README_Uedge_vers

                            LEGAL NOTICES

This work was performed at the University of California, Lawrence Livermore
National Laboratory (UC LLNL) under contract no.  W-7405-ENG-48 (Contract 48)
between the U.S. Department of Energy (DOE) and The Regents of the University
of California (the University) for the operation of UC LLNL.  The rights of
the Federal Government are reserved under Contract 48 subject to the
restrictions agreed upon by the DOE and University as allowed under DOE
Acquisition Letter 97-1.

UEDGE is released for DOE-related work, and recipients are required to contact
the author before further dissemination of the package or any of its
components.

This work was prepared as an account of work sponsored by an agency of the
United States Government.  Neither the United States Government nor the
University of California nor any of their employees makes any warranty,
express or implied, or assumes any liability or responsibility for the
accuracy, completeness, or usefulness of any information, apparatus, product,
or process disclosed, or represents that its use would not infringe privately
owned rights.  Reference herein to any specific commercial products, process,
or service by trade name, trademark, manufacturer, or otherwise, does not
necessarily constitute or imply its endorsement, recommendation, or favoring
by the United States Government or the University of California.  The views
and opinions of authors expressed herein do not necessarily state or reflect
those of the United States Government or the University of California, and
shall not be used for advertising or product endorsement purposes.

*****************************************************************************
This file was committed $Date: 2021/05/12 23:23:14 $ and tagged $Name:  $

Version 3.0	Corresponds to CVS archives as of 12/14/99

Version 3.1	Corresponds to CVS archives as of 01/13/00

Version 3.2	Corresponds to CVS archives as of 02/20/00
		- Fixes bug in visx flux-limit term msh
		- Allows boundary conditions to be properly relaxed with
		  stiff rate equations for svrpkg=nksol and isbcwdt=1

Version 3.3	Corresponds to CVS archives as of 08/07/00
		- Substantial changes to the parallel current at the plate
		  boundaries (including kappar,l) giving better ExB behavior
		- Force feex to always be into divertor plates
		- Remove guard-cell values of Te on plates for determining
		  parallel electron heat conductivity there
		- Include momentum sources from ionization, recomb., & cx
		  for multispecies equations if full momentum eqns. solved
		- Added option to use lgmax for flux limiting gas diffusion
		  using this scale-length - more robust
		- Change radial convective flux for ion momentum eqn to use
		  the harmonic average of two pairs of densities; can be
		  important for neutral momentum near ionization front
		- Added coeff. cfloye,i to change radial convective heat flux
		  from 5/2 to 3/2 for modeling ExB turbulence; default=5/2
		- Allow simultaneous albedo, chem sputtering, and recycling
		  for gas side-wall boundary conditions
		- Added model for wall impurity evaporation based on the input
		  temperature profile tvapo.
		- Allow gas outflow measured on inner wall to be injected on
		  outer wall, and vice versa
		- Fixed potential bug defining ixc=max(0,ixpt1+1) in bouncon
		- Fixed bug in definition of xcwi,o used in determining wall
		  source location; also in geometry.m, gradb2 had incorrect
		  cos(angfx) factor.

Version 3.4     Corresponds to CVS archives as of 11/28/00
                - changed the interpolation (for isgindx=1 default) to be 
                  independent in poloidal segments of leg, core, and then leg
                - changed variable for radial pressure-gradient velocity from
                  vycd to vycp
                - enabled the use of ADAS ionization and recombination rates
                  that can now depend on electron density even for impurities
                - corrected sign on nurlxe,i for iste,ipfc=0 boundary
                  condition - only affectd svrpkg="vodpk"
                - corrected iv indexing problem for isutcore B.C. for
                  isphibcc.ne.1
                - added time-dependent inertial radial-ion-current called
                  fqydt; comes from m*dv/dt term in perp. ion momentum eq.

(Below t_Ver... means a temporary version which is not saved on all platforms)
t_Version 3.41	Corresponds to minor modification to Version 3.4 where a call
		to the allocate routine is force on the first call to exmain
		- added variable icallall to indicate when allocate called once
		- initialize ifexmain to zero (previously unset)

t_Version 3.42	Only one change from V3.41; the default value of fvapi is
		changes from 1 to 0.  This can affect chemical sputtering on
		the inner (private-flux) wall and fvapi=0 should be used.
		V3.41 can still be used if one explicitly sets fvapi=0 before
		executing a run.

t_Version 3.43	Only changes from V3.42 are that chemical sputtering is now
		included on the wall regions even when matwallo,i=1, just as
		as the albedo is always included independent of matwallo,i;
		and the use of energy-transmission boundary conditions on the
		inner wall is not automatic for matwalli=1, but must be set
		by iste,ipfc = 3, if desired.

