Reprocessing - When will it end?--Ryszard Pisarski,Edward Seufert, Sumant Krishnaswamy
The traditional reason for reprocessing of data is to "fix" problems that exist with the current data set. These problems generally are caused by either having bugs in the current processing software or by not having all the necessary information incorporated into the system, such as a better attitude solution, etc. The other reason for reprocessing is to produce a uniform data set in one rational format that will be archived for future generations to use.
The processing of satellite data usually accomplished by successive approximations with the hope that convergence will occur. The reprocessing of ROSAT data was no different, and these successive attempts manifested themselves in multiple releases of SASS (Standard Analysis and Software System). SASS is the primary software used to produce data sets that can be input into standard science analysis software such as PROS, XSPEC, etc. The new SASS that will be used for the reprocessing will incorporate tighter screening parameters so that errors like the aspect drift will be eliminated. Also, some accepted times for the PSPC that were wrong due to erroneous acceptance flags will be fixed. Some data sets will have an improved attitude solution through the inclusion of the WFC startracker in the solution. The raw electrical PSPC detector coordinates of an event now will be included in the SASS output. This inclusion will enable better gain determination to be achieved.
The ongoing data analysis over the years has yielded improved detector response matrices. These matrices will be used in the standard analysis to produce better source fits. Also, the very useful SIMBAD source correlations that are available for PSPC objects will be included for the HRI. These changes are some of the "fixes" that will be incorporated during the reprocessing of ROSAT data.
Aside from the above internal changes, there will be format changes. These changes are needed to simplify data management, such as better naming conventions. The HRI output has been improved by using an MPE-developed plotting package so that the user will get a cleaner output. Finally, both the European and US communities will have an identical set of data by using the Rationalized Data Formats (RDF) (see the article by Mike Corcoran in this issue) for the output of SASS. These formats can be used by a wide range of analysis tools (PROS, XSPEC and EXSAS),thus enabling users to have an easier time comparing results on both sides of the Atlantic.
The current plan is to start reprocessing of all the ROSAT data, starting with the PCV data, on or before April 1, 1994. MPE has started reprocessing and again is acting as the pathfinder to find and correct mission timeline problems. The processing capacity at GSFC can be broken down as follows: one workday will be for ongoing new data processing; three workdays will be for reprocessing; one workday will be for down time, sw problems etc. There are 38 PVC ROSAT days (RD) that require about 13 weeks for reprocessing. The 1132 AO RD will be reprocessed in 62.8 weeks (1132/6/3=62.8). Therefore, the combined reprocessing will require a total of 75.6 weeks to complete. This schedule assumes that no major problems will be encountered. This reprocessing will be known as rev 2 so as not to confuse it with other minor reprocessing.
Reprocessing - when will it end? Probably 76 weeks after the start of rev 2, since at this point the project doesn't foresee any further reprocessings.