iPhone’s Keyboard Trying too Hard?
Thursday / 28 June 07
Tomorrow is the big day. Yes, I am talking about the Apple iPhone. Now, I am not talking about the iPhone from a fanboy perspective, but from an interface designer perspective. The iPhone is one of the first consumer products that sports a sleek multi-touch display as its primary way of interfacing. As such, it is the first time we will be able to interact with such a product in a completely new way. Also, we will be able to see Apple’s take on some of the inherent problems and their solutions.
I was especially interested in the iPhone’s virtual keyboard. Due to a lack of tactile feedback on a touch-screen it is necessary for them to create an interface that does a better job than a regular keyboard, if they wish the user to accept that one problem. Having seen the demonstration videos of iPhone’s keyboard, I instantly noticed some problems. Instead of waiting for people’s experiences, I wanted to throw this out in the wild and see if other people have different ideas.
Inherent Problems
Touch-screens are old and even multi-touch screens date back from 1982. In the eighties people have conducted usability studies and designed many of the interaction models we see in the iPhone today. There are several problems with touch-screens and none of them have been solved yet:
- Finger-operated touch-screens are excellent in the speed department, but slow in accuracy with which an user can operate it (Albert, 1982). This is especially a problem with smaller screens where interface elements are relatively small.
- Stylus-operated touch-screens are comparable to operating the interface with a mouse in terms of speed and accuracy (Mack & Lang, 1989).
- Lack of tactile feedback requires the interface to rely on other means, such as visual or auditive feedback. Especially visual feedback can prove to be a problem on smaller screens where hands or fingers may block a clear view to the operated interface element.
- “Gorilla arm”-syndrome was observed mostly with larger touch-screens where movements cause strain on the arm and shoulder after a long period of time.
Only the first and third apply to the iPhone. These are the two problems Apple needed to attack in their interface and it shows in some of the features of the virtual keyboard. I consider accuracy in one of the most important aspects in interface design. Accuracy defines the difference between hit and miss — between getting it right or wrong.
Apple implemented two types of solutions to attack this problem: suggestions and expandable touch areas in the keyboard. I couldn’t help but notice something seemed wrong in both solutions.
Word Suggestions
A keyboard, especially on a mobile device, is usually strongly connected to an intelligent text recognition system. Systems such as T9 provide suggestions based on the user’s current input and behavior in the past. A user can choose to accept or ignore the suggestion.
The iPhone implements a similar system and likely only supports an English dictionary right now. In the demonstration video they show the user entering the word ‘ouzza’ and suggest the word ‘pizza’. The suggestion itself can be perfectly correct and there is indeed not a word such as ‘ouzza’. However, that is not the point I am going to make. The interface dictates that the spacebar is used to accept the suggestion. The question I ask is: what if I want to insert a space, but not accept the suggestion?
This scenario is not uncommon. Such dictionaries are based on plain English, but we all know users have their own dictionary as well. Besides the textspeak, it is not uncommon for people to use words specific to an industry, sport or whatever. Sometimes, we add such words in the dictionary, so they will appear as suggestions the next time. The convention across mobile devices seems to be that a separate interface element can be used to accept the suggestion — in the case of the iPhone I would have thought it to be more logical to actually select the suggestion (although this isn’t perfect either). This keeps the behavior of the spacebar singular and clear and does not require an extra character to accept an already written word.
Expanding Touch Areas
Another seemingly clever feature in the virtual keyboard seems to be the expandable touch area. This means that the area with which a certain key on the keyboard can be pressed expands, when surrounding keys would not form a word with the current input. In the video the user writes the word ‘tim’ and because there are no correct words to be formed with a following ‘w’ or ‘r’, the touch area of the ‘e’ is expanded.
In theory, this is awesome. Some people immediately said it was a perfect application of Fitts’ Law. I disagree. This has nothing to do with Fitts’ Law and I think that in its current implementation it will provide more frustration than an advantage in the speed/accuracy department.
