Title: Athwart The Touchy-Feelies
Author: Adam Moore (LÆMEUR) <adam@laemeur.com>
Date: September 27, 2013

Athwart The Touchy-Feelies

The user-interface field seems to be gripped by a kind of mania over the paradigm of touch-screen direct manipulation. Direct manipulation, the visible interaction with graphical items displayed on-screen, has been around as a design principle for decades and manifests itself in now-ubiquitous forms such as draggable scrollbars, clickable buttons, resizable windows, drag&drop icons, &c.. The new twist is the capacitive touch-screen as the sensor device and the deprecation of other "user proxy" devices such as mice and styli.

The thinking is: all the mouse is ever used for is pointing, clicking, and dragging, and the keyboard is just an array of buttons; touch-screens allow pointing, clicking, and dragging with the finger, and we can eliminate the need for a keyboard by drawing an array of clickable buttons on the screen. Two birds with one stone!

There's an additional, more philosophical layer to this, and that is the belief amongst some designers that tactile direct manipulation with the fingers is inherently superior and more intuitive than the use of any proxy device.

A lot of people seem to have difficulty with the concept of intuition.  Clever, novel ideas that make a kind of contextual sense are not "intuitive".  Pinch-zooming is not intuitive, and neither is swipe-scrolling.  They make sense, and they seem sensible once shown to a device operator, but they are actions with no real-world analog(1), and they are not inherently superior to other means of performing the same task.  I actually prefer the scroll-wheel on a mouse over swipe-scrolling; more than that, I prefer scrolling with the keyboard and having Home and End keys to get to the top and bottom of documents immediately.  I find this highly preferable to flicking my finger for five or more seconds to scroll, scroll, scroll to either end of a document(2).

So, we'll call that my first argument against the touchy-feely people: your interfaces are no more intuitive than any other, as long as they operate on abstracted models of interaction that have no real-world analogs.  In fact, I might go a degree further and suggest that interfaces that rely on inherent interactive capabilities of objects, but have no explicit indicator that such capabilities are present, are less intuitive than interfaces that rely on visible UI signifiers of the presence of capabilities.  Pinch-zoom is less intuitive than an on-screen "zoom" widget, because until the user has been inculcated with the pinch-zoom paradigm, they've no reason to believe that any facility for zooming is available to them.

My second argument is this: if you've taken-away the hardware device, but then have to provide a modality that visually simulates it on the screen, you haven't actually obviated that device, have you?  You've only virtualized it.  So, instead of making systems simpler and more elegant by introducing new interface paradigms that replace old ones, you've actually made systems more complex by introducing new interface paradigms that partially replace the old ones, and invoke the old ones in the old ways where and when the new ones are inadequate.  Take for example the problem of text-cursor placement on iOS devices.  Because a finger is not a precision instrument and can't be relied-on to A)exactly touch the 0.5mm space between two letters, or B)not cover-up the area that you need to see to make a precise touch-contact, the user is forced to invoke a mode, a magnified area on the screen unobfuscated by the user's finger in which the user's finger movements are translated into the movements of a virtual mouse.  And, again, this is invoked by an unintuitive, invisible mechanism (i.e. placing the finger in the general-area where a precise placement needs to be made, then waiting an unspecified number of milliseconds for the magnifier/mouse sub-screen to appear).  

Now, I actually like touch-screens, and multi-touch screens are, in fact, a new interface mechanism that provide capability unavailable with traditional pointing devices.  My gripe is with the push against traditional input devices, against keyboards and mice and styli, in the naive and wrong-headed belief that the touch-screen not only obviates all prior input mechanisms, but surpasses them.

Fingers are not precision pointing devices.

Obfuscating the display is not always ideal, or even acceptable.

Less is sometimes actually less.



1. No, pushing a piece of paper on your desk is not the real-world analog of swipe-scrolling, and stretching images on Silly Putty is not the real-world analog of pinch-zoom; pinch-zoom is an extrapolation from handle-resize, and swipe-scrolling is an extrapolation of the same hand-cursor scrolling that's been around in the world of computers since the invention of the mouse.  These operations seem intuitive because we've seen them before — on computers.

2. Yes, I recognize that this is the grumping of a codger set in his ways, but the codger an important social archetype, and I take considerable pleasure in playing the part from time to time.

comments powered by Disqus