67
votes
Why does it seem that FOSS users don't value user-friendliness very much?
The vast majority of free and open source software available is well known for being clunky, having very unintuitive UI/UX and being very inaccessible to non-nerds.
We can see this in Linux distros, tools, programs and even fediverse sites.
I understand that a lot of it is because "it's free", but I also feel like a lot of people who make and use FOSS don't actually value user-friendliness at all. I feel like some of it is in order to gatekeep the less tech savvy out, and some of it is "it's good enough for me".
What are the best theories for why this is the case?
EDIT: A lot of replies I've been getting are focusing on the developers. I'm asking more why the users seem okay with it, rather than why the developers make it that way.
I just assume they're probably more tech oriented people. They're more focused on setting things up and making things work.
Understanding and designing UI/UX requires a different skill set and thought process. It needs you to understand design principles and human behavior to a certain extent.
They're two very different departments.
Edit:
In response to the edit, I think that the people who develop also tend to be the ones using FOSS. The ones that don't develop would tend to be more technically minded and be inclined tend to think the same way when it comes what they need or want out of the UI/UX. They just don't think there's anything wrong with it.
At work, I've had to edit some of the small documents the programmers wrote for casual people to read because it's sometimes not non-tech human friendly. Not because of gate keeping or anything, they just genuinely don't see why it's not that intuitive.
I feel that we need to appeal to UI/UX designers to help contribute to FOSS and respect their contributions.
Sure, sometimes UI/UX designers make things "too clean" and the program loses it's "charm". That's something that should be avoided. But I think that, if a respectful designer with an eye for flaws and illogical design is on the team to help identify problems of improvement, it will be a great benefit to any FOSS program.
I wonder if a fund should be set up that would employ and send UI/UX designers on attachments to FOSS projects. Would that be practical at all? I wonder how plausible it is to convince designers to do volunteer work - thus the fund idea. And a fund could promote the idea of designers contributing to FOSS and how they can do it in a way that is constructive and helpful to the project.
Why would designers get paid from a fund, as programmers work for free?
My concern is that there are fewer designers who would volunteer for free than developers. I thought of a fund because I think it would motivate, or at least motivate by employing, more designers to contribute to FOSS, since that I feel that there might be a shortage of free design talent.
I'm mainly thinking of Musescore's employment of Tantacrul here. It'd be nice to see more projects get help like that.
I would be happy to contribute to FOSS for free. I've been looking for years but they're so rare and difficult to contribute to (specifically in terms of UX, usability, UI, etc.).
If you know of any, please do share.
Yup, that's precisely the issue. Even if volunteer designers wanted to help, how can they help? It's so daunting for them, when the contribution system seems so biased to programmers, haha.
I feel like an organisation which guided designers to the right places and got them started contributing would be useful.
I also want to add that the dynamics and structures of true power are different in FOSS projects.
In FOSS, devs and repo owners have all the power.
In companies, designers are given power by management. In design-forward companies like Apple, designers are given immense power because someone with power like Steve Jobs decided that design is important and bestowed their power accordingly.
Software engineering is quite empirical. While there's some science to it, design is much more a matter of taste. I think in good orgs, there's fundamental alignment in taste through careful recruitment. But ultimately, if differences of taste cannot be resolved through conversation, then they are resolved through power.
So, an FOSS project is basically an anti-Apple environment where the powerful people (the anti-Steve Jobses) much more often than not have tastes that are opposite to the tastes of designers — who, if good, are proxies for the tastes of the user market at large, and it's not rewarding or worthwhile to spend the time and energy fighting a vertically uphill battle to achieve so little.
One thing you should be aware of is that a lot of modern UI/UX design takes into account more diverse groups of people. Perhaps they don't read well; let's make the icons clear to understand. Perhaps they have poor vision, let's make the buttons and text larger, and make sure the colors contrast enough. Perhaps they don't understand where to click; let's eliminate the extra buttons and streamline the experience.
A lot of FOSS developers are most likely college educated, middle to upper class, and have at least a working knowledge of English. While the majority of the world population maybe illiterate in English, they may have limited experience with computers, and need to have easy-to-navigate prompts.
I suggest taking the Google UX course on Coursera. It really opened my eyes to how I view modern UX design choices.
Just to confirm - is this the course that you're referring to? It's a six month commitment so I want to be sure it's the right one.
Can anyone see what the price is? Coursera says
but doesn't list the actual price.
Edit: It looks like they charge monthly and the actual cost depends on where you live.
