
2008.04.20

  I am trying to update the opihi vector code for two important upgrades:

  1) I want to convert the float vector to double vectors
  2) I want to allow the option of int (unsigned int) vectors, especially for flags

  I've define opihi_float and opihi_int types.  A major portion of the
  upgrade is to convert all of the places where a vector is allocated
  or manipulated as a float to use 'opihi_float' which can then be
  re-defined.  The other trick here is to identify the functions which
  require a vector to be a float and set assserts or error reporting.

  The other major change is to modify the dvo math code.  for the
  moment, all functions should result in opihi_float vectors except
  for the explicit cases of the db extractions which return flags.

  Some math operations on the vectors will force conversion to floats,
  but some are sensible as functions from int to int, and those
  results can be kept in their native format.
