2. INFORMATION WORLDS
2.1. The Free Rescalability of Pure Information Worlds
Using the language of classical physics, the world that surrounds us involves three
fundamental dimensions, mass (M), length (L), and time (T). A physical quantity that
involves only length and time is said to be kinematic, e.g., velocity = LT-1, or acceleration =
LT-2. A quantity that involves all three dimensions is said to be kinetic or, synonymously,
dynamic, e.g., force = MLT-2 or energy = ML2T-2.
This distinction allows us to draw a sharp contrast between the real, physical world and the
so-called electronic, or pure-information worlds that we create with modern computers.
Whereas the former is made of atoms of matter, the latter are made of bits of information
(Negroponte, 1995). The real world is dynamic in essence, whereas electronic worlds, which
have no mass component, are worlds of pure kinematics.
Galileo Galilei was the first to realize that real world objects can be rescaled only up to some
point. Taking the simple example of a wooden beam, he argued that such an object cannot
have an arbitrarily large size: as the beam is scaled up, it would gradually lose its solidity
because its resistance (a function of its cross section, dimensionally L2) would grow at a
slower rate than its weight (a function of its volume, dimensionally L3). Therefore, were the
beam scaled up, say, a thousand-fold, it would collapse under its own weight.
Galileo was considering the physical world. In an electronic world, in contrast, everything can
be freely scaled up and down, without ever experiencing Galileo’s problem. Not only can the
objects of an electronic world be of any shape and size (e.g., a 20-km tall sky-scraper), but the
user can interact at all scales with these objects. For example, the user of a zoomable
electronic atlas can perform a complete orbit of the planet in a second or so, which involves a
velocity on the order of 50,000 km/s, one sixth of the velocity of light. Since the atlas and the
virtual camera used to view it are purely informational, mass-free replicas of our planet and
an observation vehicle, the interaction suffers no upper speed limit. In fact, the only limit to
navigating such a world is the user’s ability to process the inflow of information to control the
navigation. As we will see, this has interesting implications with respect to the boundary
conditions that heavily constrain the classical version of Fitts’ pointing paradigm.
2.2. Fitts’ Target Acquisition Task: A Reduction Paradigm for
the Study of Electronic Navigation
One major difficulty for the scientific study of human navigation in information worlds is the
broad diversity of user tasks, for example, looking for a particular item, going to a particular
location, or browsing serendipitously. In this research we focus on one specific task, pointing,
or target acquisition, which was operationally defined half a century ago by Fitts (1954).
Fitts’ target-acquisition paradigm, like any research paradigm, excludes many interesting
research problems. We deliberately ignore the cognitive problems related to way-finding,
spatial orientation and the like. In this article, the target is supposed to have been defined
unequivocally from the outset and the user is assumed not only to know that the target exists,
but also to know where it is and how to get there: We are just concerned with the overt act of
reaching the target. We also assume that the computer user will work hard to comply with