Type design with flexible constrained font characteristics

Modern computer output devices work with finite resolutions to display contours of images and artefacts, including those used to model character outlines. The modelled image itself is frequently achieved by means of a spline description, connected in a piecewise manner. The goal of a rasterizer is then to translate the modelled, spline, form into a corresponding discrete version, which maintains the originality of the font and is aesthetically acceptable. The process of converting between forms has certain problems, generally, and has particular concerns when attempting to represent continuous outlines belonging to a character onto a mesh of display points. The main consequence of this usually is that, at best, the resulting character loses some of its unique features and, at worse, is illegible. To overcome such deficiencies, the rasterising process makes use of additional intelligence about a character (such as its characteristic points) and the parts (stems, serifs, bows, diagonals etc) that are involved in its description. These embodied ingredients provide a somewhat more pleasing outcome in practice. This paper addresses the concerns of incorporating intelligence or hints within a font outline by making use of geometrical constraints. These are used to provide a further insight about the properties and relationships of components employed for a character description. The work reported here provides a more flexible environment than has otherwise been gained. The usefulness of the approach is shown by evaluating its application to a Times Roman half-serif, modelled by Bézier cubic splines.