Yes, that is the one!
My company paid for my yearly subscription. ~$400/year I believe? Covers all of the courses they have on the site.
While I'm all for providing accessibility options, so that everyone can use and enjoy a given product, I don't think those designs should be the only option. The Google product I use the most, Android, has not seen a design change in the last decade (since 4.4) that has been for the better. All the large fonts, large icons and streamlining makes experienced users frustrated, since the information accessible on one screen is greatly reduced, it takes more interactions to achieve certain actions, and is generall speaking less efficient. I didn't think I'd list Reddit as a positive example anywhere, but old.reddit.com is an examples of the kind of design I think is ideal, while the new design is closer to the principles you described - having both of them available is the ideal solution.
Designing a functional GUI is so difficult. Even many of the most popular apps around have pretty shitty GUI. Hell look at the main Reddit app that is all the topic of conversation lately. That GUI is horrendous.
I'm sure it boils down to most of the people who operate in the FOSS, realm, I assume, are programmers rather than UX designers.
UX driven product design also is often counter to how FOSS projects work. Most FOSS projects work from a kanban board and/or issue tracker. That's not exactly how you can design UX. For good UX, you need more or less a dictatorial approach to the UI/UX design, and from there create the restrictions that will shape backend development.
Most FOSS projects are also just very small, with only few active developers. This makes making large changes hard and time consuming.
Which is ironic, as it’s — in my experience — much more challenging and you would expect the average programmer to enjoy a good challenge. I consider backend development to be “solved” to some extent. There is a definite “right” way to structure your backend for most scenarios. I find that aspect of the job to be very boring most of the time.
Even "good" UX is frequently bad.
On my iPad there is a "Home" icon that has a picture of a house which looks like a "home page" of some kind but it has zero value for me because it isn't compatible with my smart home stuff. (For which I have to page through multiple meaningless icons for various vendors like Hue, Sengled, all of which seem to have the "success metric" that somebody gets a bonus every time I am forced to log in again.)
"Photos" is a multicolored flower that looks like nothing next to a camera icon which is conventional but somehow I struggle to find. The App store logo is "just another nondescript triangle" for me, the Settings look like a poster for The Golden Compass, Shortcuts look like a painting by a Russian Futurist from 1917. Supposedly getting an icon on mobile is a prize that makes your app easy to find... But I find mobile icons all look the same and it is always stressful to figure out which "H" controls my home theater, to look for the Remote Desktop icon whose edges blend into my background, etc.
It's a sock drawer for apps that makes them hard to find, contrasted to Safari which makes it easy. I type "1" and it fills it the URL for my RSS reader, "ti" gets me to Tildes. Yet everybody talks about how important it is to get an icon on somebody's mobile device robotically like the brainwashed people in The Manchurian Candidate.
The point is that the best people do a pretty bad job, on Linux we are still at the point where people argue if they are supposed to use a font size that fits inside the box a label is in, or if scrollbars are supposed to be wide enough that you can really click on them, or if a menu is supposed to be correctly aligned when you click on it, etc. It's just exhausting so I try to get by without a GUI and use the easy-to-use command line where I can tell you to
and you can just cut-and-paste it. If I tried to get you to do it with the GUI it would take a few book pages to explain it and it's not unlikely at all that the GUI will freeze up entirely on your or otherwise not work.
It's very hard to remove knowledge from your brain. Basically everything is intuitive to the developer - the mess of shell scripts for random things that lives in my home directly is intuitive to me. Ditto for any particularly heavy user of a program.
Commercial applications do extensive user-interface research with focus groups, to actually find new people to subject your interface to. That's just not something FOSS folk have the resource, time, or training for. And everyone who's already using the application already has brain worms and can't tell what's intuitive or not to a new user.
This is a really good point. As a teacher, this is probably the most valuable skill one can learn.
Also known as the curse of knowledge.
The strange consequence of this is that experts aren't necessarily the best teachers. In fact, those who have just learned something are often the better at teaching it than an expert (though of course, not as good as an expert who is also a good teacher). It's one of those things that's obvious in retrospect, but counterintuitive and easy to forget in the moment.
Yup. My wife used to ask me why I didn't want to go get my PhD. I told her that being an academic and being a teacher are two different skillsets and I much prefer being a teacher.
As a teacher - I moved from teaching physics, which I have briefly studied, to programming, which is my passion and with which I have two decades of experience - it is probably the thing I struggle the most with.
