On-Screen Keyboards

More and more projects are done concerning applications for PDA's, mobile phones a.s.o. Quite a lot of those include the use of a keyboard which often is a keyboard on-screen. Since I did a study (during my study Interaction Design) for Philips (Philips Media Benelux) and IPO (Institute for Perception Research, Eindhoven University of Technology) concerning the use of on-screen keyboards (or 'soft keyboards'), I feel urged to a few remarks on the subject.

1. Do not use the QWERTY-layout for the default keyboard.
Originally keyboards that came with typewriters were alphabetical. Those keyboards were mechanical devices that used a kind of small hammers (with a character on the end) to produce a character on a sheet of paper. When you typed with a certain speed, often the hammers would collide and get stuck. You had to stop then, so in the end the overall speed was not too good.

A different layout became popular (amongst the manufacturers), especially since the inventor sold his idea (the QWERTY-layout, I believe it was to Remington). In this layout, the most used letters were assigned to the weaker (slower) fingers, thus resulting in a slower speed, but effectively in a faster speed, concerning the whole proces.

Since the invention of the IBM electric ball-typewriter the need for the QWERTTY-layout has gone, but we got stuck with it, the alternative being reschooling every typist in the world.
A design that could deliver more speed than the alphabetical keyboard is the Dalton-layout.

2. Use the Alphabetical Layout as a default
Most people know the Alphabet quite well. Even experienced typists do know the alphabet better than the QWERTY-layout. They do know the QWERTY-layout, of course, but most of it is muscle-memory (That is what you use when you need money. Even if you know your code by heart, most people have to make the moves with their fingers to remember the exact code). The Alphabet was learned when you were very young, you do have a profound fuzzy knowledge of it. You know that the 'T' is near the end, while the 'F' is near the beginning, which is enough for fast 'hunt & peck'.

3. The underlying usability is more important than the keyboard-layout.
This was my main conclusion in the end. I felt I had the obligation to mention (not everybody was pleased with that, since officially it was outside the scope of my research. Well, in that case, call it a case of serendipity).
A great example of underlying usability is the search-functionality as used in OS X on Apple Computers. It is sufficient to provide the first characters and you can scroll from there.

Last but not least:
4. Let the user choose its favourite keyboard-layout
Test may show that an Alphabetical layout is faster, but that doesn't mean that it is in the perception of the user. If he/she prefers another layout (and that may very wel be because he/she perceives the QWERTY-layout to be faster), let him/her choose! It may be an emotional decision, but if it makes somebody happy, we should not prevent it.

More information
Different keyboard-layout in different languages (did you know there even is a Dutch keyboard-layout?)

Usability vs Experience

Usually people look upon me as a usability-enthusiast, an evangelist even. To be honest, I was. In fact I still am, but only so far. When I started as an Interaction Designer my focus was on making it easy for users to use products, make it understandable, make things transparent. Important was to use the right labels, the way the information should be constructed, aso. The reason that I started my study Interaction Design was that I always had the feeling that "things weren't right", "it should be possible to make to make things better". Then I stumbled upon Psychology of Everyday Things by Donald A. Norman. And suddenly I understood: "Wow, things can be done! You just have to design them in a better way". So I took up Interaction Design, concentrating on cognitive ergonomics, usability.

But I always had the feeling there was something more. I couldn't exactly put my finger on it, but the feeling was that I seemed to miss something. Again it was Don Norman, he then published Emotional Design. And again things fell in place. After that I discovered Train of Thoughts, by John Lenker.
So I finally grew up and understood why people used SMS, despite of its horrible interface (instead of typing 10 numbers and speaking what they had to say, people punched 10 numbers and kept punching to get the message across). Like I understood before about games, games have to be exciting, it does not matter if they look good (ouch!), it's just the box that it is sold in has to look good. And the same goes for websites, people don't visits a site because it looks good, but because they find the needed information there, preferably fast. And I could go on this way.

