What factors accounted for the shift from
random-scan displays to raster-scan displays?
In the 60's and 70's computer memory was relatively
expensive. Random-scan displays
needed only to maintain a display list in memory. This display
list was usually short enough to fit in the small memories
available to most computers then. On the other hand,
to maintain a frame buffer, a black
and white raster-scan CRT had to store bits.
Such a relatively large amount of memory was very expensive.
As memory prices dropped in the 80's, the extra cost of
a frame buffer became negligible.
The primary drawback of random-scan displays is their inability to draw 2D filled regions. For example, creating a photorealistic picture of almost any everyday scene is impossible using only line seqments. A wire-frame rendering of an object can be used in some applications such as architecture. However, most graphics applications require the full power of a raster-scan display. The focus of this course is on the problem of converting a continuous mathematical model into a set of intensities associated with a discrete set of pixels.