I teach Literature (which as I see it is essentially the teaching of logic and argumentation) and work with a cohort in two year cycles. At the end of every cycle the most difficult thing is coming in to a new cohort and realigning my expectations.
I'm actually really not as good as some of my colleagues in terms of actually teaching essay writing skills, despite having greater content knowledge, because as a student it hadn't been something I had consciously structured.
In my experience two very important things to remember about designing user interfaces are:
"There's no such thing as intuitive, there's only what you're used to"
"There's no such thing as user error, only bad design"
You don't need UX testing labs, it's no harder than any other development task really - but it is another skill to learn (or find someone to do).
Funnily enough, the often disliked (by hardcore linuxers) gnome supposedly did UI research.
(I do like GNOME and settled on it currently and I think I am quite the poweruser, having gone through probably every conceivable window manager on linux)
As a hardcore Linux user I had to be dragged kicking and screaming to Gnome 3 because hardcore users tend to customize their environment until everything works just so and even the quirks and warts become so familiar that any change directly threatens us on an emotional level.
I've always been someone who's very curious and yet I only gave Gnome a fair chance after being forced to use it because it was the only desktop I had easy access to that could handle a high DPI display without squinting and pain. Now, in hindsight, I wonder why I was so glued to the old desktop metaphor.
I think this can be a trap. Just like people are locked into MS Office because they can't afford to migrate away I've seen developers who are so married to their setup that when they want to try something new and it doesn't have emacs integration for example they just don't it, or do it in the maximum painful way. I've come to value to be a bit lighter on my feet now and having an easier time adapting to things instead of hunkering down and spending hours tinkering with settings until everything is just like I want it (which it never is anyway).
I can share your experience, and I can often find myself become the “old man yells at cloud” stereotype.
The fact is, humans naturally are inclined to like the well-known - changes without good reasons are a big no-no. Being a designer responsible for a re-design is probably a very thankless job, because no matter how objectively better it might be, old users will be upset.
Relevant xkcd: https://xkcd.com/1172/
Designing a good UI/UX is actually not easy. Companies spend huge amounts of money designing, and testing, things and changes.
A lot of open source software doesn't have the giant teams and resources to do these kind of things. Nor can they move as quickly as large companies. Doing a complete redesign of a produce, might mean effectively abandoning the existing product already in use. They don't have a separate team to both maintain the old and develop new things.
And then, yes. Some FOSS might be developed with power users in mind. Meaning that tons of options are exposed, and designing an effective UI is even harder.
I used to work in the biological sciences, and there is a funny phenomenon associated with scientific articles. When most people read one and have a hard time understanding it, they don't think, "This could be better written." They think, "This is a really complicated subject."
Another example (please don't hate me!) is vanilla Dwarf Fortress. The GUI is unintuitive and hard to learn, and a lot of people mistake this for the gameplay itself being dizzyingly complicated and challenging. Once you get used to the GUI, however, you discover that the game is really not nearly as hard at it initially seemed, barring bad luck. The simulation itself is ridiculously rich, but building a self-sustaining settlement that never has anything seriously bad happen to it is quite doable.
I suspect something similar may be happening here. If a piece of software is hard to use for UX-related reasons, people may not be complaining because they think it's a failure on their part, not a failure on the part of the developer's.
I remember reading a study that gave some file management tasks to participants and compared productivity between GUI and terminal users. They didn't measure a relevant difference, but terminal users felt more productive because using the terminal has a higher mental load, meaning time passes faster.
I think this is the answer. Stated differently, the current userset doesn’t care so much about ui/ux because they are the users who can operate it.
The pool of potential users is much larger, but they remain largely only interested, and not users, because they do care a lot about ui/ux.
Edit: incidentally, i fall into this latter category. I like the spirit of foss, but find the ux very unpleasant. I could figure it out if I wanted to, but i like looking at pretty things. Even popular commercial endeavors like discord really put me off. The interface is so noisy and chaotic, frenetic. I get a significant number of folks appreciate that, but for me it gets in the way of communication.
For me, the question is: Why are so few UX specialists and graphic designers engage in FOSS?
It then becomes: How can we get them to participate?
I think it honestly boils down to the culture. Most FOSS projects are developer lead, meaning projects are run via issue trackers and kanban boards. These are tools way more suited to programming tasks that can be broken down to individual small chunks, than UI/UX re-designs which fundamentally don't fit into small incremental updates.
