Hobgoblins and Small Minds | Cypress Semiconductor
Hobgoblins and Small Minds
There is a famous misquote, which is "Consistency is the hobgoblin of small minds". This is a misstatement of Ralph Waldo Emerson's statement: "A foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines." I am not going to pretend I fully understand Emerson's point, but I will emphatically state that he DOES NOT mean that consistency is bad, but "foolish" or inappropriate consistency is. Another writer, Oscar Wilde had a different take: "Consistency is the last refuge of the unimaginative." When it comes to a design, "A foolish imagination is the last stop of a soon-to-be unemployed designer" (original quote of Jon D. Pearson).
Consistency versus imagination - the two are neither at odds nor opposite ends of the spectrum. Take for instance how a button works. If you were to look at a pushbutton on a device, could you tell its function by sight alone (all you can see is a button)? Probably not. Now, if I let you you play with the button and view what happens, and then asked you what its function is, do you think you could tell me? Only if it behaves consistently. Now, what if it is a really low-quality button, the contacts are very bouncy, the button has a lot of play in it and the device it is connected to is not debouncing the switch? As long as the function is consistent (when pressed the light turns on and when released it turns off, or pressing toggles the light on and pressing when light is on toggles light to off) you would have no problem to "specify" the behavior or to use this switch (although you may have to practice patience).
Now, imagine the switch is very high quality and the device is sensing it very precisely, but there are hundreds of combinations of current status (button pressure, time-pressed, orientation of button) and previous state (time in state, pressure/time/orientation when entered or exited). How long would it take to exhaust all possibilities and experience them frequently enough to "learn" all the behavioral nuances to be able to specify the switch "behavior"? Answer is somewhere between now and eternity (or until you give up and go do something else).
Now try one more thought experiment: imagine you are creating a new home entertainment device, and the only switch and behavior you could make use of was a momentary pushbutton. Why? Because that is what you used before and therefore will always use, exclusively. Could you design a product with this constraint? Sure. Would it be satisfying enough to sell well? Depends upon the device and use-case complexity, but likely this artificial constraint would hamper the design and therefore the customer satisfaction and in turn sales.
So which is the right way? My suggestion is to always consider the user - if a user is better served by an imaginative re-definition, then go for it. If users are very comfortable with doing things a certain way, take great care in changing this behavior (and even greater care if improper use could be dangerous). But in every case, provide consistency at least in the local sense of being understandable and repeatable. Humans are very good at learning and training themselves to adapt to a poor design that always acts the same way. They have much less patience for an elegant design they cannot remember or understand how to use.
BTW, in preparation for this article I learned a hobgoblin is not always been a bad thing either, just like consistency. See http://en.wikipedia.org/wiki/Hobgoblin if you are interested.