Version 3.44	May be the same as V3.43, but dervived from the CVS achieve
		dated 05/06/01, rather than the working files from Rognlien.

Version 3.45    From cvs archive on 06/08/01.  Corrects a bug when running on
                LINUX machines by placing the definition of ix_fl_bc to before
                the first do-loop in subroutine bouncon.  Also adds the outer
                midplane poloidal index, ixmp, to be available from the
                parser, and improves the accuracy of the ixmp calculation.
                Finally, replaces the default for the potential boundary
                condition on the core to be iphibcc=0.

Version 3.46    From CVS archive of 07/31/01.
		Only one change for V_3.45, where Takenaga's bug in jac_calc
                is corrected, i.e., ix and iy are now defined just before
                idxphi(ix,iy) if-test.  Probably has no impact, but
                possibly it does (read out-of-bounds problem)

Version 3.47	From CVS archive of 10/09/01.
		Corrected core boundary conditions to include the fact the
		diamagnetic surface fluxes can exits.  UEDGE only uses
		the grad_B portion of the diamagnetic drift velocity
		in the "body" of the computational region since the
		divergence of the other portion is identically zero.
		However, on the boundary surfaces, one must resort to
		the full diamagnetic expression for obtain particle
		and power fluxes.  The correction for the radial
		current is available in older UEDGE version, but
		requires proper setting of switches.  For cross-field drifts
		on, the recommended settings of the switches are:
		cfniybbo=0., cfniydbo=1., cfeeybbo=0., cfeeydbo=1.,
		cfqybbo=0.  cfqydbo=1. Also changed scale length B.C. for
		lyni (isnwconi,o = 2) to apply to niy0 and niy1 variables
		instead of ni(,ny) and ni(,ny+1); this is consistent with
		definition of fniy for nonorthogonal mesh.


Version 3.48	From CVS archive of 10/22/01.
		Corrected a bug in the grad_B velocity for bphi>0 (the
		standard DIII-D case has bphi<0).  Also corrected a bug
		in the nonorthogonal-mesh representation of the cross-field
		v2 velocities [mistakenly divided v2's by cos(angfx)].  Set
		the default of del2nksol=1e-14 (instead of zero, which then
		used machine-calculated round-off - sometime too small on 
		LINUX). Corrected classical parallel viscosity coefficient
		from 1.92 to 0.92, as given by Braginskii.

Version 4.0	Corresponds to CVS archive on 01/23/02.  This is a major
		upgrade that includes the capability of unbalanced double 
		nulls developed by M. Rensink.  A number of other improvements 
		have also been made; these are as follows:
		- refinement of the multispecies capability where one can
		  now have the impurity species be dominant over hydrogen
		- tritium can now run run as an impurity species
		- generalized boundary conditions allowing separate inner (pf)
		  and outer wall scale lengths; lyup added
		- allow fixed gas density at plate boundaries for BORIS comp.
		- corrected problems associated with using f90 on the SUNs;
                  must now use SUNWspro 5.0 compiler or higher.  Thus, BASIS12
		  and f90 is now be used on all platforms (SUN, LINUX, DEC).
		- Added background source to erliz ionization energy loss
		  using ngbackg; this changes vsoree and electron eng. eqn.

		Below is a compilation by Rensink of changes to variables
		in moving from Ver_3.48 to Ver_4.0:
	Input/output variables with changed names (and dimensions):

   		csin(nispmx)  -> csfaclb(nispmx,1)
   		csout(nispmx) -> csfacrb(nispmx,1)

   		fchemypi -> fchemylb(1)
   		fchemypo -> fchemyrb(1)

   		fphysypi -> fphysylb(1)
   		fphysypo -> fphysyrb(1)

   		recypi(0:ny+1,ngspmx) -> recylb(0:ny+1,ngspmx,1)
   		recypo(0:ny+1,ngspmx) -> recyrb(0:ny+1,ngspmx,1)

   		recycfi -> recycflb(1)
   		recycfo -> recycfrb(1)

