Hacker News Reader

Hacker News

Comment by StavrosK | Hacker News Reader

Parent Full thread

StavrosK

You know, I thought I had UI design mostly figured out, using a few simple rules:

* Minimise the total number of clicks needed to access the sum of the program's functionality.

* Make frequently-used functionality easier to reach than rarely-used functionality.

* Don't have more elements in the UI than is easy to parse quickly.

* No surprises.

However, after reading this article, I get the feeling that it's a more arcane art, which not many people know. What are some simple rules about good UI design, in your opinion?

View on HN

In my opinion it isn't arcane, at least, no more arcane than other "soft" design fields. But the simple rules are complicated to follow.

Your first three principles are about efficiency. That's a big part of usability. When you reduce clutter, you make it easier to use by clarifying the next action the user should take.

But "No surprises" is rather complex. It means that you not only have to explain what you are doing, but also work with the user's mental model, which might be totally wrong.

Here's an example from my current project. We are designing a web form to upload images and video to Wikipedia. In 2010, the users' model of media uploads comes from Facebook, ImageBucket, and maybe Flickr. Even after you make them click on a box that says they give up their own rights before proceeding, they don't understand what they did. Apparently people's mental model just continues on blithely until it smacks against a major contradiction.

So we've tried various metaphors. "License your work" was replaced by "Release rights", and the next iteration is going to be "Donate your work". That wording was the suggestion of a usability testing and analysis firm we hired, by the way. So it does take a certain amount of inspiration to come up with alternatives, but recognizing where the problem areas are is just a matter of measurement.

Inspiration is aided by just having a bigger bag of tricks, too -- knowing common problems and solutions. For instance, if there is so much to do on one page that users are finding it overwhelming, it is helpful to break it into multiple steps (a "wizard"). That's another approach we have taken.

Anyway, I'm not sure if my expertise with usability is so deep that I could pronounce just a few principles. Maybe "reduce cognitive load to the minimum" is one, which has been more memorably distilled by Steve Krug as "Don't Make Me Think". Since we want to avoid onerous information transfer and cognition, there are some models of communication that might be helpful here; a classic text on human communication is The Responsive Chord, by Tony Schwartz. He argued that most communication is about activating stuff you already know (at the right moment), which you can see at play in the example I gave.

But ultimately it's about motivating human behavior, so it's only about as complex as the human heart. ;)

Thank you, that's very interesting. You can't get everything right the first time, but yes, it doesn't seem to be something magical that you don't have any control over (other than praying to some UX god), either.

One of the key things about UI design these days is that it's very easy to fall into the 'widget trap'. This is where you have some content you want to interact with, and instead of really analyzing what the best way to interact with it is, you just throw the data in a standard widget, like a listbox or textarea. It might be a really nice listbox or textarea, but there may also be a much better way of interacting with the content that wasn't explored. Widgets are the typical area where this happens, but it can be generalized to any situation where people just go with pre-established conveniences rather than try to find the best solution.

The problem with that is that widgets are your language, so you tend to think in them and analyse the problem in terms of widgets. It's sort of the Sapir-Whorf hypothesis applied to UX, it's hard for someone to think in widgets they don't have.

Right, but that's what I'm talking about. Widgets should never be your language. If you think widgets are your language then you've fallen into the widget trap.

Solid list.

Personally, I spend a great deal of time putting myself in the mind of the user, trying to figure out how we as humans visualize things and simplifying the actual underlying task at hand.

The list you summed up does work for 95% of all cases - especially if you're just thinking about a "website". However when you're trying to create something new or working with some new concepts you have to imagine how the user pictures things so that it matches his thought pattern as much as possible.

This also includes placing things were the user is most likely to look for them etc.

ISO defines usability as "the extent to which a product can be used by specified users to achieve specified goals with effectiveness, efficiency and satisfaction in a specified context of use.

So your right with the number of clicks, but you also have to make sure that the user can complete the task (effectiveness), and that they are happy (satisfaction). From our own research the three; effectiveness, efficiency and satisfaction are very highly correlated together.

If you nail all the things you mention, you'll probably be far ahead of most software out there.

Please don't take offense, but that's a very programmer-like approach to usability.

Usability is so much more about mental models than it is about efficiency, about helping the user feel confident & safe, than about a lack of surprises. Sensible hierarchy and visual relationships are a big thing that most people seem to forget.

Workflow, too, is critical. I've used a lot of software where every screen is mathematically, scientifically "usable" but the software overall was crummy because it duplicated a workflow from another piece of software which had got it wrong initially.

For example: 100% of email clients. Feed readers. Project management tools. They're all the same.

Surprises are actually very good in many cases. Even "ease of use" is something you don't actually always want.

My interfaces break all sorts of rules like "use standard widgets" and "conform to expectations" and "no surprises" -- but that surprise then results in an "AHA!", great enjoyment and power, and that's why people write love letters to our customer support account.

Reply on news.ycombinator.com