No Jaggies! The classic SIGGRAPH t-shirt design superimposing a staircase and the international NO design (red circle and slash) eloquently speaks to improving the quality of computer-generated graphics. The jaggies introduce false messages, especially easy-to-notice corners or staircases in supposedly smooth and straight lines and edges.

There are two major solution strategies: increased resolution and smoothing techniques. Increasing resolution involves changing devices, which argues both for device-independent representation of images and for a systems approach to permit access to a variety of devices. With relatively coarse-resolution display screens, anti-aliasing techniques to eliminate the jaggies have received considerable attention in computer graphics research. With laser printers that are capable of binary-valued image pixels and posses only moderate resolutions, the difficulties are considerable. The Warnock and Wyatt paper on Device Independent Graphics Imaging Model is fundamental to the page description languages that address this issue.

High quality printing increases the expectation level for typographic (or `typeset') quality. Reproducing typefaces are a crucial quality issue. Chuck Bigelow's following paper on the Design of Lucida discusses how that typeface design attacks the quality issue.

A common standard of graphic arts quality is the use of typographic fonts. The choice of type families is often subjective and driven by faddish concerns. Three type family names have become quite ubiquitous in documentation graphics: Helvetica, Times Roman, and Computer Modern. However, the quality of reproduction is sometimes so poor that the resemblence of the type to those families is in name only!

Type fonts stress the reproduction quality of systems because they contain so much visual information in such a small space. The rythym of the strokes, the swell of the serifs, the contrast between thick and thins, and the `color' of the resulting page are all subjective measures with which we subconsciously develop experience during our lifetime of reading. Attempting to match these typographic expectations on moderate-resolution devices is a significant challenge.

A fascinating distinction between low-quality images and high-quality one is often the variation of line weights. Uniform weights are a signal that little effort was made to distinguish the importance of information. (Of course, uniformity of line weight here does not refer to the poor drawing capabilities of pen plotters with nearly exhausted pens.) Specifying thin lines for dimension lines, axes of graphs, or highlight lines is a common method for demphasizing them. Using thick lines for important objects or outlining the shape or curve with a thick line helps direct the viewer's attention.

Varying the line weight introduces difficulties in properly drawing lines that intersect, and in finishing the ends and joints of line segments. The page description languages discussed later both provide extensive controls for specifying these attributes.

The contrast and dynamic range of an image indicates the quality of the reproduction process. The blackness of the blacks, the whiteness of the whites, and the distribution of the grays in between are an issue here. Copy-quality problems with some output devices reduce the achievable contrast. High-resolution devices that output to high-contrast photographic media are designed to avoid these problems. Such devices are typically graphic arts image setters or process cameras.

Digital images that contain scanned or sampled data must be produced with concern for the tone values that result from halftoning or dithering algorithms. Compensation techniques to ensure a linear response are necessary to avoid compressing all the information into the white or black region of the tone reproduction curve.

The use of color introduces many significant challenges. The perceptual impact of color is well-known, but the control over that impact is not fully understood or practiced. It is impossible to explain the garish choice of fully-saturated primary and secondary colors on many colored samples produced by manufacturers of color devices. Well-coordinated use of color requires subtle control over the color imaging process. Faithful reproduction of color across different devices and media requires continued research as indicated by Maureen Stone's paper Color, Graphic Design, and Computer Systems, later in these notes.

Visually interesting images are a challenge. Creative people can produce interesting images in any media, including those produced with computer-based tools. However, such automated tools lend themselves to producing uninteresting images faster and with more regularity! The tools do not increase the creative content, although they may liberate creative people to attempt more creative images.

Illustrations prepared for graphic arts quality publications are often edited and managed to achieve a certain appearance. This control and discipline must be considered in using documentation graphics tools. Harmonizing the choice of typefaces, the selection of line weights, the variation in colors, and the range of interesting shapes is a necessary supervisory task. Graphic designers receive professional training in this discipline. When amateurs attempt similar illustrations, they should rely on the professionals for advice or use tools that incorporate the disciplines of the professional.