Since I live in the Netherlands, and we are a cycling country, just one more example I used in my classes. When you cycle from home to school (or your office or whatever) each morning and you have a choice between two routes, one along the highway straight on for several miles, the other on the other side of the same highway but it meanders along a stream, leads you through a few trees, but also is a mile longer. When you're in a hurry, you will take the fastest route. But when people have the time, a majority will choose the longer route, just because it is more appealing, more pleasant to use.
Indeed, just so much for functionality, for usability. That's only a part of our lives.

User Testing - I

Cheap & Easy! That is, as far as costs are concerned, not the results.
Today we look at Card Sorting. Since real users are the people who have to work with what you are developing, it is important to know how they look at their tasks, what is their mental model? Which labels do they use? What is their terminology?

Card Sorting: What is it?.
Card Sorting is a method to get user-input even before you start to develop. It provides you with the user's view on the subject. Two methods: Open Card Sorting and Closed Card Sorting.
Open Card Sorting is done with blank cards, labels are provided by the end-users. Closed Card Sorting is done with labeled cards, users provide the position in their mental model.

Card Sorting: How do you do it?.
1. you have to collect some real end-users (the ones that really have to work with it), up to 15 should be sufficient, but less is OK, even 2 would help a lot.
2. you divide them into groups, about 5 persons each.
3. you provide them with cards, ca. 30 cards usually is OK and one or more felt-tips.

Open Card Sorting
4. Open Card Sorting: you ask your users to write down the issues that are important to them. Each item is written on a different card. When they are done:
5. ask your users to put the cards into stacks.
6. ask them to label each stack.
7. collect the stacks.

Closed Card Sorting
4. ask your users to put the cards into stacks.
5. ask them to label each stack.
6. collect the stacks.

Card Sorting: What does it do for you?
Card Sorting provides you with insight in the mental models that users use doing their tasks. It gives you the labels they think with, making it easier for you to taylor your design to their needs.

Tip: Provide your users with feedback, let them know they did make a difference!

More on Card Sorting:
www.infodesign.com - Usability Resources
www.boxesand arrows.com
Jacob Nielsen (www.useit.com) on Card Sorting
More on Mental Models
Wikipedia on mental models
Ruth Byrne's website on the subject
Don Norman's website, you should read anything Don has written.

User Tests

To make things usable it is important to do tests with (real!!!) users. It is important to do the tests with real end-users who will use the application, the web-site, or whatever needs to be tested. Testing by designers/developers doesn't do anything, they know the application, they are profound with computers, they are NOT the average user.
Are user tests profitable? Oh yes, they are, look it up at Jacob Nielsen's site (http://www.useit.com), you will find several items about the subject in his Alertbox. (You can sign up for his newletter)

Of course User testing is much more profitable when you do it cheaper. And you can do it cheaper! Much cheaper! For quite a long period I was hired either to do user tests or to do a heuristic evaluation (providing comments as a usability specialist). This can be quite time-consuming for a website with more than just a few pages or a complicated application, and therefor expensive. But even more important: you are very limited in what you can do. For instance, you do not make people happy telling them to do it all over again. Not if it is the result of several months of hard labour..... Nevertheless, sometimes the only reasonable advise is just to do that.
And always I thought: "why didn't you come sooner?" It would have been easier for me, cheaper for you, and much better for the poor sod that has to work with it.

The next posts will be about how to do tests (as a design team, you will not have to hire a usability specialist, although if you wish, you can mail me...) I will write about card sorting and paper mock-ups with linx for more background information.

Newbie

Well, since this is my first post and my first blog I consider myself a newbie, despite my age. (For the curious ones, I will make a profile, probably some time in the near future). Most posts will be in English, but some may be in Dutch.
Subjects will range my whole scope of interests, usually connected with the fields of my profession, Interaction Design, with emphasise on Usability, Cognitive Ergonomics, Psychology, to name a few, up to Buzz marketing, and so on. You also may find posts concerning RIA and Web 2.0, although those could be just links to interesting articles, sites, lists et cetera.
For starters: http://www.web20searchengine.com/web20/web-2.0-list.htm.
For Dutch starters: http://www.usabilityweb.nl, with several articles that I can recommend, for instance Gestaltpsychologie en Webdesign, Zoekmachinewaanzin (search engine madness) and many more).