LingURed: Language-aware editing functions based on NLP resources

In this paper we compare the state of the art of language awareness in source code editors and word processors. Language awareness refers to functions operating on the elements and structures of a formal or natural language. Language-aware functions allow users to work with meaningful units, increasing efficiency and reducing errors. While all modern source code editors provide programmers with language-aware functions, similar functions for natural-language editing are almost nonex-istent. Writers have to manipulate characters, which makes editing and revising challenging and results in typical errors. We describe the LingURed project, in which we implement language-aware editing functions for German with the goal of supporting experienced writers. Our approach is based on the combination of standard editor functionality and shallow localized natural language processing. Prototypical functions demonstrate the feasibility of the approach. Based on our preliminary experience we discuss requirements for NLP components suitable for use in interactive editing environments.