- current examples
- are any innovations needed?
- Seurat compatibility
- Internal interoperability assurances
- Concept of separability of data object and analysis object
- rhdf5 can be used with socket-level parallelism, how about forking?
- HDF5 local and remote
- Spark and other distributed data architectures
- Parquet, Arrow
- accommodating the HCA schema/terminology beyond files to features, cells, donors
- Does SingleCellExperiment need additional infrastructure to increase precision of provenance
- Integrating elasticsearch functionality for R users
- CWL/WDL