Classification of persistent system state
A table for guiding design decisions about where various kinds of information stored by Kepler should be located and organized.
Data type |
Deletable automatically? |
Accessible across multiple installations? |
Visible outside Kepler? |
Comments |
---|---|---|---|---|
Discovered and downloaded data |
Yes |
Yes |
No | Data sets downloaded through the Kepler data panel. Only data that can reliably be downloaded again as needed. |
Compiled KAR files |
Yes | OK | No | |
Library serializations |
Yes | OK (Could be hard?) |
No | |
Default configuration |
No | Yes, via shared modules |
No | Menu entries, widgets, default ontologies, etc. |
Scientists' customizations |
No. (Annoying) | No | No | Persisted preferences, e.g., what modules are active, what ontologies are enabled, etc. |
Scientific project data |
No. (Disaster) | Yes! | Yes! | Files input to or created by workflows, etc. |
Reusable intermediate data products |
No | Yes! | No | Data that could be reused automatically in future runs to optimize performance. |
Add-on modules | No | Yes |
No |
Modules not included in the base installation. |
Provenance and run logs |
No. (Disaster) |
Yes |
No | |
Scientist-applied annotations and tags |
No. (Disaster) |
No |
No | |
Actors scientists save to the library |
No. (Disaster) |
Yes |
No |