Users: I demand OSS devs and Maintainers do X
OSS Devs/Maintainers: Are you willing to contribute code or at least donate any money?
Users: Uhh its OSS, you should just do all the work for free with no funding. Also I demand that your software be as polished and complete as (premium proprietary software) I demand you do X, I demand you do Y, because im entitled to free software.
I am sympathetic but also so damn tired of seeing what essentially translates to:
“Look, [megacorpo] bought out my school’s ecosystem so that’s all I learned. It’s “industry standard”, I can’t believe this FOSS can’t even do this one niche corporate-job feature, therefore it’s objectively terrible / not ready / inferior / useless for job work.”
Which can usually be further boiled down to:
“I tried it but it wasn’t a carbon copy of my preferred corpo-ware without any strings attached so it basically sucks.”
Counterpoint: Blender was the first 3d modeling tool I tried and I bounced off that UX so hard that I haven’t touched it in nearly 20 years. Sometimes a bad UX is just bad UX.
Sometimes a bad UX is just bad UX.
Totally can be! Absolutely!
Although Blender’s amazingly usable now and has had lots of love in that regard! But it took a LOT of support to get this far.
Good UX is crazy important.
I think I’m more irritated at the people who seem to show up in so many FOSS discussions, expect FOSS alternatives to compete 1:1 with their billion-dollar corpo-ware of choice, demand the world of it, offer zero support, and then declare “it sucks and isn’t ready for the real world” because it’s not so perfect that Autodesk and Adobe are like “Well we’ve had a good run, guys.” and give up lol.
I sympathize because I know where the frustration comes from. They’re sick of their tools being held hostage by interests that constantly seek to screw them! But change requires flexibility, cooperation, and support.
I think a lot of people just don’t want to say “I want Maya/Photoshop/Excel/Solidworks/Windows/etc…but free and without dark-patterns!” (Don’t we all lol) Because they know that sounds unreasonable (yarr aside lol) , but people tend to get settled and comfortable with whatever got to them first.
But taking that out on the community isn’t helping anybody.
Constructive criticism of UI/UX is absolutely essential though, and requires a lot more understanding of how humans interact with things than simply “Well, billion-dollar-ware has always done it this way.” Haha
Absolutely this, you cannot expect Blender level UI/UX without blender level funding. The fundamental problem is that new users/inexperienced users/nontechnical users arent used to contributing bug reports or even proper constructive criticism.
Furthermore what people forget is that being a 1:1 carbon copy of a corporate software isnt inherently a good thing. For example Linux, I love Linux and I love the way it works. I use it not because its OSS but because I genuenly prefer it above Windows, I dont want Linux to be like Windows. I love tiling, I love Sway, I love Hyprland, and despite being in alpha I love the Cosmic Desktop. I dont care that tiling isnt immediately intuitive to Windows users, I absolutely love it.
While im at it I absolutely despise the idea that the Terminal is inherently not user friendly (especially with a shell like fish). The idea that just because somebody isnt used to something makes it bad. Or that having to use a wiki/look at the docs means its “not ready”. All software is new to somebody at some point, that doesn’t make it bad.
OSS: Why do so few people use our softwaree!? It’s literally free!
Engineers don’t let engineers design interfaces.
The open source music notation software MuseScore used to be really, really bad. A musician and UX designer gave it a scathing review in a humorous YouTube video. And then the company behind MuseScore hired that YouTuber and spent a lot of effort doing a major redesign, and now it’s actually quite good.
All it takes is for the people in charge of the project to put aside their hubris and trust that sometimes, programmers aren’t the best designers, and to get people who are trained in designing and evaluating user interfaces to do the job. And to perform adequate user testing.
I was a happy MuseScore user before and after the UI changes. So this post brings to mind questions that usually float in my mind:
- When I can happily use a thing whose UX is criticized: is it just because I don’t know any better alternatives, or because I’ve spent so long with it that of course I know how to work it? Or is the UX really not that bad? Or is it that there are often general solutions for most of the population, but sometimes some people take really well to things that work poorly for others and vice versa? Is it that the hated parts are bits I do not touch much in my workflow, so of course I see no problems because I am not interacting with the problem parts?
- When I have difficulty using a thing whose UX is praised or has no criticism: is it because I am smoothbrained? That I just have not had enough time trying to figure it out, so of course I struggle and just need to apply myself more? Is it something that works for most, but it will not work for everyone? Am I in a really niche use case with bad UX that nobody else has bothered to complain about?
I do not have enough UX knowledge to criticize or make objective evaluations here. I only have how easy it is for me to navigate applications. Though I would like to work on gaining it someday, especially so I can help out FOSS targets of “bad UX” complaints.
