IPAC 2MASS Working Group Meeting, #67

IPAC 2MASS Working Group Meeting #67 Minutes 8/08/95

Attendees: R. Cutri, T. Evans, J. Fowler, T. Jarrett, B. Light, C. Lonsdale, S. Terebey, J. White

AGENDA

Subsystem Design Documents
Postprocessing of April Data

DISCUSSION

Subsystem Design Documents -- A review of most of the 2MAPPS technical subsystems is scheduled for the middle of November. The review committee is expected to be primarily interested in algorithms to be employed for source extraction (point and extended), image construction, and position reconstruction. Because the 2MAPPS documentation is behind schedule in several areas, and because the Subsystem Design Specifications (SDS's) are supposed to contain the information needed by the review committee, it was decided to focus effort on producing at least rough drafts of the SDS's. Although not expected to be placed under as much scrutiny by the review committee, the EXEC and TAPELOAD SDS's are also to be stressed; the latter's first draft is past due and the former is scheduled for October 1 and requires considerable input from T. Conrow, who will not have much 2MASS time allocation (if any) after October 1. Frame flattening supports source extraction and image construction, and so the corresponding SDS is also needed.
The goal is to have the SDS's by October 15 so that they can be sent out to the review committee in time for several weeks' study before the November review. SDS's are needed for the following subsystems: EXEC, GALWORKS, PIXCAL/DFLAT, PIXPHOT/PICMAN, POSMAN, PROPHOT, and TAPELOAD. These will be INITIAL RELEASES with lots of TBD's; the main goal is to support the review. It is not anticipated that the review committee will have time or interest for lists of subroutines, FORTRAN unit numbers, etc. General data flow (e.g., input and output information) and algorithms are the essential areas to be done as completely as possible at this time. Algorithms should be presented as an executable sequence of operations as opposed to statements such as "then we solve Maxwell's equations", except a little bit of the latter may be unavoidable early on, at least as regards widely understood computations (e.g., least-squares estimation). It seems unlikely that any derivations should be needed; the existence of applicable formulas should be assumed, and if necessary, derivations can be provided later and/or elsewhere.
It is highly desirable to have as consistent a format as possible for all subsystems. A sample SDS from IRAS/SDAS will be distributed to the 2MAPPS cognizant engineers for consideration. It is also desirable to avoid unnecessary obstacles such as fancy equations unless such niceties can be obtained with no significant excess delay. Some of these issues will be addressed at next week's meeting (Tuesday 8/15, 3PM). The cognizant engineers are requested to look over the sample before then so that we can discuss what changes to make to this basic format for our own SDS's and how individual subsystem(s) might need special considerations.
The other area of documentation needing attention is the Software Interface Specifications (SIS's). These are generally one-page documents, so the time needed to produce one is dominated by the time it takes to design the interface. Such interfaces must be understood correctly by all parties making use of them, and last-minute corrections often require code changes that may introduce bugs, so it is important to get these straight from the first moment any cognizant engineer needs the information (of course, subsequent modifications are common, but the idea is to maintain coherent understanding of the interface). A list of the SIS's needed in the short term was emailed to the Working Group by J. Fowler, and that list is attached to these minutes. Proposed due dates for first versions were intended for discussion, but this was postponed because of the absence of several of the individuals involved.
Postprocessing of April Data -- R. Cutri reported that the postprocessing of the April data has begun. This will involve applying certain photometric corrections and rerunning some phases of the processing (e.g., merging), but it does not involve full scale rerunning of the pipeline.

================================= ATTACHMENT ==================================

 
    Highest-Priority Outstanding Software Interface Specifications (SIS's)
 
                   Proposed
Written By  Read By  SISID  Due Date   Comments / Questions
----------  -------  -----  --------   ----------------------------------------
   PICMAN   PROPHOT  PIC02   9/01/95   Mechanism for feeding flattened frames
                                       to PROPHOT
   PICMAN   PROPHOT  PIC03   9/01/95   Mechanism for feeding point-source
                                       detections list to PROPHOT
   PICMAN  GALWORKS  PIC01  10/01/95   Coadded images
   POSMAN  GALWORKS  POS01  10/01/95   File interface; position-reconstructed
                                       band-merged point source list
BANDMERGE    POSMAN  BMG01  11/01/95   File interface (?); band-merged point
                                       source list
    STATS   PROPHOT  STS01  11/01/95   PSFMAN interface
           GALWORKS
  PROPHOT BANDMERGE  PRO01  12/01/95   File interface; point-source extractions
                                       list for each band
 TAPELOAD     DARKS  TAP01   3/01/96   Does TAPELOAD write FITS files of dark
                                       frames? Invoke DARKS as subroutine that
                                       calls RDFRAME?
                                       
Notes: a number of important interfaces are not listed here because both sides
       of the interface belong to the same Cognizant Engineer (e.g., STATS-to-
       COFF-to-PICMAN, PROPHOT-to-CALMON, TAPELOAD-to-RDFRAME, STATS-to-
       PSFMAN); many interfaces to DBMAN are not listed here because they will
       not be needed until sometime in 1996. All of these will need to be
       defined in early 1996.
 
-------------------------------------------------------------------------------
 
            Existing Software Interface Specifications
 
Written By   Read By   SISID     Comments 
----------   -------   -----     ----------------------------------------
 RDFRAME      DFLAT    DFL01     Two-way communication
 DFLAT        STATS    DFL02
 PCP          PIXCAL   PXC01     Command-line parameters for PIXCAL