Graphics Reference
In-Depth Information
was most effective. The 8 corners, 12 edges, and 6 faces of the cube each cor-
respond to specific views. A user can click on zones near any of those with a
mouse to warp to the predefined viewpoints relative to the center of the cube,
or click and drag to rotate the cube to an arbitrary orientation (in the style of the
arcball [Sho92]). The outlines of the cube are stroked as solid paths when the cube
is at one of the 26 canonical views and dashed for intermediate views. In addition,
small arrows (not shown in Figure 21.11) point to the four peripheral faces (which
may not be visible) and support 90
◦
roll rotations in the plane of the current view.
When you are in free-transform mode in Photoshop and you select an image, its
bounding box is shown with small square “handles” at the corners and edges. As
the cursor moves over these handles, it changes to a double-headed arrow, disclos-
ing that you can click and drag the handles. Corner drags reshape the bounding
box (and its content) in both
x
and
y
; a shift-key modifier makes the changes in
width and height be proportional. A control-key modifier lets the corner (or edge)
be moved to any position, so the image is no longer rectangular. Edge drags move
the selected edge; a shift modifier makes the opposite edge adjust as well so that
a shift edge drag on the top scales the image around its horizontal centerline; a
control-shift modifier lets the user shear the image (i.e., move the edge center
along the line containing the edge).
When the cursor is slightly outside the bounding box, it becomes a curved
double-headed arrow, indicating that you can rotate the box and its contents (see
Figure 21.12). Finally, if you click on a corner and press appropriate modifier
keys, you can apply a perspective transform to “keystone” the bounding box in
either the horizontal or vertical direction, giving the appearance of perspective
(see Figure 21.13).
Figure 21.12: Rotating an image
in free-transform mode.
Chateau [IH01] is a system for rapidly creating highly symmetric forms. User
input is processed to search for symmetries. For instance, if the user recently cre-
ated a cylinder of length 5 and radius 1, and begins the gestures to create a new
cylinder, indicating a length of approximately 5, the system offers up a completed
cylinder in a thumbnail view, which the user can click to confirm that it's what's
wanted. If there are multiple possible completions, the most likely (according to
some heuristics) are offered. Once the second cylinder is placed somewhere, the
system may propose a third cylinder, offset from the second in the same way the
second is offset from the first, thus making it easy to create a row of columns, for
instance.
While the particulars of this program are not especially relevant, the notion of a
suggestive interface,
in which candidate completions of actions are offered, lever-
ages the “recognition is faster than recall” idea: The user can recognize the correct
completion rapidly. Similar ideas are used in keyboard input for Asian character
sets, where each character is represented by a quadruple of ASCII characters,
but once the user types one or two ASCII characters, several “likely” choices are
offered as completions, with likelihood being determined by things like recent use
in the document, or even surrounding vocabulary or sentence structure.
Figure 21.13: The image has
been keystoned by dragging
a corner along the left side;
the bounding box remains un-
changed, however.
