While traditional workflow management systems have proven useful in supporting well-defined organizational processes, they are less suitable for the support of less well-defined, emergent processes. These kinds of processes are better supported by systems, which are based on the notion of workspaces. The research prototype LiveNet, which allows workspaces to be defined and supports their dynamic modification and evolution, has been developed and successfully applied in a number of domains. Based on the experiences with LiveNet, current research is focused on analyzing the structures supporting cooperation defined by workspace system users. Such an analysis can reveal structural deficiencies, which can be brought to the users' attention. It can also aid the accumulation of a library of successful structures for later reuse. Finally, it can aid in extracting personal styles or preferences from observed structures, which can be applied in creating new ones.