I don't think this is the problem. You can make incremental changes to interfaces, it doesn't all need to be huge overhauls. It's better if not, most of the time. You can run UX projects on a kanban board just fine. UX should be as fundamental to the development process as programming is - interface designers are developers. The programmers can write the best code in the world but if nobody can use it, it's worthless. Equally you can make the slickest interfaces in the world but if the code sucks, it's worthless.
I think one part of the issue is just that most UX designers aren't really aware there is a FOSS scene at all (because let's face it, most of them are using Macs and aren't highly techie people) that they could volunteer their time in, and most software engineers don't really think UX matters.
The latter is demonstrated by the fairly common attitude of "we shouldn't make things too easy" and "people can work it out if they're not stupid" neither of which is conducive to straightforward user interface design. The frustrating thing about that is the "power users" have got it backwards - the smarter users (aka them) can easily figure out to go into a menu, use shortcut keys, scripting or the console to get the advanced things done, while the everyday users benefit from the top level being streamlined and easy.
There's also a strong resistance to change in a segment of the community - look at the proliferation of Gnome 2 forks and clones which came about when 3 was released, and the rage. Yet Gnome Shell is one of the best user interfaces I've ever used on a computer. Ironically, the people complaining Shell is too simple just haven't put the time in learning how to use it properly..
Fewer UX specialists are engaged because interface design has far less "scratch your own itch" value, because by definition the interface doesn't expose new functionality and the designer of an intuitive interface already knows how the interface works.
What's more, "do-ocracy" works against UI designers as their interface doesn't do anything on its own.
By the way, we already know how to get them to participate: pay them. I know this breaks from the FOSS anarchist ideal, but frankly capitalism isn't disappearing anytime soon, and in the mean time starving artists are more than just a cliche.
Getting into this deeper would be very off-topic, but I will say that anarchist =/= absence of governance, at least not in this context. It's an ideology that is skeptical of centralized authority and seeks to abolish unnecessary hierarchy, but that doesn't mean a lack of governance or organization necessarily. One can definitely question how anarchist FOSS actually is in practice, but the ideology at the center of FOSS (especially at its most idealistic) definitely contains a lot of anarchist principles.
To this I have to repeat myself:
Why pay designers and not the programmers?
We should also probably pay the programmers, tbh.
Programmers are good at using complex tools and so have difficulty making not complex tools, and there is no strong financial incentive to make said tools accessible to a broader audience.
As for the users: I don't think users are okay with it! But contributing to a new codebase is a monumental task in and of itself, much less rewriting a GUI and well enough that other users used to the existing interface (incl. the tool creator) think it's better.
IMO its because of a few reasons
good UI isn't a goal from the start in most projects
good UI is HARD
Speaking to why FOSS users tend to accept it...
they tend to be nerdier
sometimes when you don't have money you just have to accept it
sometimes good UI comes at the cost of features or the ability to hack stuff together
I'd also like to add that in the FOSS space there has actually been a big push asking for major projects to overhaul their UI. GIMP in particular is always used as an example.
Edit: also, often if projects become too mainstream they then add telemetry or weird stuff that really pisses off the community.
I was waiting for GIMP to be mentioned, which is the poster child of “FOSS is for inhuman robots who hate users” design … and likely responsible for turning off countless users to Linux in general.
If “GIMP” is what free software is, who wouldn’t happily pay money to never use it again. It’s sort of spectacular in its awfulness.
I've used GIMP since the late 90s. It's come on leaps and bounds and I'm happy with how it works.
You can certainly feel it's age though, and lack of UI enthusiasm. The issue is, there are no developers who devote their day job to fixing and/or improving it. It costs you nothing and they get nothing out of it except saying "I helped make this."
You're made of sterner stuff than I; used it once and never tried it again. It was like being punched in my eyes over and over and having my hands encased in jello trying to do the simplest things. Photoshop, it is not.
I love me some Linux, but for visual creatives it's still way behind. I've never found anything remotely comparable to Photoshop / Lightroom. The interfaces are horrible. Every one. It's maddening.
Not FOSS but Affinity Photo, Designer and Publisher to replace Photoshop, Illustrator, and InDesign respectively. It doesn't work on Linux (someone suggested it works via wine?), and they aren't as full featured as Adobe software is... but THAT is how you do UI/UX better than Adobe.
one time purchase, no subscription
updates for years (1.x saw updates and improvements for 8 years before you had to buy 2.0 for new stuff)
iPad apps for Photo and Designer that aren't stripped out husks (looking at you again Adobe)
No more Adobe software always running in the background slowing things down
They just need to do a Lightroom equivalent but I don't believe it's on the roadmap.
