The SAS provides a number of tools for analysis of data from the RGS1 and RGS2 spectrometers, starting from the raw data in the ODF and finishing with one or more high-resolution X-ray spectra and their associated RMF response matrices and background spectra. The RGS meta-task rgsproc which runs all the way from ODF to RMF, is broken down into a sequence of component tasks that deal with successive stages of the work required, allowing the user flexibility to make adjustments where necessary. While it is possible simply to type rgsproc at the command line and get reasonable results, the investment of some thought and advanced planning usually gives significant improvement in the quality of the final products. Help is always immediately available through the SAS on-line help system:
of which the rgsproc entry is a suitable RGS starting point:
Detailed descriptions of the methods used and parameter specifications are given there and liberal use of those pages is recommended. As a complement to this SAS manual chapter, RGS data analysis threads at http://xmm.esac.esa.int/sas/current/documentation/threads/  and the RGS chapter of the NASA/GSFC ABC guide  may also be consulted. Please do not hesitate to contact the helpdesk with any comments, complaints or calls for improvements in any of these documents.
In common with the vast majority of XMM-Newton data , RGS files are in FITS format so that it is useful to have at hand general-purpose tools such as FTOOLS including fv  in particular to inspect files. The spectra and response matrices that are the final products of the RGS SAS procedures are intended for further analysis outside the SAS with standard tools such as Xspec , SPEX  or ISIS .
The results of an automatic version of rgsproc that is run on all ODFs are also available via the archive and will have been sent to observers with the original ODF distribution. These automatically generated results, known as the PPS products, are discussed in § 5.19 below. Naturally, these procedures are run without human intervention and thus are not expected to emulate the quality that can often be produced by the individual attention of a well-informed and creative astronomer. Experience shows, however, that they are useful, especially when used in conjunction with the other XMM-Newton instruments. It is most likely the case, though, that an RGS observer will benefit from running the analysis again starting from the ODF. This chapter is organised roughly along the lines of a typical session of RGS work.
Immediately, the most important things to be considered are
SAS v8.0 in 2008 is the first version to offer accumulation of spectra on a wavelength grid as opposed to the dispersion angle grid that remains the default. Such wavelength-based spectra are expected to allow easier and more robust analysis.