This is the starting page for the KPML-compatible resources for generating German

German was the second language to be addressed in considerable detail within the Penman-style of computational systemic-functional grammar after the original English grammar. The grammar was originally developed as part of the KOMET natural language generation project at the then GMD institute IPSI in Darmstadt. Following on extensive development of English within the Penman project and the more experimental developments for Japanese undertaken at Kyoto (cf. Matthiessen/Bateman, 1991), the KOMET project required extensive German generation capabilities.

The grammar took its original design from Erich Steiner's work within and around the Eurotra project and incorporated several aspects of Robin Fawcett's brand of systemic theorizing. Development of the grammar was mostly carried out by Elke Teich and formed the basis of her book length description of certain aspects of the developed grammar. Subsequently the grammar was partly merged with a parallel multilingual generation project, the TechDoc system of Dietmar Rösner. For this subsequent phase of development, the main developers were Brigitte Grote and Elke Teich. The final projects that the German grammar in this form were used in were the multilingual biography experiments of KOMET and the multimodal spoken language information retrieval information developed in the Speak! project. After this, around 1998, the grammar was not used for some time and fell into a certain degree of disrepair as bit rot set in. Meanwhile, the KPML-development continued and other languages were at the focus of attention. Maintainance and extensions for German were attempted by Renate Henschel (1999 in Edinburgh) and most recently by Serge Sharoff (2001/2 in Bielefeld).

The last generated set of examples from that earlier grammar version can be seen here (generated 1999).

We are now (January, 2003) beginning to fully update the German grammar, extending its coverage as convenient, for some newly started projects. The current state of this work will be reported on here, and current versions of the grammar can be downloaded from the respective example pages.

Apart from a general dust-off and re-checking of example sets, the most important updates are/will be the following:

The first example pages available are reached here (January 12th, 2003).

The system network, organized by regions, can now be inspected from the Region Overview Report (December 5th, 2004).