to learn it, because every feature is already there, no one is forcing you to upgrade, and no one is moving things around and deleting things because it is more profitable to them.

Commercial UX had better be good enough to learn basic functionality quickly, because it's going to be completely recycled every year. It also doesn't want you to learn anything too deeply, because then you might become a (gasp) "power user" and make unprofitable demands.

Imagine using the same type of software (e.g. an mp3 player) for a couple dozen years and thinking that it's a hassle to figure out how to learn it. Imagine being boiled like a frog to relearn how your mp3 player works every couple years for a couple dozen years, and dreaming you've saved some sort of time or effort. Imagine you've used mp3 players for a couple dozen years, you still don't know what an mp3 is, you don't know what a file is, and you don't know how to get your mp3s back after switching your mp3 player without rebuying them. Eventually you switch to streaming, give up making any decisions, and take whatever the marketers prefer.

This is "good" UX.