I never actually put any serious effort into using MuseScore myself before the changes, so I can’t comment from extensive personal experience.
But as a musician, I did use scores written by someone in MuseScore, as well as ones written in Sibelius. And I could always tell when it was MuseScore. I’m sure it was possible to write good looking scores in MuseScore 2, but it clearly did not make it easy. The scores were obviously inferior in terms of layout and design compared to those produced in Sibelius. Basic things like spaces between notes not being the right proportion, or dynamic markings appearing as plain italic text instead of the usual bold dynamics would be wrong in MuseScore far more often than in Sibelius.
As a general rule, a good UX should:
- Make it very, very easy to do (or discover how to do) the most common basic things, and should result in them being done in the way a user expects
- Not slow down a power user from accomplishing basic tasks at speed
- Allow easy discovery of and access to less common tasks
A lot of designed-by-software-engineer FOSS applications do a good job of 2 and an ok job of 3, but fail at 1.
I was prepared to hear the story that the youtuber got sued into oblivion for defamation. Glad to hear they actually worked on improving it instead
Love tentacrul. I re-watch that video from time to time just because it’s so good. It was also really funny watching a later video of his where he just casually dropped that he was working on musescore.
I want him to do FreeCAD.
Omg yes. I actually tried using that after my Fusion360 student license expired and the amount of time it took me just to extrude a basic shape was insane.
You want to make a midplane, a reference plane midway between to parallel planes. You click the Datum Plane button, and then the two parallel planes, and you get a plane that is perpendicular to both, not parallel halftway between. I found it easier to forget how to need midplanes than to get FreeCAD to make one.
There’s some cool concepts in there, I make heavy use of the spreadsheet function, but I swear to every god in every pantheon that Autodesk is paying the FreeCAD development community to keep their UX at least this horrible to preserve their business model. I can’t explain this level of incompetence any other way without relying on rock chewing stupidity.
I love that FreeCAD exists and use it often, but I really hate how much you have to “fight” the UI. I even briefly considered learning OpenSCAD out of frustration with the ui and the toponaming problem (before realizing that switching to OpenSCAD would just shift my frustration onto Javascript fuckery).
Now that the latter is fixed, though, I have just forced myself to forget how to do things the “normal” CAD way (i.e. using patterns and flows that most other software has standardized around) and instead how to do them the FreeCAD way.
I hope at some point we get an overhaul of the UX, but in the meantime I’ll grin and bear it since I have yet to find an even remotely comparable F/OSS CAD software that works the same on both Linux and Windows.
That whole series is absolutely brilliant, but it’s hard to go past the Sibelius one if I’m gonna go back to one. And I say that as a long-time Sibelius user who can comfortably work much faster in it than in any of the alternatives.
The conclusion I got from the video was that, while it’s very powerful software, it’s very difficult for new users to start using the program due to unintuitive placement of options. That’s how he, for a lack of better word, reviews each piece of software, from the eyes of a new user.
But on the other hand, the video is also 7 years old at this point so maybe sibbelius has fixed some of the stuff that he pointed out. I don’t really compose music so I honestly wouldn’t know.
Unfortunately Sibelius’s development has basically stagnated since 2012 when the new corporate owners fired the entire original development team, with only one noteworthy release of the core app (not counting side-projects like an iPad app) since then, in 2014.
I first learnt Sibelius on its pre-ribbon interface, which I think was much better (even though I loved the ribbon in MS Office). That certainly made the transfer to more modern versions easier. Still, although Sibelius has a number of specific hangups in its interface that make fairly common activities awkward and unintuitive, I really do think it has the best basic flow. When you’re just in the zone inputting notes, it’s so easy to use in a way MuseScore isn’t.
I actually take some issue with Tantacrul’s design process, because it feels like he fundamentally doesn’t understand how intermediate users like myself use the app. At one point he sent out a survey asking “how many keyboard shortcuts do you use?” in Sibelius/MuseScore etc. The problem was that he didn’t define what a keyboard shortcuts is, and when people asked for his definition, he just snarkily responded that it would be obvious. But it’s not. In Sibelius, you use your left hand on letters A–G to enter the note pitch, and your right hand on the notepad to enter rhythm values and common articulations. Slur lines and some other things can be entered during this process as well (slurs with the letter S).
Does this count as keyboard shortcuts? To me, everything I described above except maybe the slurs is actually the musical equivalent of typing text into a word processor…or a browser text box, like I’m doing right now. Does it become a “keyboard shortcut” just because it can also be done by clicking a rhythm value in a toolbar, and then clicking a location in the staff to choose pitch? I have no idea if Tantacrul thinks so, because he chose snark rather than clarifying.
