Wish list of additional features
-------------------------------
o convergence of amplitudes should be measured with residuals,
  not updates. (RJH)
o Norms of amplitudes & related measures
o Largest amplitudes
o CCD (because its there)
o Out of core DIIS
o Control to switch off DIIS entirely
o AO-driven stuff should use Texas blocking interface

Notable changes to code
-----------------------
o Began user's guide documentation
o PRINT input directive renamed to IPRT to make room for standard
  NWChem PRINT controls
o Starting to add NWChem-style print controls to objects
o Made term debugging switches available via input deck (not sure if
  this is really a good idea or not!)
o Commented out calls to rtc (doesn't exist on Solaris, links, but
  does not work as intended on IBM)
o Eliminated RSize & ISize
o Made DIIS subspace size an input directive (DIISBAS)
o /para1/ replaced with called to GA_NodeID() and GA_NNodes()
o /len/ turned into ccsd_len.fh include file
o /debug/ turned into ccsd_debug.fh include file
o Implicit NONE in all routines
o Default convergence threshold 1.0e-8 --> 1.0e-6 at suggestion of RJH
o All routines have RCS/CVS $Id: NOTES 19708 2010-10-29 18:04:21Z d3y133 $
o Triple transformation now uses Texas blocking [ATW]
o DIIS storage compressed (using permutational symmetry) [RK]
o Frozen orbitals work for energies [RK]
o Convergence threshold changed from integer to real (to conform with
  the rest of NWChem)
o Tol2E made user input parameter. NOTE: Effects only 3-/4-virtual
  terms and triples, which are "AO based".  MOINTS transformations
  used elsewhere still have hard-wired 1.0e-12 value!
