My SAX software : historical-conceptual summary

• Since the beginning I insisted that the language of election for XAS be Fortran (given the larger experience base), namely, awaiting a f8x standard which never came out, a commonly accepted (and strictly controlled) superset of f77 (see reports S14-15 69-70)

• Another key issue was portability (although at the time almost nobody had a VAX and those who didn't have one were lounging to have one, portability was designed essentially to be between VMS and Unix(es)). That was originally planned as both operating system and device independence, in the "dream book" (report S22 94)

• The same document originally planned a complete X-ray astronomy package (divided into data accumulation and reduction, image analysis, spectral analysis and temporal analysis, similar to my Exosat system. This was soon descoped to the essential mission specific accumulation part.

• Despite this the other fundamental idea of mission and instrument independence was not abandoned, and rather enforced ...
• ...via the definition of a data file format which was instrument independent and natural, while, for efficiency reason, used the native binary representation of each system unlike FITS (to which was however designed to be mappable) as defined and specified in a basic document (report S35 126). Further practical elements for not using FITS for telemetry data are given in another report ( S31 118)
• ... via the description of (instrument specific) telemetry files in a packetcap descriptor file (see report S40 137) allowing new data modes to be supported without recompiling accumulation programs.

• A report ( S36 127) defined a simple and not messy arrangement of directories where to keep the software for contrast with many other overcomplicated packages

• The client-server approach for graphics was defined in a further couple of issues of a document (report S38 133)

• Further concepts were : one command, one program ; no overarching environment, but exploitation of the operating system environment via the VOS ; user interface defined as a further refurbishment of the one used for Exosat (84) (see e.g. status report S37 132)
• all the above allowed to have the basic modules for FOT filing (see report S46 143) and data accumulation (see report S49 151) ready in prototype form to be used for ground calibration analysis, and in shape for flight data analysis a few months after launch as planned.

• while I used my own Real Programmer Tool (see R16 114) for developemnt, a tool designed by Daniele Dal Fiume was used to generate makefiles for distribution (this is described in report S41 138) reusing some part of the RPTool.