Modeling of Problem Solving Processes
The goal of this effort is to study and comprehend the typical problem solving processes of different scientific domains in order to identify general workflow requirements and potential target areas for the development of collaborative environments. An auxiliary goal was to gather user requirements for a potential workflow management tool. Workflow analysis sessions were performed with representatives of five scientific domains with the objective of understanding the common types of processing / tasks performed across domains and where we might develop general capabilities that can be used in many domains. Issues explored in each domain include the processes users carry out, tasks involved in the processes, decision and collaboration points, dependencies, and linkages to external resources. Analysis of collected workflow data is continuing, but has already yielded a wealth of valuable information. At the conceptual level, discussions about the circumstances under which researchers consult with local colleagues has led to the use case of collaboration as a means of reviewing the exceptional case, not as standard procedure. This suggests that interfaces to collaborative environments that require any decisions about collaboration (will a process be done collaboratively, with whom, what information will be shared, etc.) at the start of a process may be less effective than those that provide a means to support a collaborative review of the decisions and results made by an individual or subgroup. Information at the level of the specific data acquisition, analysis, modeling, and visualization codes, data formats, and procedures used in these domains was also recorded. From the general workflow data and these specifics of current practice, we are identifying potential design and technology options for the design of future collaborative environments. A full description of this work can be found in Participatory Workflow Analysis: Unveiling Scientific Research Processes with Scientists.

