The Defense Advanced Research Projects Agency (DARPA) sought a common data language to support their Deep Green research program - a program which aims to develop a simulation-based Course of Action (COA) analysis tool for brigade commanders. Given a limited set of available languages, Military Scenario Definition Language (MSDL) was proposed as a logical choice for the program. However, MSDL had never before been implemented to run fully automated simulation software. Thus, Deep Green stakeholders were both optimistic and apprehensive about developing software around a prototype data language.

Despite the fact that MSDL was on the verge of standardization, it quickly became evident that the existing version 1.0 was not adequate for representing the brigade-level courses of action (COAs) that were at the heart of Deep Green. In order to make MSDL viable, a new version (designated MSDL-DG) came into being. This new MSDL variant took months to define and even longer to implement.

This paper gives an introduction to the Deep Green program, and describes some of the problems encountered when encoding brigade-level COAs into MSDL. Included are some of the encoding conventions that were adopted, MSDL schema modifications that were implemented, innovative tools and processes used in the encoding process, and insights offered for future data representations of COAs - with a focus on Battle Management Language (BML).