I can clarify a few things. We are using the Arden Service with OpenMRS 1.7.x. We have a new version of the parser that uses Arden2Bytecode that is much better. The old version has some problems and is not Arden compliant. I am familiar with the parser code and can answer questions about it if needed. Vibha wrote the grammar for the parser and I wrote most of the java translation to logic rules. We have the code changes for OpenMRS in an private IU repository at the moment. If you are interested, we can share it. It is branched off of OpenMRS 1.7.x.