Graphics Reference
In-Depth Information
that happen, by 2000 the situation was entirely reversed: The user had become the
precious resource while the computer was a relatively low-cost item. The UI is
the place where user time is consumed, even in large and slow-running programs:
Once the user sets the program running, he or she can do other things. Hence, we
should concentrate more and more effort on interfaces and interaction.
1
0
-1
-2
-3
What sorts of issues affect UI design? Many of them are related to psychol-
ogy, perception, and the general area called human factors. It's one thing to use
color in your UI; it's another to make sure the UI works for color-deficient users
as well. It's one thing to have all necessary menu items present; it's another to
order and group them so that a typical user can find what s/he is looking for
quickly and select it easily: The menu items must be organized, and each item
must be large enough to make the selection process easy. And it's still another
thing to be certain that your UI is appropriate for whatever kind of device you
might be using: a desktop machine, a smartphone, a PDA, or a video game
controller.
Despite the importance of interfaces, we will not discuss them much; UI
research is now its own field, related to graphics but no longer a part of it. In some
cases, there are interface elements for which those with experience in graphics can
offer particular insight. Chapter 21 discusses some of these as applications of the
modeling and transformation technology developed earlier in the topic.
From this discussion, it's clear that the goals of computer graphics are not
purely based on physics or algorithms, but they depend critically on human beings.
We don't merely compute the transfer of light energy in a scene; we must also con-
sider the human perception of the results: Was the extra computation time used in
a way that mattered to the viewer? We don't merely create an application program
that provides functionality and performance that are appropriate for the some par-
ticular endeavor (e.g., playing music from a library or helping a physician main-
tain notes on patients); we also concern ourselves with whether the interface the
program presents makes the program easy to use. Ease of use is obviously closely
tied to human perception. We therefore present an introduction to perception in
Chapter 5.
1960
2010
Year
Figure 1.6: The log of the ratio
between the cost of a computer
and the salary of a person using
the computer (roughly amortized
for
multiuser
systems),
plotted
against the year.
Graphics research has followed a goal-directed path, but one in which the goal
has continued to shift; the first researchers worked in a context of limited pro-
cessor power, and thus they frequently made choices that got results as quickly
and easily as possible. Early efforts were divided between trying to make draw-
ings (e.g., blueprints) and trying to make pictures (e.g., photorealistic images). In
each case, many assumptions were made, usually in concession to available pro-
cessor power and display technologies. When a single display cost as much or
more than an engineer's salary, every picture displayed had to have some value.
When displaying a few hundred polygons took minutes, approximating curved
surfaces with relatively few polygons made a lot of sense. And when processor
speeds were measured in MIPS (millions of instructions per second) but images
contained 250,000 or 500,000 pixels, one could not afford to perform a lot of
computations per pixel. (In the 1960s and early 1970s, many institutions had at
most a single graphical display!) Typical simplifying assumptions were that all
objects reflected light more or less as flat latex paint does (although some more-
sophisticated reflectance models were used in a few systems), that light either
illuminated a surface directly or bounced around in the scene so often that it
