Languages

The Structured Metrics Meta-Model (SMM) is designed to specify measures, apply measures to models, and process and collect measurements. These three areas are represented in measure libraries, observations and measurements in the SMM. However, it is complicate to build SMM models by hand and the distinction between the definition of measures, the application of measures and the evaluation of measures is not that strict in SMM. Therefore, we designed two languages (so far) to address these fields. First, the Measure Definition Language (MDL) is used to construct measure libraries targeted for specific meta-models (e.g., KDM). And second, the Measure Query Language (MQL) us used to describe the application or configuration of measures and the evaluation of measurement results.

Basic Modeling Scenario

The MAMBA workflow takes SMM models and models under observation and processes them together with external measurement data into a resulting SMM model. This process can be seen in the following figure. As SMM models comprising a measure library and the application of measures are hard to develop by hand, they can be generated out of MDL and MQL model files.

These MDL and MQL models are used to generate parts of an SMM model. As shown in the following figure, the MDL is used to specify measures in a measure library, while MQL is used to prepare an observation and to steer the evaluation. In addition the MQL is also able to create derived measures and add them to a measure library. However, this is not shown in the figure, as it would lead to a less comprehensible illustration.

The figure shows on the left the meta-model of SMM and on the right in blue a meta-model of a model to be observed. In that particular case, it is the meta-model of the quickstart example. In general any meta-model could play the role of the system meta-model from the figure. In conjunction with a DSL the meta-model of that DSL play that role, or when software is represented, by  Knowledge Discovery Metamodel (KDM).