Agree that Vanilla GIMP is somewhat bad but it has customisable icons and UI. There was a Photoshop overhaul add-on available for those similar with Photoshop. Using an icon pack like Breeze Dark or its Gnome counterpart also makes a big difference. Once you get familiar with the icons it gets better.
I'm not a developer just regular users of FOSS and using GIMP and Paint.Net for years. A while back GIMP used to have three seperate main windows only and no option to change that behaviour. It's free so I just assumed dev knew how to make three windows work together but not how to join them.
Edit. Sorry I was replying to @bd_rom
I'm surprised by this, I'm a regular GIMP user and while its interface does seem a touch dated it has always seemed logical enough to me and if you're at all familiar with Photoshop it's very obvious that it's attempting to clone (older) versions of PS in how it is laid out.
I do agree it needs work though, I have spectacularly broken my GIMP panel layout before with an errant click/drag (crappy touchpads) and despite being a long time user was unable to fix it. That sort of thing needs work.
I think in general it just needs cleaning up and refining, it's not a lost cause though. Some inspiration might come from looking at the enormous leap the people behind blender made for 2.8, while I still use pre-2.8 version of blender in addition to much newer ones for some reasons I am always amazed at just how big a jump they made with the new versions and for such a powerful piece of software they have done amazing things as far as usability with 2.8+.
I’m glad you mentioned blender, because with how good it is there’s no excuse why after 25 years of development GIMP is still so bad.
I guess we'll agree to disagree on GIMP, it's not fancy but I would never call it bad.
Blender was one of those programs that people used to joke about because of its interface though, it's still a massively complicated program by its very nature but they've definitely put some great work into making it make a lot more sense.
In my experience the main issue is when you combine a bunch of different pieces of FOSS software together, each developed by different people with different ideas. This is especially obvious with something like Linux distros, which are made up of hundreds of different components developed by separate teams. Unfortunately the current software paradigm isn't particularly well-suited for this style of development (maybe some sort of ubiquitous containerization or advanced hypermedia could make this better).
Although I'll say I don't think I agree much with the main point. Commercial software doesn't really value user-friendliness, it values user-give-us-moneyness, which is particularly obvious with the whole Reddit debacle going on right now, but results in dark patterns widespread commercial software.
I think poor but usable UI is the default, whether it's open source or commercial software. There are plenty of clunky web apps out there from big companies. If it's not literally unusable then someone who has learned its quirks can use it to get the job done. People who really need that functionality will use it unless it's so confusing that they can't figure out how to do it.
By contrast, consumer apps that are inessential and have lots of competition often need to be easier to use or their customers will give up, give it a low rating, and choose a competitor. But that still doesn't mean everyone's able to do make a great UI; there are many more apps that fail than succeed.
Those are vague guesses. I think you'd learn more by picking specific examples and researching how they ended up that way?
This is where a lot of FOSS comes from: folks scratching their itch. If it does what they need and they decide to share that’s more than we should be asking of them.
Folks could then contribute fixes or polish but even that can be a lot of extra work for someone who just wanted to share their itch-scratcher.
I’ve only done super small contributions or tools over the years and even the PRs or emails for supporting those things was more than I wanted to deal with. (If you ever write a script to uninstall something and share it with the maintainers of a larger package don’t put your personal email in the script or you’re on the hook for support for years!)
I hope you're talking about the 3rd party apps here because the OS itself such as Ubuntu, Linux Mint, etc. seem very polished and user friendly.
A few popular apps like LibreOffice, Firefox, etc. are also very polished as they are built and tested for all platforms.
Regarding the open source third party apps, you can try finding alternatives with better UX, if you can't find one, consider hiring a freelancer on Upwork or Fiverr who can fork its source code and improve the UX to your liking. You'll be hitting two birds with one stone: Contributing to FOSS and help someone with employment.
LibreOffice's UI is pretty mediocre and especially lags behind MS Office.
KERNING
Nobody takes it seriously when it gets brought up.
I get that its a tough problem and it doesn't help that good kerning is HARD. IMO the fix needs to start with acknowledging that there's a problem.
Issue with writer itself: "pair kerning" makes things worse
UI issue: the pair kerning setting is hard to find and should be disabled by default not ON
OS specific: for me in MacOS everything thats suggested to fix it doesn't work. I'm willing to admit that this may have just been me doing it wrong... back to the first point.
