Andrew,
I addressed the scope and tried to "beef it up" to make it a little clearer all of the things that we have discussed. Everyone, please update as you see fit:
-
Standard descriptions and definitions of scientific data.
-
Processing and management of 1D, 2D and 3D data including both structured and unstructured grids.
-
Plotting and visualizations of data in 1D, 2D and 3D.
-
Workflow algorithms and their visualization.
-
Machine learning, artificial intelligence and data mining.
-
Modeling and simulation projects related to the physical sciences, including but not limited to physics, chemistry, biology and geology.
-
Modeling and simulation projects related to the social sciences, including but not limited to sociology and psychology.
-
Applied mathematics projects such as common math libraries and mesh management tools and with the exception of cryptography.
-
Infrastructure to support scientific computing including those tools found in the Eclipse Parallel Tools Platform including job launch and monitoring, parallel debugging and remote project management for scientific computing.
I wrote bullet 9 as a starting point for the infrastructure/PTP scope. I wanted to make sure that we consider structured and unstructured grids for the data management. I also wanted to expand the machine learning section to include AI and data mining, which are the first things that a lot of data scientists will ask about if they don't see them listed. I didn't want to exclude agent-based simulations, such as Eclipse AMP (
http://eclipse.org/amp/), which can be used for simulations in the social sciences, so I included them in addition to the blurb on the physical sciences. Likewise, I didn't want to leave out applied mathematics because if we start building a big capability we will not doubt be writing our own math library or grid/mesh management tools. I left cryptography out because cryptographic routines are export controlled.
I hope this doesn't bother anyone. Like I said, please edit as you see fit!
Jay