For standard product granules versioning information has four types and is
reported in four places:
End of standard product filename (and the ECS Metadata object
LocalGranuleID for Level 2). Numbers capturing format changes and
reruns/reprocessing.
ECS Metadata object LocalVersionID. Has format: "Executable Name:
Clearcase Label". Clearcase is the configuration management tool used by
the TES project.
ECS Metadata object PGEVersion. TES Software Release number, e.g.
"R6.1.2".
Production History. Text block in standard product file. See the TES Data
Products Specifications document for the contents of the Production History
text block.
Versioning in Filenames
The versioning issue covers two major aspects reflecting changes in the name
of a data file:
Data file format changes.
Data file content changes (coming from reruns or reprocessing)
The following naming conventions should be used to denote changes due to
file format/content changes:
<file name>_Fff_cc...
where:
F = File Format placeholder
ff = 2-digit version number reflecting file Format changes
cc = 2-digit version number reflecting file Content changes
In creating these version numbers, the following ground rules were
observed:
The two sets of identifiers may be incremented independent of each other.
The start default value for each set is 01.
Neither set can ever revert back to 01 and must always be incremented
throughout the file span of the file.
The value of each set must always be incremented by 1.
The following are the definitions for Format and Content changes of any data
file:
Format change:
Data type changes (new or redefined)
New/deleted/renamed data fields/structures
Dimension changes of data fields/structures
The basic rule defining a Format change is any change that must be reflected
in the Data Products Specification (DPS), and therefore requires a DPS update.
These DPS updates will be disseminated via change pages and captured in
Clearcase, with complete releases of a new DPS version at certain time
intervals.
Content change:
Data file created during a PGE re-run with a different set of input files
or different environment.
Data file reprocessing where contents are updated for completeness or
better accuracy/algorithm improvement.
The purpose of the content versioning number is only to reflect a particular
run version, not the version of the TES science software used during the run.
However, the implication would be that any re-processing run of the same PGE
using a different input file set would generate product files with new
contents, thus indicated by a new value for the cc field of the products'
names.