Fitts’ Law states that the time it takes to interact with an interface element is defined by the distance to and the size of the element. Looking at this feature in the demonstration video, it seems as if the expanded touch area is not visually communicated. How can the user possibly know when this occurs? The user will still want to focus on the normal touch area, because that is what is communicated. There is no visual difference in the size of the target, thus no dynamic application of Fitts’ Law.
Now, the first thing you might say is that is excellent for error handling. This transparent expanded touch area allows for the right input in cases where accuracy was low, usually in the first days of using the virtual keyboard. I could agree, but this is mostly irrelevant. Exactly because the expanded area is not visually communicated and the algorithm is complicated, the user cannot predict nor rely on its behavior. If this is not possible, how is he supposed to use this behavior to his advantage?
Finally, I must briefly mention another edge scenario. What if I do want to enter a ‘w’ or ‘r’? The video does not make this clear and suggests that the touch area mostly covers these keys.
Share Your Experiences
I have to admit this was a bit of a rant. I can’t imagine Apple did not foresee these problems. So, I am looking forward to be proven wrong and hearing about your experiences with the aforementioned functionality in particular.
Judging by the video, Apple did succeed in improving a regular keyboard experience. One of the strengths of touch-screens is that it allows for contextual interfaces and Apple uses this by providing variations of the keyboard depending on the task you are trying to complete.
As for the iPhone itself, I would love to have my hands on one just for the interface alone. Unfortunately, I couldn’t see myself actually using one.
1Jochem posted:
29 June 07, 03:58:59 AM
Finally, I must briefly mention another edge scenario. What if I do want to enter a ‘w’ or ‘r’? The video does not make this clear and suggests that the touch area mostly covers these keys.
I think you would probably have to complete the following sequence:
1) Cancel the suggestion by tapping it. This is - especially for everyone that's ever used another PDA, where tapping the suggestion accepts it - counterintuitive, takes your fingers away from the main control element, which also increases the chance of mistakes, either in tapping the suggestion or returning your fingers to the keyboard.
2) Type a space to convince the iphone you're no longer working on a word, and resetting the 'smart area' thing.
... Then, either ...
3a) Type the letter you want to type, move the cursor one right (by dragging), pressing backspace.
... or ...
3b) Typing backspace, then the letter you want to type.
I think the first option is probably the more likely, seeing as the iPhone might assume you're going back to editing the previous word if you remove the trailing space and reactivate the suggestion.
Another thing that I noticed when watching the video is that it apparently takes the iPhone forever to come up with the suggestion. The demo-dude had typed 'availabl' before the iPhone came up with 'available', not a very good suggestion, as it would only save you a single keystroke.
The way suggestions work on my windows mobile device (with it's bad tiny keyboard, and the slow stylus and all the other things wrong with it) is imho a lot better.
After typing a few characters, it comes up with a suggestion that you can tap to accept. If you don't, and keep typing, after two or three more characters, it'll start suggesting another possible word. Afterall, you didn't use the first one, it's clearly wrong (or, you're not interested in the suggestions). Admittedly, it doesn't contain a 'full english dictionary' (I wonder how full the iPhone's really is), but it's still something that's saved me quite a few keystrokes.
Another thing I'd like to remark on, is that Microsoft or any number of third party developers, could practise the old tradition of stealing ideas from Apple, and simply implement an installable plugin for the standard ms keyboard.
My last point is how huge the iPhone is. I have a windows mobile PDA myself. I don't have wide fingers, but when I try and put both thumbs on the touchscreen, they cover the entire width of the device, not leaving any room to a) see the keyboard, b) differentiate where I push the screen.
Even if the man in the video has disproportionately tiny thumbs, the iPhone is still the size of a small refridgerator. That's a major disadvantage for me. I, like so many <strikethrough>geeks</strikethrough> phone users with me, prefer small formfactor phones. It's only the hamfisted, nearsighted, and older-than-middleaged in the world that, on average, want to have a larger phone. Apple's key demographic of the hip college student is also going to be turned off the iPhone if it's as big as it seems in this video.