The Collaborative Computing exemplar enables computational tools to be more proactive in performing analytic tasks of potential benefit to an analyst’s workflow. Traditionally, the initiative to execute a computational analytic task (such as a query or a mathematical computation) comes from the analyst. The analyst determines and defines the task to be performed, and then the computational system performs the task and returns the results. This analyst-initiated workflow can create frustration (if the analyst does not know how to define the requested task) and friction (if the analyst or computational system must wait for the other to perform their task).
Our vision is to enable a model of analyst augmentation where the system proactively identifies beneficial computational tasks, performs them, and seamlessly integrates the results into the analyst’s workflow. These computational tasks can be both assistive/amplifying in nature, where the task is one that the analyst would have asked for as part of their current workflow, or broadening/complementary in nature, where the task is one that may benefit the analyst’s workflow but not one they would have thought or known how to ask for.
The Collaborative Computing exemplar develops fundamental understanding of and explore novel approaches to what we see as three core pillars of collaborative computing:
- Instrumentation and Observation – Understanding and capturing what the analyst is doing with minimal interference to the analyst
- Tasking and Execution – Using the instrumentation information to determine which computational tasks should be performed and executing them
- User Experience – Integrating the computed information and products into an analyst’s workspace without unintentionally destroying an analyst’s natural workflow
The research efforts in these areas are grounded by data capturing existing and potential analytic tasks. Our preference is that this data be generated through the instrumentation of tools and prototypes that analysts can use to provide real mission value. Therefore, a significant part of the Collaborative Computing exemplar is the design and development of the instrumentation approach to generating, collecting, and structuring analytic workflow data for further analysis. Of additional interest is the development and implementation of workflow recommenders.
This exemplar also develops user-facing, value-adding, collaborative computing prototypes based on the understanding gained by analyzing workflow data. The main objective of these prototype development efforts is to quickly iterate capabilities that provide mission analysts real value through collaborative computing. These prototypes will also be instrumented so that as their value is demonstrated through their increased usage, they will capture more data about analytic tasks and workflows.