Computational workflows have become essential for data processing and analysis across diverse scientific domains, including materials science. While dedicated workflow systems are gaining popularity, many tend to simplify workflow composition merely at the syntactic level, e.g. by utilizing graphical languages substituting textual code. This keynote takes a look into semantic composition support for workflows, which complements syntactic composition support by leveraging additional semantic meta-information. Such semantic support can, for instance, suggest suitable components for a workflow under development, or identify conceptual inconsistencies within it.

Based on several years of experience in developing methods and tools for user-level design of computational workflows, this talk will discuss opportunities and challenges of advancing workflow composition through semantic approaches. Examples will mainly be drawn from bioinformatics and geoinformatics, fields that have witnessed significant development of ontologies and semantically annotated tool collections in the past years. The talk will review the crucial role of domain ontologies in enhancing semantic workflows, and discuss the challenges associated with creating, maintaining, and applying domain ontologies.