
DVO : Outstanding Issues (20081204) TBR before survey start

* fundamentals : 
  * Commit to and Require C99?  
  * Integer scalars and vectors to s64?
  * force float for / for scalars (and vectors?)

* Indices / IDs

  Do we have all of the necessary IDs defined?

  * Image IDs : we need to be able to link back to the IPP tables, but
    also guarantee an internal unique ID even if images from multiple
    databases / other sources are injects.  Note that the full 3pi
    survey will consist of ~ 6000 * 60 * 60 chips = 1.3e7 images.

    * imageID  : Internal image ID (32bit -> 4G rows)
    * externID : chip_id, warp_id, stack_id, diff_id
    * sourceID : database + table

  * Object IDs : Our internal ID needs to be unique across all
    objects.  We enforce this by generating the object ID when the
    object is created with the objID seq number from the table in
    which is was created + catID for the table.  We also need to
    generate and track the external (PSPS) ID (64bit).

    * objID : ID for object in a table
    * catID : ID of the source table 
    * extIDlo : high bits for external ID
    * extIDhi : low bits for external ID
      (or commit to a 64bit version)

  * Detection IDs : internal only; these are generated as the seq
    number on the image + the IPP image ID.

* Add PSPS obj ID code

* Orphans / Objects: PSPS considers detections to be orphans until
  they are correlated with enough (2?) other detections.  IPP needs to
  send the detections with their IPP objID, but only supply a list of
  the new PSPS 'objects' when the object is no longer considered an
  orphan.  The detection-to-object correlation is performed during the
  addstar process.  IPP needs to generate a table of new non-orphaned
  objects at that time for distribution.  How do we complete the
  hand-shake?  

* Schema updates

* External Image IDs :  

  * IMAGEID in header is the ID from chip_imfile, warp_skyfile, etc.
  * SOURCEID in header maps to the database + stages (chip, warp, etc)

