Hacker News new | past | comments | ask | show | jobs | submit login
A Tale of Two Users, or How Design is Tough (georgesaines.com)
13 points by gsaines on Sept 15, 2010 | hide | past | favorite | 13 comments



Are UX discussions now really so played out that this little anecdote is considered insightful?

To summarise for comment readers "some people have a different understanding of complexity and how something you see as trivial might be confusing to others".


I think so. For the non UX crowd or those starting out. You have to start understand the issues at hand somehow.

Anecdotes are stories. Stories are great ways to communicate complex issues in an easy to digest manner.

It may not be ground breaking insightful; but still relevant.


Reminds me of the elevator button example:

You're in the fourth floor, the elevator is on the fifth floor (indicated by a display above the door) and you want to go up.

Do you press the up, or down button?

At first this seems obvious, since you want to go up, but it's only obvious because you know how elevator interfaces work.

Another valid way to interpret the elevator interface is to press the down button, since you're in the fourth floor and want it to come down to you.

Yep, interface design is hard.


Creating available, consistent and learn-able interfaces is not that hard. Making them completely ignorance proof may be though.

Having to teach something is a reasonable part of any interface. Such as having a sign near elevator buttons that says "press the direction in which you want to travel". Or a cashier that says "no, use this pen." Or a website support person that guides an elderly man over the phone on the differences between a login box and a password recovery box.

Of course, it would be the ultimate aesthetic if interfaces would just do the magic, decide what the user really wanted, no thinking needed. I think this is the unreasonable ideal some designers dream of.

Even better, when we are born we could get a dialog box that says "Lead a Long, Fulfilling Life. OK, Cancel". But then we would have to learn to read to understand the dialog box. Maybe someone could press a button for us that makes us learn to read first. Or maybe an even more magical button could be designed, that when pressed, no one ever would have to read dialog boxes and press buttons ever again, or think or live. I think the Russians actually have a button like that.


You press the single button on the panel, hypothetically labelled—hypothetically because a label is completely unnecessary when there's only one—"come to me." A lot of what makes interface design "hard" is that our minds are filled with cached solutions to no-longer-valid problems (like the problem of directing around a purely-hydraulic lift that cannot indicate its relative position to you.) The elevator now has a computer in it; it knows which direction it needs to go to come toward you, and you now know if it's coming from the "wrong" direction in order to pick an alternate lift out of a bank.

Some of these things are so codified that people would react badly if we took them away, though. For example, JRPGs no longer need to display numeric damage tallies—but it persists because that's what makes a JRPG a JRPG.


But that's wouldn't solve the problem that elevator buttons are designed to solve.

You have to tell it "I'm going up" so that it doesn't stop downward-bound elevators on your floor and slow the system. Your single-button interface would make things considerably worse.


That assumes, though, that there is more than one elevator. There is already well-known and battle-tested solution for that case: put a destination-floor-number panel outside the elevator bank (not connected to any single elevator, but rather offset), with the current floor's button removed. To remove the possibility of thinking that the panel of buttons has anything to do with the numbers the elevators display, remove the external displays from all the elevators. When you press the button, the first elevator available (and going in the right direction) opens at your floor, already set to go to your destination floor. (If you really insist, you can pulse the backlight of the activated destination floor button as the elevator nears, in a sort of marco-polo way; I've seen a very upscale hotel do just this.)

Now, what I was referring to in my previous post was a single-elevator scenario (which is most elevators outside of hotels and apartment buildings.) In the case of a single elevator, there's really no difference between the elevator either:

1. passing you, turning around, arriving at your floor, picking you up, and heading to your destination floor; or

2. opening at your floor, taking you along with whomever else is inside to their destination floors, then turning around and heading for your destination floor.

Either way, you spend the same amount of time waiting (arbitrarily either inside or outside of the elevator), and the elevator opens the same number of times. When there is one elevator, all that really matters is getting inside it.


Capacity is often a bottleneck in large buildings. The elevator might be full as it passes going the wrong way, so getting in isn't even an option. Also, more people might show up on your floor by the time the elevator passes going the right way, necessitating another stop.


Firstly, if capacity is a problem, there really shouldn't just be a single elevator serving this supposedly "large" building. But since a lot of places can't afford all the elevators they need... you know how a full bus will just ignore everyone at a bus stop, and only stop when someone wants off? Elevators know when they're full.


Isn't this more an example of when it's best to let a customer go?

Is the cost associated with meeting this small percentage of customer's needs worth actually meeting them?


Depends. Not with mission critical processes like checkout IMHO. What this piece points to is a flaw in design processes. As designers and developers we see things through a logical lens of our own making.

What we don't realize is that the real world and other factors influence our target users. It's the mark of a good designer when he/she pushes user testing to understand what people will do with what we build.

In this specific case, we all know you use the magnetic pin but not everyone does. It's hard to believe but the sometimes don't.


I agree that if your site metrics show large numbers of people set to buy but abandoning their purchase because the checkout process is confusing, then you should work to make it more user friendly.

But what I'm saying is every designer should be asking 1. How many people are having this unique difficulty and 2. Is it worth the time and COST to fix it? Moreover, will adding the information/steps to needed to clarify the process bore/put off the majority of customers who already know how to use this (for example, a tutorial that either adds an additional step or doesn't allow users to bypass it)?

A good designer should be able to plan for contingencies, but a good businessperson should be able to draw the line between pleasing a customer and watching their bottom line.


Case in point:

I had a friend's mum who asked me once (in reference to vertical scrolling on ie. browsers and Microsoft Word):

Why do you press down to make the page go up?

('down' is the down cursor button)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: