Slashdot Mirror
Ask Slashdot: A Point of Contention - Modern User Interfaces
Reader Artem Tashkinov writes: Here are the staples of the modern user interface (in varying degree apply to modern web/and most operating systems such as Windows 10, iOS and even Android):
- Too much white space, huge margins, too little information
- Text is indistinguishable from controls
- Text in full-CAPS
- Certain controls cannot be easily understood (like on/off states for check boxes or elements like tabs)
- Everything presented in shades of gray or using a severely and artificially limited palette
- Often awful fonts suitable only for HiDPI devices (Windows 10 modern apps are a prime example)
- Cannot be controlled by keyboard
- Very little customizability if any
How would Slashdotters explain the proliferation and existance of such unusable user interfaces and design choices? And also, do you agree?
You think it looks bad on screen, try printing it!
Its really the proliferation of the horrid iPhone UI. iOS has a horrid User Interface that is really difficult to use and everyone seems to be very quick to copy the least usable portions of it :/
It's a little sharper than that—the current generation of interface designs was a direct reaction to the previous decade's tradition of absurd skeuomorphism. The moment Steve Jobs died, Apple did an about-face and started following Microsoft's Win8/Modern/Metro UI lead. It may look like a step backward to those who from the Windows 2000 and Gnome 2 era, since there's a loss of visual cues, but the flatness of current interfaces is way better than what the classics became in the post-Windows XP era: bloated, overdesigned, pseudo-real-objects cluttered with mismatched shadows and conflicting perspective angles. You couldn't tell what was a button there, either! At least now there's a consistency and a return to the actual use of design guidelines.
That said, there are still a lot of cases where literacy in idioms dominates: for example, the largely inexplicable convention of swiping sideways on a list to reveal 'delete' or 'edit' buttons in mobile apps. That's probably where you and the UX designers run into the most difficulty. But two decades ago, every "how-to-use-a-computer" class targeted at seniors started with how to operate a mouse—so, as I think you've already recognized, it's important to try to take these things with a grain of salt, and recognize that no one is completely objective when it comes to understanding the culture of computer operation.
Bio questions? Ask me to start a Q&A journal. Computer analogies available for most topics!
Scott Meyers calls this the The Keyhole Problem and has a paper with a bunch of good examples.
My "favorite" modern example of the problem is Chrome's omnibox auto-completion, you get six results at maximum, they don't even give you a scroll bar or a "Show more" link, six results only. There used to be a command line option to increase it, but they removed it some years ago, it's now a hardcoded constant in the source code.
The problem is that there's a glut of "UX" designers convinced that if someone else has successful, and you copy the superficial hallmarks of their design, you'll be successful too. Take Facebook's "infinitely scrolling" page design for example - suddenly you have every damn app and website using an infinitely scrolling layout, even things like weather apps where the information is finite and is best presented using another paradigm such as tabs. Combine this with the prevailing attitude that if less is more, then even less must be even more, and you get the current mess we're in now.
This is not only the case with the current "flat" design epidemic ("Apple went flat and look at how successful they are! If we go flat we'll look modern and we'll be successful too!") but in many other elements that have been taken to an extreme at the cost of usability and accessibility:
- The use of razor thin fonts
- White text on monochrome, pastel backgrounds
- The loss of critical UI elements like scroll bars and button outlines, because apparently they just clutter things up
- The use of "hamburger" mystery meat menus
- Loss of status bars (which attempted to at least give some idea of percentage completion of a task) in favor of things like dots that twirl, spin, and dance in circles
Developers traditionally make efficient, functional, ugly interfaces. They did this by using standardized UI controls. They were largely constrained. Today, without those constraints, those same developers make inefficient, semi-functional, pretty interfaces. And with the focus on form over function, they are pushed in this direction by management. (Thanks Apple, for telling me that I want to get rid of all the jacks in my laptop so that it can be be 0.00001 inches thinner.)
A good UX person -- not the kind of BS "UX" that I see lambasted here -- but a real one -- can improve the look and feel of an application, optimize the workflow, and make it pretty too. I work with a UX engineer who uses statistics on the average hand size of our target demographic, and can quote the average size and resolution of the displays they are using. On touch-screen apps, our UX team optimizes for right handedness, and organized the screen so your hand doesn't cover the things you are looking at and so you make minimal movements. A few years ago we even created a mock-up, and had actual users go through a workflow and timed them, counted number of clicks, etc. This is good UX. It's human factors engineering + graphic design.
