Please note: These file names are the ones created by standard processing for everything that goes into the archive, but not necessarily what the user sees; files may be renamed when transferred to the user (e.g., dropping digits in the start time, leaving off processing run versions).
The prime motivations for the convention are:
<i><s><t><v>_[<f>_]<c><l>[<sl>].fits <i> instrument: "acis", "hrc", ... <s> data source: "x", "f", "t", "b", "s", "u" <t> TSTART (integer part): "#########" For higher level products: ObsId and observation interval <v> processing run (version): "N###" <f> optional discriminator (e.g., FEP Id, TU) <c> contents: "evt", "prf", "hst", "src", "win", "bias", ... <l> processing level: single digit: "0", "1", "2" <sl> sublevel: single lower case letter; e.g., 1.5 is "1a"Example: acisf123456789N000_evt1.fits
acis AXAF CCD Imaging Spectrometer subsystem hrc High Resolution Camera subsystem ephin Electron-Proton-Helium INstrument subsystem pcad Pointing Control and Attitude Determination subsystem tel Telescope subsystem sim Science Instrument Module subsystem obc On-Board Computer software subsystem ccdm Communications Command and Data Management subsystem cpe CPE hardware and software subsystem eps Electrical Power Subystem thm Thermal control subsystem sms Structure subsystem prop Propulsion and pyro subsystems misc Other AXAF signals orbit Orbit ephemeris sun Solar ephemeris moon Lunar ephemeris angles Viewing angles clock Clock measurements axaf Certain mission-level itemsData Source Values
x XRCF f Flight t TRW b Ball Bros c Other lab calibration s Simulated u UnknownProcessing Version
It is assumed that decisions about reprocessing of production data will be made at a kind of CCB level and that the processing version number will hence be provided "manually" to the processing pipeline.
On purpose, I have not defined the format of the optional discriminator. I would urge developers, though, to design them as compact as possible, preferably without underscores. Similarly, it is assumed that common sense will be used in defining the instrument and contents strings. Where necessary, changes to ICDs may be suggested. The value "TU" is reserved for files containing quantities in telemetry units that duplicate other files containing the same data in engineering (calibrated) units.
We will not finalize this convention, yet, but one might envision shortening the above example to: