In yesterday’s TechCrunch post, “Hey Apple, What The Next iPhone Really, Really Needs Is A Much Better Keyboard“, Natasha Lomas argues that the iPhone keyboard is lagging behind its competition. I disagree.
I’ve spent the last six weeks using a combination of Android’s built-in keyboard and SwiftKey, and it’s been an interesting opportunity to think about the nuances of keyboard design. Of course there’s room for Apple to improve; but I believe they still have the best keyboard in the business.
Qwerty keyboards owe their letter layout to early mechanical typewriters…Qwerty letter order made sense when the metal legs arranged around the typewriter’s amphitheatre needed to be positioned to avoid clashing with one another as they can-canned their letters up and over to stamp on the ink-soaked ribbon…Such mechanical thinking is clearly redundant in today’s digital word [sic].
QWERTY was, indeed, devised to avoid jamming typewriters. But a digital world doesn’t eliminate the need for mechanical thinking; it just changes the machines we think about – in this case, our hands.
QWERTY’s design kept keys apart that were likely to be pressed in quick succession. Think about how we operate our mobile keyboards when we want to type fast: two hands on the phone, thumbs tapping. Repeatedly tapping one thumb is slower and less comfortable than alternating thumbs. A design like QWERTY’s that maximizes the distance between successive taps increases the frequency of thumb alternation, arguably improving both typing speed and ergonomics.
Next Word Prediction and Cognitive Load
Next word prediction, for starters, has broken the traditional letter by letter rhythm carried down from the days of mechanical typing. It’s a big step up from auto correction (which, yes, the iPhone has).
For those who haven’t tried it, next word prediction (which most Android keyboards have) provides the user with a set of possible completions for whatever she’s typed so far. The most common implementation is a horizontal bar, just above the keyboard, containing three completions. It may not even require that you start a word: it can predict based on the context you’re in and the words you’ve already typed.
Which is great, right? Every time it guesses your word you can avoid typing it. Except that it’s not that simple. Typing efficiency is about more than the number of taps required. It’s about decision-making or cognitive load. Looking at a set of three words takes time and brainpower. How many taps’ worth? I don’t know. But there’s a cost to pausing whatever you’re typing, looking at a set of choices, and picking one.
And that’s not the only decision involved. With autocorrect, you have either zero choices (no suggestion is present) or a simple binary choice in the context of your current thought process: is the word in your head best represented by whatever you’ve typed, or by the thing that’s popped up next to it? With next word prediction, you’re faced with a constant, less contextual decision: do I keep typing what I’m typing, or do I take a look at the choices the keyboard is offering me (which, for some reason, tend to appear in a location I wouldn’t otherwise look at)? That, in turn, can be further broken down into several decisions: Am I typing the word correctly? If so, is it long enough that it’s worth checking to see if there’s a completion available? Or if not, do I think it’s worth looking for the correct word amongst my choices?
This isn’t just theoretical stuff: when I’m typing on Android I honestly feel like I’m constantly interrupting myself to use the completion feature.
The iPhone’s keyboard is a relic of the past in more ways than one. It has barely changed since the phone was first introduced, way back in 2007 — the proverbial ice age in technology terms.
The iPhone keyboard’s UI hasn’t changed much, but it’s improved tremendously – which is sort of ideal, since users didn’t have to adapt to anything new. I celebrated the day it stopped autocorrecting “for” to “fir.”
I can think of one UI change since the original iPhone: when you backspace after typing a word, you get a popup with alternate completions, including whatever you originally typed. I’ve never gotten the hang of using it, and I think it’s the exact problem I described above: it’s easier to correct manually than to make the decision to use the popup.
In the touchscreen era, the most disruptive text input technique that has gained significant traction was devised by Swype…Instead of tapping, the Swype keyboard lets the user drag a finger to chain letters together to form words.
While whole-word gestures are conceptually intriguing, I have some fundamental usability issues with Swype:
- Traditional tap keyboards provide feedback for each tap above the key itself. Swype draws your gesture as you move, but that drawing is largely obscured by your finger.
- The atomic nature of a word gesture can be problematic. Swype corrects for missed keys just as a tap keyboard does, but you don’t have any way of knowing what it’ll come up with until you’re done with the word. It’s up to you to decide if it’s worth bailing mid-gesture – at which point it’s two taps to get rid of your mistake.
- Lengthy drag gestures can be awkward when operating a phone with one hand. But Swype’s word gestures don’t support two-handed operation: whatever hand isn’t gesturing is just idle.
Cargo Cult Design
Again, there’s obviously room for Apple to improve – and for others to leapfrog the iPhone. But like so much in the tech industry these days, this feels like an example of cargo cult design. Innovators come up with great ideas but often don’t seem to analyze their competitors’ solutions on a detailed, task-driven level; to dig in and ask, “Why is this the way it is?” As a result, they eliminate or misinterpret areas of the product where they’re not innovating. The result is often one step forward, two steps back. And that’s a shame.
By way of disclosure: I have not tried every product Lomas brings up. This is intended as a general critique, not a comprehensive multi-product review.