I agree with some of the points (like ambiguous icons and vague error messages), but not all of them.<p><i>‘If you are an X, then you have to fill in Y and Z. If you are not an X, please only fill in Z, unless Z=1, then you should fill Y too.’</i><p>This could be simplified considerably: "Fill in Z. If you are X or Z=1, fill in Y." Although I'm not surprised that refactoring boolean expressions is something that a lot of people seem to have trouble with.<p><i>In some cases it can be better to have a more human, relative notation, e.g. ‘within the last hour’, ‘yesterday’, ‘next week’…</i><p>Actually I really, <i>really</i> abhor this style of date formatting since e.g. seeing events all dated "yesterday" - possibly on different sites - make ordering/comparing them nearly impossible. Is ISO 8601 really beyond comprehension for most of the population?<p><i>The user interface is totally empty when starting out</i><p>Unless you mean <i>completely empty</i> (as in devoid of any buttons or other widgets), I don't think that's a bad thing. It's funny that it mentions adding an explicit instruction "Click ‘add contact’ to add your first contact." when the next point is "Visual clutter", since it could very well be the case that the users are confused because after removing that "Visual clutter", "add contact" doesn't even look like a button anymore! Incidentally, this trend of "flat" UI designs is another thing I loathe.<p>Am I the only one who thinks it's rather sad that, despite society spending great effort over the last few hundred years to improve literacy (which was quite successful), we're now dumbing-down software to encourage basically the opposite?
Recent stupidity: UI asked for postal code (knowing the address being entered was Canadian.) Without thinking, the user (not me) put it in as L6A 3Z3. When the form was submitted, it was rejected with the field flagged in Red. the error message said like "please enter the Canadian postal code in the correct format, AnAnAn with no spaces." (AnAnAn? Cryptic to non-coders; the user was baffled.)<p>Postal codes <i>are</i> written with the space:<p><a href="https://en.wikipedia.org/wiki/Postal_codes_in_Canada" rel="nofollow">https://en.wikipedia.org/wiki/Postal_codes_in_Canada</a><p>If you're already validating the format, how much effort would it take to recognize and accept the common variant which has the space in it? A [ ]? in the regex, and a strip spaces function call.
>Use of image-only icons/buttons that have no distinctive meaning<p>There's nothing worse than this, along with no alt-text on web pages. With font-icons being so common nowadays, designers want to remove all text and revert back to hieroglyphics.
EDIT: disabled zoom?<p>It's weird to me that Google fails, hard, on so many ui things.<p>Here's one example from the Gmail app on iOS, from the feedback screen.<p><a href="http://imgur.com/a/koKRb" rel="nofollow">http://imgur.com/a/koKRb</a><p>The [done] button is where the [return] button usually is; and it's directly over the [send] button. So if you don't notice that return is now done, and tap it twice to get a new paragraph, you just sent your unfinished feedback.<p>I blame Google, but it could be Apple being stupid with the dreadful iOS keyboard.<p>> possible to get a pretty good impression of whether the user interface was thought about or rather just an afterthought.<p>I'm sure Google spends a lot of time and money on the UI. Which just makes it more frustrating to me. EG Chrome's (on iOS) address bar does not let you select and copy an address until the page has nearly finished loading. And to paste an url into the bar you don't tap the cursor but some mid-point of the address box - except it's flat UI so you just guess where to tap. Items in the burger menu take five seconds to become clickable.
> "Too much text/instructions"<p>I disagree that too many characters or too much instruction is a sign of a bad interface.<p>For users trying to complete their desired task, an interface lacking instruction/explanation will often cause lower task completion rates and longer average task completion times than an interface with verbose instructions. This is especially true for new users.<p>Don't get me wrong: I dislike wordiness and consider myself an advocate for simplicity, minimalism, and concision – and that might be the author's sentiment. However, while bloated instruction may hinder task completion and user engagement, incomplete instruction will prevent it.<p>I believe more useful measures of good instruction are when it is presented, where it is presented, and how well it is understood by users.<p>For example, a well-designed mobile app avoids a text-heavy interface by disclosing the appropriate amount of instruction at the appropriate time. Limited screen real estate happens to be a handy constraint in this case, and transfers more instruction responsibility onto a well-designed onboarding wizard. That wizard will really only get one shot to educate users, usher them to their 'aha' moment as quickly as possible, and hold their hand through the few actions that strongly influence successful user engagement and retention.
These are always useful reminders. It's easy to get so lost in designing a system that you forget the perspective of the user, which is what really matters.<p>One of my current favorites is Rocksmith (guitar learning software), where upon exiting you are prompted with "Do you wish to continue? Yes/No". You as the user need to essentially state, "Yes, I would like to continue...exiting the program". This is right next to another menu option that I sometimes accidentally select which prompts me to "Quit", which in fact just brings up the "select profile" screen. It does boggle the mind how some things like this get all the way to production.
I mostly agree, however one point I disagree very strongly with is the first: <i>An Excel/development IDE – like user interface</i><p>IDEs are horrible, horrible examples of interface design I can easily agree with that. However,<p>> Excel is the most unconstrained application most people know, and a lot of software starts out as an Excel sheet before growing up into ‘real’ software.<p>Usability comes first and foremost. If most people are familiar with Excel that means that the Excel interface language is a good language (if not the best). I don't care if it doesn't sate the desire for an "elegant" user interface. The fact is that nearly all users already know how to use the interface and are extremely proficient with it: what more could you possibly want out of an interface? Scrolly fiver-finger gestures and other complete shit that you don't need? YAGNI.<p>We make huge sales based on usability. Our bigger customers have historically placed us into a competition during our initial sales pitch. If the sale ends up in that situation <i>we always win.</i> Why? Because by the time customers are using our software to build toy/test solutions our competitors are still training the customer. Why? Because usability comes first and foremost, because our UX guys don't let pet peeves about Excel get in the way of making great and usable interfaces.<p>If you can "unconstrain" your solution enough to make it fit into a spreadsheet then I say, "go for it."
