The forms of visual language are woven from many threads. Traditionally, the tools and materials used in making these forms have been an intimate part of their structure, and this is true for digital typography as well. But, it is also true that the forms of digital type are now generally represented by mathematically described outlines and, of course ultimately, by a sequence of zeroes and ones. These outline representations were christened “device independent” by Adobe Systems in the mid 1980's. The forms exist in an abstract space which seems quite separate from their physical instantiations.

Conceptual structures for letterforms have a history. The formal structure of stroke-based written characters, known as a ductus, is often decipherable from the forms themselves, and in sixteenth century Europe the order and direction of strokes are explicitly spelled out by writing masters in their manuals. Such systems have another conceptual level which has both functional and aesthetic components, namely visual integration.

The tendency toward visual integration is a common theme in writing systems. It is evident in the development of Sumerian cuneiform over its life span of more than 3000 years, and in the evolution of cursive writing in Egypt, a system with even greater longevity. Indeed it is the basis of the entire lineage of stroke-based scripts which lead to the Roman alphabet. The repertoire of strokes and their composition seems to magically result in a visually integrated system, but the system itself is the result of a design process which necessarily must involve both conceptual thinking and craft. It is also inevitably part of an even more complex structure—a writing pedagogy.

Typefaces are another matter. The structure of calligraphic forms which have been translated into type is frequently possible to understand if the typographic versions are an attempt to faithfully render the written ones. The early roman types did not strictly follow this practice, and neither have their progeny.

The internal structures of roman type designs can, therefore, be considerably more difficult to comprehend than their written predecessors. The visual integration that is achieved in roman type design is the result of an individual's ability to make a coherent fabric from a complex collection of fibers. The system, if there is one, is private. Unlike formal writing, a type design is not expected to be part of a tradition which carries with it instructions for its reproduction. There are few first hand accounts by type designers about what they were up to. To quote Harry Carter, "In the darkness of typo-graphical history even a feeble light...is a great help."

There are passages from the manuals of Pierre Simon Fournier and Giambattista Bodoni which cast light on the concepts which influenced the graphic structure of their letterforms. Fequently, however, the underlying structure of typographic letterforms is undocumented. One must attempt to infer the mental patterns of type designers from the letters they designed and from what we can discover about their historical and cultural contex.

Many attempts have been made to concoct geometric constructions for Roman capital letterforms. One of the appealing things about these schemes is that they offer an explicit, systematic set of procedures for arriving at the forms. Unfortunately, none of these has produced letters that are credible copies of those in the most refined Imperial Roman inscriptions.