Database Reference
In-Depth Information
Chapter 12: What Paper is (and Isn't) Good For
Saying that paper prototyping is a good technique is like saying that hiking boots are a good type of
footwear—you have to specify for what purpose. Hiking boots are great for a walk in the woods, but
they'd suck for ballet. In this chapter I discuss the kinds of problems paper prototyping is likely and
unlikely to find in an interface. I also describe some issues for which usability testing in general
(whether paper or computer) is not the best means of discovery.
Dimensions of a Prototype
So exactly what is a prototype, anyway? I know that's a funny question to be posing this far through a
topic on paper prototyping, but understanding the nature of a prototype helps you decide what method
is suitable for what you need to learn. In essence, a prototype is a representation of a design concept,
but that definition is too high-level to provide much insight, so let's break it down.
Prototype Fidelity: A Misleading Concept
First, I want to describe a way of classifying prototypes that I have found to be more problematic than
useful, although I still hear it used fairly often. Years ago I was taught to think in terms of "high-fidelity"
(hi-fi) and "low-fidelity" (low-fi) prototypes. In his paper, Tullis (1990) defines fidelity [ 1 ] by how it appears
to the user, which may or may not reflect the extent of the working technology supporting it. Hi-fi
prototypes were coded to look and act a lot like the real thing, perhaps using some kind of prototyping
software. They were often good enough to be mistaken for the actual interface. Low-fi prototypes, on
the other hand, were obviously faked—no one who sees a Computer shuffling pieces of paper thinks
they're looking at software.
Although fidelity is a useful concept, it's misleading to apply it to an entire prototype because fidelity
takes what is actually several dimensions and rolls them up into one. This can lead to
confusion—person A's definition of a low-fi prototype may be significantly different than person B's
without either of them being aware of it. To avoid this confusion, I've stopped using the term fidelity
when I'm referring to an entire prototype. I use it only when I'm referring to a specific aspect of the
prototype; for example, a screen created using graphics software has a high-fidelity look. And
whenever I hear "low-fidelity prototype," I'm careful to find out what the speaker or author means by that
term.
Four Dimensions of a Prototype
So let's expand this one-dimensional view of prototyping. There are different ways to look at prototypes,
but the one I've found most useful consists of four dimensions—breadth, depth, look, and interaction (I
got this idea from the 1996 paper by Virzi, Sokolov, and Karis). As you'll see, some of these prototype
dimensions are more important than others in terms of helping you answer the questions you have
about your interface.
Breadth
Breadth [ 2 ] refers to the percentage of the product's functionality that is represented in the prototype.
Some prototypes contain all the functionality that will be present in the real design, whereas others
have only a subset. In terms of usability testing, a prototype needs sufficient breadth to cover the tasks
that have been created, but usually not much more.
Breadth is a relatively easy concept to understand, but it's not especially interesting to our discussion
because any kind of prototype can be broad or narrow. Regardless of the method you're using, there is
a fairly linear relationship between breadth and the effort required to create the prototype—if you want a
broader prototype, you'll have to do more work. Although this factor is certainly important to an
understanding of prototypes, for the most part I'm going to ignore breadth because it doesn't help us
differentiate and choose among prototyping methods.
Depth
Search WWH ::




Custom Search