Incidentally, his MuseScore design replicates this flow, but without the visual reference of the keypad toolbar that lets you learn and easily see what number to press, without requiring sheer memorisation. It’s been a while since I last tried it, but I vaguely recall having other issues with the flow being hard to work out with a keyboard. Great if you’re just slowly mousing around everywhere, but not for the intermediate user trying to get in the zone.
Which is such a shame, because he did such a fantastic job of the other stuff. The user onboarding, score setup, page layout management, etc. The attention to detail even with small things like music fonts and symbol design is impeccable.
I’d say those are keyboard shortcuts.
Interesting. That would make his survey of rather limited value, in my opinion, because just by doing notes (including rests), durations (just from semiquaver to semibreve, including tie and dot), and accidentals, you get 18, right off the bat. Considering the ranges offered in the poll he made were 1–5, 5–10, 10–20, and 20+ (never mind the overlap if you happened to use exactly 5 or 10…), that makes it very hard for anyone who types their note input instead of hunting around slowly with the mouse to get into anything other than the top bucket. Especially since he quite explicitly said “including typical ones (like Ctrl+S, Ctrl+Z, etc.)”
This is a huge victory, the big takeaway for me is that the person who smack talked the software was willing to get in the room with the designers and help them out. It’s easy to complain, it’s a lot more work to complain, run through user tests, file bug reports, etc. So bravo to that person, and hopefully we can see this sort of outcome on more software.
Blender’s UI overhaul from 2.7 to 2.8 was a much needed and very welcome update as well
There’s a quote along the lines of “User error is not a thing, the system allowed for the error through bad design”
Which can be true depending on how far you stretch it. I’d say that if a chunk of the user base is having a problem, it’s a design problem
I recently had a case at work where you could move an object by holding the left mouse button and delete it with the right mouse button. If you deleted it while moving, you got an error message and the program would crash. It was an easy fix but afterwards I had a one hour discussion with our usability engineers if what I had fixed was a bug (my opinion) or a user error (theirs).
That one’s easy. Is the crash part of the program’s design?
If not: It’s an implementation bug, the program is not behaving as intended.
If yes: It’s a design bug, crashes shouldn’t be intended behavior.
Their argument was along the lines of “The requirements and design don’t specify what should happen if you move and delete at the same time so it can’t be a bug. Behavior that doesn’t violate the design but also doesn’t lead to the result the user wanted is a user error”. My argument was that we can’t always specify the interaction between arbitrary features other than “If the user does two things at once, at least one of them should be executed, ideally both” and “the program shouldn’t crash just because the user did something unexpected”. Otherwise our design document would be ten times as long.
I think that there is always an implied design requirement of the program shouldn’t crash.
You would think so, right? But that doesn’t have a requirement ID so apparently it can’t be referenced in the incident report.
Sounds like the devs are cowards. Or maybe their pay counts on it not being a bug
Software for a medical device. Everything needs to be done exactly right and documented in three different places or else the regulatory agencies from at least three countries get really angry at you and worst case pull your device from circulation. Less cowardice and more cover your ass. Still annoying though.
Yeah, that’s basically the kind of logic you use when designing a low-level programming language: If we didn’t define what happens here then anything that happens is correct behavior and it’s up to the user to avoid it.
Of course applying that logic to a GUI application intended for a comparatively nontechnical audience is utter madness.
That’s the kind of logic people historically used when designing low level programming languages. It’s not the kind of logic you should use or that people nowadays usually do use. Undefined behavior is widely seen as a Bad Thing in the programming language design community.
Oh, don’t get me wrong, I fully agree. Undefined behavior is terrible UX and a huge security risk.
Undefined behavior was kind of okay when RAM and storage were measured in kilobytes and adding checks for this stuff was noticeably expensive. That time has passed, though, and modern developers have no business thinking like that, even ones working on low-level languages.
I should’ve phrased my comment differently.
That is the type of thinking that causes a massive amount of CVEs in those languages.
Hey, the design specs never said the program shouldn’t blast out and air raid siren at full volumn every time the user clicks a button. Cannot be a bug, must be user error.
They were holding it wrong, obviously.
But IMO that’s one reason weird UX/design is not uncommon and can persist in dev ecosystems. The intended users are more proficient than average and most are able to work around most issues.
God really should get on that and fix the design issues with humans
I dunno, if we’re really designed in that dudes image then I wouldn’t trust him at all personally.
And here I am trying to convince my sales team that supporting a workflow where users run our app with
sudois a bad idea.This sadly excludes the majority of bad UX decisions that are done entirely to maximize users time inside of the app as well as display advertising.
So many functional apps are destroyed by these incentives. There is literally a “skill issue” but in the opposite direction. The design is either purposely malicious in a subtle way with “dark patterns” (something Amazon is insanely guilty of. Literally just go try and return and item.) or is purposely annoying trying to ensure the user purchases the “free trial” to actually make the app functional. Knowing a lot of users will be charged at least once for the free trial.