A sad anecdote: A few years ago I had the pain of designing a UI with a bunch of managers. It was a screen to add/edit/delete users who had access to an account. I drew-up a typical text box with a list, and then an add/edit/delete button at the bottom. You could fit 50 users on a typical screen, quite readably. They HATED it. Their design fit about 10 users on the screen. Big margins all around. Each row had a separate add, edit, and delete button, a large single-color icon of a person. All the icons were the same, so they communicated nothing. The text was so large that long names needed an ellipsis to fit. The add/edit/delete buttons were tiny icons without text. It was pretty, wasteful, and slow. They loved it.
On another project, which was an industrial machine, they wanted icon buttons. Their previous version used 16-color EGA graphics so it needed an update. So I used actual 3D renderings of the parts as icons. Initially everyone loved it because it was clear what the icons did. 3 years later, it laughed-at because it is too "realistic." So on the next project they replaced the realistic icons with single-color conceptual representational icons. Unless you were on the project, you had no idea what the icon meant. The customers came-up with names for the icons: the "one-eyed cat" let you search. The "disney castle" was to load a tray into the device. The "laser broom" was the barcode scanner. This interface is loved by development because it is so pretty, and is the new standard moving forward. The customers (and training department) complain that unless someone uses the device regularly, they forget if they should start the workflow by clicking the "one-eyed cat" or the "laser broom."
At with the next project, they are using text under the icons again, so users know what they are.
I will relate my own experience. I have used technology products since Decwriters and CRT terminals on big computers behind glass walls. And everything since then.
I have used numerous candy bar and flip phones. I used an Android phone. When I was handed an iPhone to do something, I was absolutely baffled at how to do certain basic operations. I would even consider this is because I could be an ignorant idiot. But I don't think that is so. I could make certain fits of progress, but then get stuck at some basic operation. (Don't remember details, it was a few years ago.)
I'm sure I could learn how to use an iPhone / iPad just fine. I look at some of the things I have had to learn. Back in the day you had to memorize a stack of computer manuals that you could not remove from the computer room because they were physically bolted to a table. And it was uphill both ways. I practically brain downloaded the entire Common Lisp The Language (1, and partially 2) in the very early 1990's.
What bothered me was that I was a huge Apple fanboy back in the day. Apple was all about user friendly. Human Interface. Things should be intuitive. What you can do should be directly recognizable from what you can see. Even if what you can see is a control that reveals more possible operations. There weren't hidden gestures. Magic handshakes. Etc.
That's just one person's experience. It is not a generalization to everyone. But you did ask.
I'll see your senator, and I'll raise you two judges.
Back in the day Apple had Human Interface Guidelines. And I understand that Microsoft did too.
IBM had "Common User Access" (CUA), and Microsoft had "Consistent User Interface" (CUI) guidelines, which were roughly comparable to Apple's. Following those guidelines might not be as visually attractive as some of the crap being designed today, but at least it meant that people could get acclimated to your product quickly and with a minimum of confusion. In the world of UIs today, there's way too much frosting and not nearly enough cake.
Please stand clear of the doors, por favor mantenganse alejado de las puertas
Today all of that has gone out the window. I'll just give one example. Google's Material Design. Not that I'm criticizing it. But just criticizing the NAME. The name screams it is all about the aesthetics and not how well it interacts with human beings.
Ah. Well, then, you might want to read more about Material Design than the name. It actually has quite a bit about human interactions. Even if it were just about aesthetics, a lot of visual design is about how humans interact with colors, shapes, fonts. No visual designer I've ever worked with picks colors purely because they like that shade of blue.
The automobile has gone through quite a few control redesigns which are continuing. If you jumped into a Model T (maybe only early ones), you'd find it hard to figure out. Besides the controls that have been automated away, choke and ignition advance, the parking brake was operated by your left hand, along with ignition advance, throttle by right hand and gears by the pedals along with the brake being the right pedal.
The steering wheel started out as a tiller and as late as 1899 was introduced in America as a wheel. Since various controls have migrated to the wheel or right beside. The turn signal operated by your left hand, which has acquired more and more functionality such as operating the lights, wipers, high beam. On the right, there was the gear shifter for quite a while before mostly migrating to the floor. And all the various controls that can be found on a modern steering wheel. Even my old truck has the cruise control buttons on the wheel. The shifter pattern has also changed at times. Had an early 5 speed where reverse was where 1st usually is.
