Two barriers to computational reproducibility are the ability to record the critical metadata required for rerunning a computation, as well as translating the semantics of the metadata so that alternate approaches can easily be configured for verifying computational reproducibility. We are addressing this problem in the context of biomolecular NMR computational analysis by developing a series of linked ontologies which define the semantics of the various software tools used by researchers for data transformation and analysis. Building from a core ontology representing the primary observational data of NMR, the linked data approach allows for the translation of metadata in order to configure alternate software approaches for given computational tasks. In this paper we illustrate the utility of this with a small sample of the core ontology as well as tool-specific semantics for two third-party software tools. This approach to semantic mediation will help support an automated approach to validating the reliability of computation in which the same processing workflow is implemented with different software tools. In addition, the detailed semantics of both the data and the processing functionalities will provide a method for software tool classification.