Skip to main content Skip to main navigation

Project | SKATE-W

Duration:

SAP Knowledge Authoring and Text Enhancement

Ensuring consistency and quality in large-scale, multilingual, distributed documentation infrastructures is a complex task - available software applications are not able to guarantee the consistent use of terminology and style between authors. Building on the successful BMBF-funded FLAG project, SKATE - a development project sponsored by SAP - aims to develop tools and linguistic resources to support a consistent style of authoring and translating documentation within SAP. SAP has one of the most sophisticated and successful automated document authoring and localisation infrastructures of any industry and the SKATE software is intended for use by technical writers, translators and copy-editors. Support is currently only available for German and English, but further languages are planned for the future. SKATE is tightly integrated with SAP's current standard offline editor, Microsoft Word, but other editors are also planned. The main emphasis of the project for SAP was to provide an environment for standardised style checking. The FLAG project delivered an environment for developing language checking components, this is an open system designed to be used by organisations who wish to integrate their own language standards and guidelines. In SKATE, the style rules specific to SAP are developed in house at SAP, whilst the DFKI is developing more generic grammar checking rules. Once implemented, these rules are available as modular checking components. During the project, SAP put particular emphasis on usability and user interface issues to ensure acceptance by users. As well as highlighting problems for authors, the system provides simple access (on a right-click of the mouse) to extensive customisable help and guidelines to help improve the quality of the text. The secondary aim of the project was to aid the consistent use of standard terminology. Existing SAP terminological resources were converted and hyperlinking technology was used to link non-approved terms with their approved variants. The system supports a modular terminological resource organisation, where up to 10 different termbanks can be accessed.

  • An integrated terminology and language checking application
  • Creation of language resources for ensuring consistent documentation
  • Flexible client/server architecture
  • Integration with standard authoring tools

Publications about the project

  1. Text Generation in the SmartWeb Multimodal Dialogue System

    Ralf Engel; Daniel Sonntag

    In: J. Hertzberg; M. Beetz; R. Englert (Hrsg.). KI 2007: Advances in Artificial Intelligence. 30th Annual German Conference on Artificial Intelligence (KI 2007), September 10-13, Osnabrück, Germany. German Conference on Artificial Intelligence (KI), Pages 448-451, LNCS, Vol. 4667, Springer, 9/2007.