When users edit large amounts of text, however, errors can be hard to spot. Outlier finding is a way to reduce errors by drawing the user's attention to inconsistencies. When integrated into the multiple-selection editor and tested in a user study, outlier finding reduced errors.

Both selection inference and outlier finding depend on lightweight structure, a new way to represent and manipulate text structure. Lightweight structure is to pattern matching what procedural abstraction is to programming --- enabling the construction of a library of reusable, composable patterns and parsers. Theoretical contributions include a formal definition of the region algebra that underlies lightweight structure, the data structures and algorithms that implement it efficiently, and a characterization of the classes of languages recognized by region algebra expressions.

These ideas are implemented in LAPIS, a freely-available, open-source text editor/web browser written in Java http://www.cs.cmu.edu/~rcm/lapis.

<< Back