The CDF file contents are fully defined in the document `Reference Document for CSDS CDF Implementation', DS-QMW-TN-0003. Much of the file content defined in that document is fixed in advance and provided via the skeleton files. The skeleton files are generated by QMW for CSDS wide use and do not interface to the PI provided software. Other information in the files is provided by the CDF writing software itself.
This document will only explicitly deal with information that must be generated by the PI provided data reduction software, the RDA modules or control script, and passed to the CDF writing software via the interface files (PPIF and SPIF, collectively referred to as .iff files).
Files will contain metadata in the form of attributes. Two classes of attributes are required, those that apply to the whole file and those that are attached to each variable. Data will be treated as records comprised of a time tag and a value for each CSDS variable in turn. Records will be numbered for convenience, commencing with zero.
When it is not possible to write a value for one variable in a record a fill value must be used as a place holder. When a whole record cannot be written (even if it is only a single missing record) this is treated as a data gap and no values are entered. A data gap is identified by one (or more) lines containing the letter N in the first column (see section 4 for details).