Nonetheless, the problem of letter-identity has received attention from typographers as well as ergonomists. The designer Eric Gill, for instance, decided to draw a number of different, simplistic versions of the letter ‘a’ and compare them to each other. However, this kind of exploration does not solve a basic problem; we still do not know which are the determinant features. After all, people might agree which of the characters Gill has drawn are acceptable as a letter and which are not, but there is no analysis of how people arrived at such a conclusion. In an experiment, Ovink (1938) has compared sets of typefaces per letter of the alphabet in order to pinpoint the typical parts that made people identify a character. This kind of research resulted in conclusions about how near to typical a particular typeface is for one particular letter only. This knowledge may be useful for typeface designers, but does not contribute much to the general question about the determinant features. Despite the fact that we can name all the features of a character (figure 1) and use some of these features as a basis of typeface classification (table 1), we are still looking for the truly distinctive features of type form that prompt us to make the ‘a’-sound when seeing the ‘a’- form. Of course, the real challenge for the type designer lies in shaping a character into a typical member of its typeface-alphabet. In other words, the character’s form not only has to approximate the archetypal ‘a’, but should also conform to the general typeface form, which it shares with the ‘d’, ‘z’, and all the other letters in the same family.