   		ndati(ngspmx) -> ndatlb(ngspmx,1)
   		ndato(ngspmx) -> ndatrb(ngspmx,1)

   		ydati(ngspmx,50) -> ydatlb(ngspmx,50,1)
   		ydato(ngspmx,50) -> ydatrb(ngspmx,50,1)

  		rdati(ngspmx,50) -> rdatlb(ngspmx,50,1)
  		rdato(ngspmx,50) -> rdatrb(ngspmx,50,1)

  		albpi(0:ny+1,ngspmx) -> alblb(0:ny+1,ngspmx,1)
  		albpo(0:ny+1,ngspmx) -> albrb(0:ny+1,ngspmx,1)

 		fngxsi(0:ny+1,ngspmx) -> fngxslb(0:ny+1,ngspmx,1)
 		fngxso(0:ny+1,ngspmx) -> fngxsrb(0:ny+1,ngspmx,1)

 		adati(ngspmx,50) -> adatlb(ngspmx,50,1)
 		adato(ngspmx,50) -> adatrb(ngspmx,50,1)

 		ngplati(ngspmx) -> ngplatlb(ngspmx,1)
 		ngplato(ngspmx) -> ngplatrb(ngspmx,1)

  		sputti(0:ny+1,ngspmx) -> sputtlb(0:ny+1,ngspmx,1)
 		sputto(0:ny+1,ngspmx) -> sputtrb(0:ny+1,ngspmx,1)

 		albedopi(ngspmx) -> albedolb(ngspmx,1)
 		albedopo(ngspmx) -> albedorb(ngspmx,1)


	Input/Output variables with changed dimensions only:

   		newbcl -> newbcl(1)
   		newbcr -> newbcr(1)

   		phi0l(0:ny+1) -> phi0l(0:ny+1,1)
   		phi0r(0:ny+1) -> phi0r(0:ny+1,1)

   		yylb(0:ny+1) -> yylb(0:ny+1,1)
   		yyrb(0:ny+1) -> yyrb(0:ny+1,1)

   		upxpt(1:nusp)   -> upxpt(1:nusp,1)
  		nixpt(1:nusp)   -> nixpt(1:nusp,1)
  		visyxpt(1:nusp) -> visyxpt(1:nusp,1)
  		vyhxpt(1:nusp)  -> vyhxpt(1:nusp,1)
  		vyvxpt(1:nusp)  -> vyvxpt(1:nusp,1)
  		fmihxpt(1:nusp) -> fmihxpt(1:nusp,1)
  		fmivxpt(1:nusp) -> fmivxpt(1:nusp,1)

   		fdiaxlb(0:ny+1) -> fdiaxlb(0:ny+1,1)
   		fdiaxrb(0:ny+1) -> fdiaxrb(0:ny+1,1)

   		bcel(0:ny+1) -> bcel(0:ny+1,1)
  		bcer(0:ny+1) -> bcer(0:ny+1,1)
  		bcil(0:ny+1) -> bcil(0:ny+1,1)
   		bcir(0:ny+1) -> bcir(0:ny+1,1)
  		kappal(0:ny+1) -> kappal(0:ny+1,1)
   		kappar(0:ny+1) -> kappar(0:ny+1,1)

   		fqpsatlb(0:ny+1) -> fqpsatlb(0:ny+1,1)
   		fqpsatrb(0:ny+1) -> fqpsatrb(0:ny+1,1)

	NOTE on option isnfmiy=1:

	In previous versions of UEDGE, isnfmiy=1 had the effect of setting
	fmiy=0 for velocity cells straddling the x-point because the spatial
	factors gyhxpt,gyvxpt,sxyxpt were not calculated.  Now, these spatial
	factors are finite for all EFIT-based configurations, so in order to
	recover previous results you must set the new coefficient cfnfmiy=0
	when isnfmiy=1; the preferred (default) setting is cfnfmiy=1.