Upstream issues: Not sure what packages could all cause that issue but fonts maybe fit this category (?) and can be an issue. Maaaybe just work on making the default font work at a bare minimum.
Dammit! Gets me every time.
Hasn’t almost every OS settled on some Harfbuzz nowadays?
I never said UX is easy, I'm just saying that LibreOffice's UX is mediocre. I don't blame them, I know exactly why MS Office's UI is amazing: they spent hundreds of millions of dollars focus testing and polishing the crap out of it.
I don't blame LibreOffice for lacking that money; I blame the FOSS community for pretending this is something that we can be competitive on with a handful of volunteers and no budget.
The #1 thing holding back FOSS is its lack of emphasis on funding FOSS projects and paying developers. There's a damn good reason why Firefox has such a well-polished UI, and we all know it.
I don't think the premise is valid. Some open source tools have a bad UX but from my experience there's often care put into UX. Some examples like Element Chat, Home Assistant, Emulators, Moonlight streaming, playnite, and a bunch of other tools I use everyday.
If I had to guess, the reason this comes up is because FOSS people often recommend people move from market leading and highly polished tools to open source tools that are much smaller and still in development. That UX isn't going to be as polished, but that doesn't mean the team doesn't care about it.
I don't think it's that they don't care, but rather that they care about other things more. They care about something being free and open source more than good usability. They're willing to accept a level of user friendliness that might be less than others because the software is open source.
Maybe the users of FOSS software are from the same demographic as the developers; maybe they're other developers, with similar opinions about the aesthetics and features of software. Developers developing software for other developers aren't going to get the same pushback as if they tried to release their software to the general population.
And, of course, the userbase of FOSS software is going to be self-selecting to be in favour of its aesthetic. People who like what they see will sign up, while people who don't like it will quietly move on to another alternative.
Good UI and UX is difficult. And people rarely work on such things free.
It's not like we maliciously make things difficult. It's just a skillset that many programmers lack.
I believe that part of it has to do with how many FOSS users have associated the terms "user friendliness" and "good UX" with "dumbing down" or oversimplification and don't want the software they enjoy to suffer from that.
This association isn't without grounding in reality, but oversimplification does not actually go hand-in-hand with good UX. In fact the software mostly highly lauded for good UX excels in progressive disclosure, which provides a UI that puts the basics first and surfaces power user features and complexity as users come to need those things, rather than dumping everything on them at once and overwhelming them.
Yeah I don't really understand it either. As someone who's really tech-literate, I love reading comments/posts about non-techies sharing their experiences on the world to provide me a different perspective on things.
I don't believe this is supportable by evidence. In fact, as I've written previously:
I just wanted to say that GIMP is my favourite example of software where devs never cared about UI or UX. I am 100% sure they don't even use their own tool because if they would it wouldn't look and work like it does.
One thing I think about a lot when I elect to use FOSS is that designing and developing software (especially GUIs) is incredibly hard. Having developed some basic applications before for my own personal use or for use by people I personally know, it can be a massive struggle to convey a simple and pleasant user experience without any formal training or experience. I think that understanding manifests in the form of a lot of patience for software that I know respects me as a user.
Additionally, the audience for these kinds of software generally have a lot of experience with computers and design languages in general, and so it can be much easier for us to intuit the intention behind specific design decisions and use a FOSS product the way it was intended by the developer.
I'm a non-techie and I don't find FOSS software unfriendly at all. iOS is probably considered the most user friendly operating system, yet the software is super boxed in, limited, and usually subscription based. Also, the hardware is expensive. FOSS software is most always user-respecting, private, and customizable. I use LibreOffice and VLC every day.
I'd have no problem recommending a machine with Ubuntu or Linux Mint to grandma. Maybe she'd buy one with the OS preinstalled, but it's not like she's installed MacOS or Windows from scratch either.
Survivorship bias. The users who aren't ok with it, don't use it. The only projects that stick around are the ones that have the right attractiveness to a niche to support them. Usually that's a techy niche that isn't using mainstream projects for whatever reason.
Because many of the projects are niche and niche things always turn into a 'club - members only' mentality.
That's why it doesn't improve, anyway. Why they start out difficult to use is because most of them are worked by pure programmers with little experience or interested in UX design. Once the users work out the software they become invested in their successful implementation and believe everyone should go through the trouble, gatekeeping as you say.