I guess my point is that there is so so so so much wrong with UX design today. But for the majority of people that’s not because of a bad programmer with no design knowledge. It’s on purpose in most cases.
You can’t “skill issue” yourself out from every situation
If you can’t do that - that’s a skill issue tbh.
I was about to post this and just thought I should check if someone else had got to it first.
people always mention blender when talking about good ux in open source software, but i feel like the godot game engine doesn’t get enough love. it’s miles above of unity in terms of intuitiveness for me personally. plus it’s entirely customisable since it’s built in godot itself.
Godot is something I can still be super newb at and yet straight up admire. The nodes tree / scene system is a work of genius and I love it so much.
I do feel like a lot of inspector bits suffer from unintuitive “hard to distinguish menu to sub-sub-sub-sub menu” UX, but I think the editor’s “expand all inspector headings” (or something) option is really handy for knowing what you’re working with, and mitigates that a little.
Remember when selecting something was done with the right mouse button in blender. That was great UX for beginners.
Point not made. First time around I read this as “elitist devs looking down on other devs thinking the latter can’t figure out a good UX”, which instantly gets countered with “it is a skill issue alright, but with all the people who design to UX to be exactly that (looking at you too, Android)”
Ok, so this is about devs making software on their own and producing bad UX? Wow, news flash: most devs, even good ones, are not good UI/UX designers, that’s a completely different skill
deleted by creator
I like GTK and it’s really simple to make good looking functional UI with GTK4, but apparently people have a hate boner for anything good looking, GTK or Gnome related
From what I’ve heard about it, it’s because the default gtk style only fits in with gnome, and gtk4 made it really difficult to customise it and is also really buggy on anything not gnome.
That’s what I’ve heard anyway, I’m not a distro dev and the distro I last used is still on gtk3
Nowadays “buggy” is not how I’d describe it, though there were certainly teething issues at the beginning. By now other DEs have learned to deal with it.
However it’s still true that the GTK4 design is ill-fitting, and very opinionated. Quite exemplary of this are the applications that hardcode the GTK file picker (like Firefox and chrome) even though it’s inferior in every way to the Qt file picker and forces the infuriating GTK “design” choice of doing fuzzy search when you type in the file list instead of jumping to the relevant file. Very annoying when dealing with organized directories especially when no other file browser on my system works that way!
Thank you!!!
I’m so glad someone else hates when applications hardcode the file picker, especially to the GTK one. I always have to remember it’s single click to open…
From what I’ve heard from devs who touched GTK/Gnome, that iskind of caused by the GTK devs.
GTK is the better looking girl at the sock hop. QT’s dress is a little ratty and she’s still got that lazy eye.
QT has a certain “Ah that’s good enough for now, I’ll fix it later” feel to it, while GTK makes things that look done. It’s such a shame they wasted all that potential making something as rectal puke as Gnome out of it.
Gnome is good, their extension system is ass but I love the rest of Gnome
Every Microsoft product.
But also Gimp.
smh skill issue
But it is a skill issue, just UI/UX design skill. Not software development skill.
cough Inkscape cough …
Why are y’all looking at me like that?
Inkscape isnt bad, Ive used it and preferred its ux to that of adobes even 10 years ago.
now freecad…
FreeCAD has improved a lot though with the 1.0 update.
There’s no cad somewhere on this planet with good UX.
OpenSCAD has pretty nice UX (though massively outdated UI look & feel) but of course describing your part in code is a very different use case from most other CAD tools.
Counter point: KiCAD
Yes I know it’s schematic capture and PCB layout, but I’m giving it as an example for two reasons:
-
The UX is genuinely really good and easy to use even for a novice following YouTube tutorials because it follows the norms of a schematic/PCB software package you’d expect to pay for (OrCAD, Altium, etc.)
-
It’s open source and used in industry so GIMP and Inkscape have ZERO excuses for their horrific UX which is the prime reason industry professionals don’t want to spend an age re-learning all of their workflows.
There I said it, I’ll get down off this soapbox now.
-
But freecad is nonsensical. Why do i have to change modes? Why can’t i sketch it all out first and then do body work? Why is snap to grid not on default? Why is the constraint system basically non functional?
Its good but you really need to rtfm and forget 30 years of ui and ux evolution.
Inkscape’s UX is horrendous. Looks alright, but it’s unusable garbage.
I thought it was just me! I’ve been using Inkscape for a long time now and I always feel I’m wrestling with the damn thing. I understand the principles behind vectors but I’ll be damned if I can consistently achieve what I’m attempting to accomplish.
