Index: /branches/eam_branch_20080223/Ohana/src/addstar/doc/notes.txt
===================================================================
--- /branches/eam_branch_20080223/Ohana/src/addstar/doc/notes.txt	(revision 16633)
+++ /branches/eam_branch_20080223/Ohana/src/addstar/doc/notes.txt	(revision 16634)
@@ -1,2 +1,32 @@
+
+2008.02.23
+
+  features I need to add / finish in dvo:
+
+  * detID / objID / imageID
+
+    some questions:
+
+      db_id
+      table_id (chip / warp / stack / diff)
+      chip_id / warp_id / stack_id / diff_id
+      
+      detections from stand-alone analysis would require an alternate sequence?
+      detections without an associated image get image_id = 0
+      
+    * imageID : we will define a new, unique running ID within dvo
+      against which joins are performed.  in addition, I will keep the
+      externally supplied ids as a reference (32bit id + 16bit source)
+
+    * detID : the detection (measure) should carry the imageID + detID as a
+      unique value. 
+
+    * objID : objects will get a unique ID when they are created
+      (sequence in table + table ID? or use the same bits as Maria?)
+
+  * extended measurement tables
+  * SDSS ingest
+  * 2MASS flags
+  
 
 2008.01.31
Index: /branches/eam_branch_20080223/Ohana/src/libautocode/def/average.d
===================================================================
--- /branches/eam_branch_20080223/Ohana/src/libautocode/def/average.d	(revision 16633)
+++ /branches/eam_branch_20080223/Ohana/src/libautocode/def/average.d	(revision 16634)
@@ -30,5 +30,13 @@
 FIELD dummy,          DUMMY,      char[2],          padding
 
-# Pan-STARRS uses a 64-bit detection ID.  keep this in two 32 bit ints for backwards compatibility?
+# Pan-STARRS uses a 64-bit detection ID.  keep this in two 32 bit ints
+# for C89 compatibility.  The objID is constructed based on the
+# position of first instatiation.  this is actually quite expensive
+# because we need to include the uniqueness test to construct this,
+# which requires a select for each new object.  Therefore, I will use
+# a table based ID (table ID + object ID), and we will have to
+# re-number the object IDs if we change the table density, OR treat
+# all subdivisions as entries which are from a foreign table.
+
 FIELD objID_hi,       OBJ_ID_HI,    unsigned int,   ID upper bytes
 FIELD objID_lo,       OBJ_ID_LO,    unsigned int,   ID lower bytes
Index: /branches/eam_branch_20080223/Ohana/src/libautocode/def/image.d
===================================================================
--- /branches/eam_branch_20080223/Ohana/src/libautocode/def/image.d	(revision 16633)
+++ /branches/eam_branch_20080223/Ohana/src/libautocode/def/image.d	(revision 16634)
@@ -48,5 +48,7 @@
 FIELD 	  code,             CODE,                 char,           image quality flag
 FIELD 	  ccdnum,           CCDNUM,               unsigned char,  CCD ID number
-FIELD 	  imageID,          IMAGE_ID,             unsigned int,   image ID
+FIELD 	  imageID,          IMAGE_ID,             unsigned int,   internal image ID
+FIELD 	  externID,         EXTERN_ID,            unsigned int,   external image ID
+FIELD 	  sourceID,         SOURCE_ID,            unsigned short, analysis source ID
 # 48 bytes 
 
Index: /branches/eam_branch_20080223/Ohana/src/libautocode/def/measure.d
===================================================================
--- /branches/eam_branch_20080223/Ohana/src/libautocode/def/measure.d	(revision 16633)
+++ /branches/eam_branch_20080223/Ohana/src/libautocode/def/measure.d	(revision 16634)
@@ -30,5 +30,5 @@
 # Pan-STARRS uses a 64-bit detection ID.  keep this in two 32 bit ints for backwards compatibility?
 FIELD detID,          DET_ID,       unsigned int,   detection ID
-FIELD imageID,        IMAGE_ID,     unsigned int,   reference to image
+FIELD imageID,        IMAGE_ID,     unsigned int,   reference to DVO image ID
 
 # do we need more resolution than a short? should this be a log?