Chapter 20. Real-time Processing of Cursive Writing and Sketched Graphics
Arnold J.W.M. Thomassen, Hans-Leo H. M. Teulings · 1988
The advances that have recently been made with respect to intelligent worksta tions and software also Involve highly sophisticated recognition algorithms. The latter open up attractive possibilities for accessing the computer by means of the Unguistlc communication modes of speech and writing. Yet the large-scale introduction of keyboard-and-screen text editors with the many human-machine interaction problems associated with that revolution, and the spectacular potential of speech processing soliciting huge and lasting research investments in that area, seem to be responsible for the relative oblivion of the study of handwriting and drawing as efficient modes of human-computer interaction. The present contribution intends to point out the attractiveness and feasibility of using pen and paper as a natural communication device in an office work environment. Admittedly, as compared with the keyboard, handwriting does have some seri ous drawbacks due to its relative slowness, attention demands, and fatigue. These disadvantages are more prohibiting the longer the text involved and the more routine the writing task. However, for shorter notes, for editing com ments, and for the composition of nonroutine texts, where writing speed is secondary to careful formulation and accurate positioning, handwriting may be the most efficient mode. This is also the case, of course, for filling out forms, addressing mail, writing out cheques, and other writing tasks which are difficult to achieve with typewriters and text editors. Moreover, the natural use of the pen can be most effective in situations where various marks and symbols are used in combination with text, with editing comments in existing documents, both handwritten and from line printers, and with graphics, as we shall see in the present paper. Neglecting in our research and development the use of the pen as a processing mode in the modern office would therefore be a denial of the possibility of a universal communication facility that is in fact within reach. It would, moreover, overlook the enormous economic interests involved in satisfactory solutions for the automatic processing of the large variety of handwritten documents in our society. Finally, it would be a denial of the contributions made by several promising research projects dur ing the past 25 years. In the paragraphs to follow, we shall refer to some of this earlier work as well as to some recent research including that at our laboratory.