A Study on Modeling of a Lightweight Scientific Workflow System Using the XML Schema

Modeling of the processes of scientific research activities is a fundamental problem that confronts scientific-workflow applications. This paper analyzes the internal structure of a data-oriented scientific workflow in both user operation form and computer processing aimed toward the Chinese Virtual Observatory. Based on the analysis modelings using the pipe technique and XML are studyed separately. To overcome the flaws of the pipe-based modeling, we propose a Scientific Workflow Description Language (SPDL) based on the XML schema. The SPDL has characteristics of high-level abstraction and high representing capability to support a lightweight scientific workflow system. Results from analysis and experiments show that the language can describe coarse-grained scientific research activities and satisfy the requirement of dataflow-oriented modeling.