It would be interesting to review/extend the standard commenting operation in regard of QC/PP tests over the history field in the NetCDF exported files. This would help, for example, on external reporting scripts (or any other external tool) that need to inspect what has been done within a particular file.
Ideally, the history would provide complete string representations of what has been done, where applicable. Optionally, the history may even have reversible/parsable string representations of the operations performed.
Complete means that all arguments/options used are explicitly stated/referenced. This is mostly the case, but maybe missing some information that is contained in variable comments only.
Reversible means that the history comment can be read, processed, parsed and converted to be re-used to obtain the same results. The NetCDF history field is not ideal for that, but a lot of conventions used some kind of string introspection for conversions anyway (e.g. mapping transformations). This kind of functionality may help in reprocessing from the file itself (no DB, no previous session files).
A deeper change in formatting, like timestamps on each operation or toolbox transition information (loading -> PP -> Metadata edits -> QCs -> export) would be good candidates too. Maybe we need a more powerful and structured log?
It would be interesting to review/extend the standard commenting operation in regard of QC/PP tests over the
historyfield in the NetCDF exported files. This would help, for example, on external reporting scripts (or any other external tool) that need to inspect what has been done within a particular file.Ideally, the
historywould provide complete string representations of what has been done, where applicable. Optionally, thehistorymay even have reversible/parsable string representations of the operations performed.Complete means that all arguments/options used are explicitly stated/referenced. This is mostly the case, but maybe missing some information that is contained in variable comments only.
Reversible means that the
historycomment can be read, processed, parsed and converted to be re-used to obtain the same results. The NetCDF history field is not ideal for that, but a lot of conventions used some kind of string introspection for conversions anyway (e.g. mapping transformations). This kind of functionality may help in reprocessing from the file itself (no DB, no previous session files).A deeper change in formatting, like timestamps on each operation or toolbox transition information (loading -> PP -> Metadata edits -> QCs -> export) would be good candidates too. Maybe we need a more powerful and structured log?