A Case Study in Modernizing Software: The IRAS Scan Processing and Integration Tool (“Scanpi”)


Paper:	A Case Study in Modernizing Software: The IRAS Scan Processing and Integration Tool (“Scanpi”)
Volume:	411, Astronomical Data Analysis Software and Systems XVIII
Page:	33
Authors:	Alexov, A.; Good, J.; Khan, I.R.
Abstract:	The all-sky far-infrared sky survey performed by the Infrared Astronomical Satellite (IRAS), launched in 1983, remains of exceptional value in astronomy. A tool developed during the IRAS mission, the Scan Processing and Integration Tool (Scanpi) has proven indispensable in maximizing the scientific value of the IRAS data. It performs weighted average fluxes of 1-dimensional (in-scan) IRAS raw survey scans; these averages provide sensitivity gains of 2–5 over the IRSA Point Source Catalog (PSC) in the fluxes of extended, confused or faint sources. It has recently been used in research areas as diverse as searches for planetary debris disks and star formation in low surface brightness galaxies. The aging code, now under maintenance at the NASA/IPAC Infrared Science Archive (IRSA), has proved ever more difficult to maintain and build. Scanpi was written in FORTRAN66, and over the years became unwieldy with the addition of wrappers and patches to keep apace with changing platforms. In 2007, IRSA delivered a modernized version of Scanpi, designed for long term maintenance and offering new functionality. Scanpi¹ was rewritten in C and deployed on a Linux server. A major part of the development was to integrate Scanpi into the IRSA software architecture, which has been in operations for nine years, has supported over 22 million queries and is under active maintenance. Scanpi is written largely in C for performance and maintainability, and supports VO protocols. The architecture is designed as a set of stand-alone and reusable modules with simple program interfaces. Thus existing modules which perform tasks such as coordinate transformations and table filtering have been incorporated into Scanpi. We describe lessons learned and list best practices for modernizing software.