Speaking of my 25 year old truck, while most of the pedals are standard, on the left there's the parking brake release and the hi-lo headlight dimmer button on the floor. Turn signal only operates the turn signals with a knob on the dash that you pull to turn on the lights and turn to dim the dash lights and turn on the interior light. The wiper switch is besides it, turn one way for normal wiper operation, further for high speed, turn the other way for intermittent operation, push for squirting cleaner (still have to turn the wipers on manually).
Another set of controls that seemed somewhat standardized for a long time and now are in flux are the climate controls and radio/sound system where automakers keep screwing around with stupid touch controls. Stupid due to breaking the paradigm that the driver should be able to operate everything by feel while watching the road.
It took close to 50 years to standardize just the pedals on the car UI, while the modern computer UI is at the most 30 yrs old. Hopefully in another 100 yrs, things will have mostly settled down, but as the automobile has shown, new tech such as touch screens, still puts basic interface into flux, often with stupid design decisions such as trading easy to feel buttons for hard to use, changeable, touch screen.
https://en.wikipedia.org/wiki/Inverted_totalitarianism
Our screens are way bigger than they were back in the old days, so we have plenty of room for things like menus and toolbars. Yet the trend in modern UI design is to make things magical and non-discoverable.
Just yesterday I helped my father with a problem: the menus and toolbar from Thunderbird were gone. I was on the phone with him for a while. The task was to find the one magic part of the Thunderbird window where he could right-click and find the context menu with the checkboxes for hiding/displaying the main menu and toolbar. Thank goodness I have him running MATE so every window has a title bar... "find the blue bar at the top that says 'Inbox - Mozilla Thunderbird' Now right-click in the dark grey area underneat that, to the right of the tab that says 'Inbox'..." "It didn't work" I'll spare you the back-and-forth, he had multiple tabs and was clicking in a tab to the right of "Inbox". Once I got him over to the correct magic spot, he found the context menu and restored his menu and toolbar. (The stupid hamburger menu is part of the toolbar, and hides with the toolbar... which means it's possible to hide all the menus! And my dad somehow did so by accident!)
The original UI spec for the Macintosh required menus all the time for every app, and the menus had to be in the same place. And I learned very quickly that I could browse the menu, find the command I wanted, and the keyboard shortcut was documented right there in the menu. Hidden menus are far too magical, and if you are going to have them, the very least you should do is to make every context menu have the ability to unhide them, rather than requiring the mouse pointer to be hovering over a particular magical few pixels of your screen.
I also remember the 45 minutes it took to help my dad un-mute YouTube videos. First I had him use the MATE sound preferences dialog to test his speakers, which just took a couple of minutes. Then I had to walk him through moving the mouse pointer over the YouTube video window to make the controls un-hide... (he wasn't full-screen, why do the controls hide when there is plenty of screen real estate available?) Then he had to move the mouse pointer to touch the audio control (and a slider pops out when you get it right) and click to un-mute... and when it's un-muted it says "MUTE". Because when it's un-muted the button becomes the "MUTE" button, and when it's muted the same button becomes the "Un-mute" button. The old-school solution would be a checkbox labelled "MUTE" that's checked when it's muted; the newer way would be a GUI toggle that slides left for un-mute and slides right for mute. There's plenty of screen real estate for either of these.
I know, I know, on mobile devices these magical hiding tricks are not so pointless because screens are smaller. But desktops are not mobile devices and trying to treat them the same is a bad idea.
My dad is not stupid and I don't want to sound like I'm making fun of him. I'm just annoyed over the modern trend in UI design where everything is so magical that it's tricky and weird.
lf(1): it's like ls(1) but sorts filenames by extension, tersely
As the saying goes, a picture is worth a thousand words, or in this case 73 words.
And touch controls are stupid because knobs and buttons allow you to rest your hand on them while you use them. This means your hand does not leave the control when you hit a bump in the road. With touch controls, you have to keep your hand floating in front of the screen, where every bump and jiggle causes it to shake around relative to the screen. It's actually worse than just having to take your eyes off the road to use them. You also have to concentrate on keeping your hand aligned with the screen while the car bumps along.