The Design of everyday things [1] (Found on Coding Horror's recommended reading [2]) was crucial in understanding how people intuitively interact with the world around them - including your web application - based on the visual cue's you provide.<p>While this article provides some metaphorical fish - I found the Design of everyday things helps you become a fisherman.<p>EDIT: Swapped the order of references.<p>[1] <a href="http://www.amazon.com/Design-Everyday-Things-Revised-Expanded-ebook/dp/B00E257T6C" rel="nofollow">http://www.amazon.com/Design-Everyday-Things-Revised-Expande...</a><p>[2] <a href="http://blog.codinghorror.com/recommended-reading-for-developers/" rel="nofollow">http://blog.codinghorror.com/recommended-reading-for-develop...</a>
Well, this is nothing new. It's called heuristic evaluation and it's been around at least since 1990 [1]. Almost every item in the article's list maps to a simple and well known heuristic.<p>It's very strange (or sad) that the author doesn't mention heuristics in his post and talks about his conclusion as something "new": "it dawned on me that a trained eye can often spot unfriendly software a mile away." It may have "dawned on you", but it's nothing new.<p>It never ceases to amaze me, how oblivious of basic usability the people in IT are.<p>[1] Nielsen, J., and Molich, R. (1990). Heuristic evaluation of user interfaces, Proc. ACM CHI'90 Conf. (Seattle, WA, 1-5 April), 249-256.
This feels like one of those articles that the people who need to read it never will; developers who don't care to improve their craft, and are churning out software to check off some box in a list of features demanded by management.
I agree... especially with the comment about terse and unclear error messages. I recently got lured into the customer-facing side of things, and nothing annoys me (or our new users) than a message like "Incorrect syntax" when I used a unsupported symbol for my password.<p>First, be clear with your design and instructions. In my hasty password example - "Please create your password (6 or more characters, letters and numbers only)"<p>Then - if the user messes up, explain a bit... "Whoops, please only use letters and numbers"<p>So, so much nicer. :)<p></rant>
A great user interface is dangerous when it's disconnected from the internals of the software, because the user starts to feel the software "is only pretending to be nice" and is fighting against them.<p>When an app loses hours of your work, you start to wish you used an app that took more clicks to use but actually focused on the internals.<p>I'm making daily videos at Rate My App (.com) and the latest version of ScreenFlow 5 (screen capture software) is known for the best user interface, but in fact impossible to feel safe recording even the most basic videos with.<p>- Forces you to delete your videos with endless prompts or forcefully shut down the app if you do a File -> Rename...<p>- Let's you record a new project but will not actually save it if it recently had an issue (above), so you lose that one, as well<p>- Does not explain that clicking Remove in the infinite prompt actually deletes all the video content. Does not backup for you, either<p>- Doesn’t always show YouTube categories, but allows you to click Start Upload, in which case the uploads remain in the upload queue forever. Issues or errors with video sites are not mentioned, the upload just stays in the queue.<p>- Warns about uploading videos greater than 15 minutes long on YouTube even though you’re exporting a 5 second range<p>- Their most important feature, video actions (keyframes) don't work. Modifying one randomly modifies another one, so I stopped using it. This was a major bug until it deleted a 1.5 hour project of mine. Now, it seems minor.
> Unneeded use of separating lines, grouping boxes,.. less of these = easier on the eye.<p>I seriously disagree with this particular assertion. I find that "flat" design trends make things much <i>less</i> usable, and this is no exception. It is far easier to group things with visual references than with lack-of-visual-references. And it is far easier to find things like buttons if there are, say, actual indications that they are buttons.
A bit dated, but relevant: <a href="http://interfacehallofshame.eu/www.iarchitect.com/lotus.htm" rel="nofollow">http://interfacehallofshame.eu/www.iarchitect.com/lotus.htm</a><p>And its complement: <a href="http://hallofshame.gp.co.at/fame.htm" rel="nofollow">http://hallofshame.gp.co.at/fame.htm</a>
> The user interface is totally empty when starting out. When seeing a ‘clean’ installation or account, there are no instructions to get you started.<p>This is such a big problem, and such low hanging fruit, it sucks to see products fail to implement proper onboarding. I understand why... writing user interfaces is a pain in the ass. Onboarding is usually done after "finishing" the code (of course the code is never actually finished), so generally most programmers are probably so exhausted after writing the code that implementing an additional onboarding process seems like too much trouble, because "JUST SHIP IT!"<p>It's like the programmers do 95% of the work, then stop caring about the last 5%. But it just means that the 95% goes to waste.
I've never heard the term "hallway testing" before, and I'm assuming I'm inferring its meaning correctly. I'll be using it in the future.
<i>On the other hand, developers have a knack of seeing everything as the user interface they’re most familiar with: their development IDE.</i><p>Along with the bizarre idea that the best interface of all is one that forces the user to write code of their own. Look folks, if everyone wanted to type then text adventures would still be top of the video game charts.
Shotgun full of buttons is your standard coder designed UI.<p>Many of the common alternatives are worse though, I prefer a shotgun full of buttons to many of the supposedly simpler systems unless they are very well thought out.<p>Good UI is very hard.<p>edit - to me, really good UI is simple shit like things that gather system stuff, then install while gathering user stuff. When that started coming in as standard with many linux installs, I almost wept ;)