some character sequences are more likely than others and
most simply don’t exist as legitimate words. The fundamental conceptual breakthrough to the word-gesture keyboard
paradigm is that valid letter combinations form a finite set
that can be captured in a language model, created by, for
example, mining emails, blogs, and the Web. A very simple
form of a language model is a lexicon—a list of all permissible words. In the case of English, a lexicon size of 20,000–
100,000 words would be sufficient for most users. The words
in a lexicon can be represented geometrically on a given keyboard layout as word gestures and matched against users’
input gesture. Later in this paper, we explain how to efficiently classify and recognize such gestures.
Of course, an individual user may occasionally still need to
write rare names and jargons, email addresses, or passwords
that are out of vocabulary (OOV). Since a gesture keyboard
enhances, rather than replaces, a conventional touchscreen
keyboard, OOV letter sequences can always be entered by
typing the individual letter keys. If these OOV sequences are
frequently used then they may be added to the system’s list of
recognized words, either manually or automatically.
Figure 2. Word-gesture keyboards can also work on alternative keyboard layouts. Shown here are Shape Writer atomiK mode on the iPhone (circa
2008 top left), Shape Writer on a Windows tablet with the atomiK layout (circa 2005, top right), and an illustration of the atomiK layout (bottom).