Version 4.1	Corresponds to CVS archive on 02/16/02. 
		- Added the additional flux limit for the neutral thermal
		  conductivity and viscosity via the input scale lengths 
		  lgtmax and lgvmax.
		- Slight change to pscx0 for first impurity charge state to
		  include niz_floor for consistency with Ver_3.48.
		- Corrected bug in Ver_4.0 only relating to the sign of
		  the albedo pumping (albrb) on the right plate.
		- Changed the radial gas scale length for isngcore=2 from
		  v_t**2/(nucx*nuiz) to v_t**2/(nuix*nuiz).
		- Corrected Ver_4.0 only bug in calculation of sygytotc in 
		  geometry.m; ixpt1,2(2) was referenced even if nxpt=1.
		- Corrected Ver_4.0 only bug in first definition of argo in
		  subroutine walsor; changed xnoti --> xnoto

Version 4.12	Corresponds to CVS archive of dce package on 05/08/02. 
                - Dce cvs -r1 - all routines are rpc. Connection established
                  via well know port (12100) first then portmapper second.
                - Dce cvs -r2 - routine rzxform, and supporting modules, linked
                  directly into Uedge instead of the client rpc stubb. 

Version 4.12_mds  Identical to Version 4.12 above but has package mds added.
                  This provides access to MDSPlus through which users may
                  access DIII-D equilibria, time series data, and profile
                  data to name but a few. (LINUX version only for now)

Version 4.13	Corresponds to CVS archieve of 05/08/02, except the dce file
		c_dce.c has be corrected (08/15/02) by Meyer and the routines
		wmodi and wmodo in boundary.m have corrected signs in the
		definition of fngysi,o used when gas is removed in one
		source region and injected in another (08/15/02).

Version 4.2	Corresponds to CVS archive of 08/26/02.  Changes include:
		- Add the ExB boundary velocity for plate flux which got lost
		  after Ver_3.48.
		- Now two options for phi on side walls; if iphibcwi,o=0, then
		  d(phi)/dy=0, or if iphibcwi,o=1, then phi=phintew,o*te/ev.
		- Allow diamagnetic and B x gradT heat flows for the plate
		  B.C. if cfeixdbo & cfeexdbo = 1.
		- Allow the density radial scale length to be poloidally
		  dependent by setting isulynix=1 & filling lynix(ix,2).
		- Corrected an error in the sign of re-injected gas on the
		  outer wall in routine wsmodo.
		- Corrected impurity isotope counter, nzsp_rt, so that it works
		  when turning off 1 of 2 or more isotopes through isnion=0.
		- New core density boundary condition isnicore=5 for dni/dy=0.
		- Allow modification of total convective gas flux at guard-cell
		  interface through gcfacgx,y; added (1-rrv**2) factor; similar
		  factor gcfacgtx,y only effects gradT convective gas flux.
		  New default has gcfacgx,y=1, but gcfacgtx,y=0 for robustness.
		- New callable subroutine pradpltwl calculates radiation power
		  flux on plates and wall; pwr_plth,z and pwr_wallh,z arrays.
		- Extend zero fluxes for geometry="dnbot" to ix=nxc+-1.
		- Change difpr meaning and add difp_use to give density diff
		  velocity prop_to (difpr+difp_use)(1/P)(dP/dr - 1.5ndTe/dr).
		- Generalized vy_use(ix,iy,ifld) to be species dependent.
	 	- New option for isbohmcalc=3 gives anomalous diffusion scaling
		  as (1/B)**inbdif with input coeff. difniv(iy), kyev(iy),etc.
		- Add array wjdote to hold the J_dot_E electron heating term.
		- Add routines write_profs_boris & read_profs_boris to exchange
		  profile data with the BORIS code.
		- Allow inclusion of only a radial portion of the inner core
		  region by setting nyomitmx > nysol.
		- Add a factor cffqpsat that scales the allowable saturation
		  current in the sheath-potential calculation.
                - Allow DCE package operation on all platforms.
		- Include kinetic Pastukhov correction to elec. particle loss
		  for sheath potential calc; set cfkincor=0 to omit.
		- Include flux limiting of parallel thermal force through
		  flalftf; default (1e20) is off, best guess is flalftf=1.

Version 4.21	Corresponds to CVS archive of 09/16/02.  Changes include:
		- Allow poloidal distribution of extra cells around x-point
		  (via non-zero nxxpt) to be manipulated independent above (u)
		  and below (l) the x-point via the parameters alfxptu,l and
		  alfxpt2u,l that replace alfxpt and alfxpt2.  Here alfxptu,l
		  controls the spacing of the nxxpt cells relative to each 
		  other, and alfxpt2u,l adjusts the spacing of the newly
		  refined region and the one adjacent poloidal cell.
		- The second "localized" boundary condition for phi on core 
		  boundary is now applied at the outer midplane rather than 
		  ixpt2; also, isutcore=2 now gives d^2(ey)/dy^2 = 0 at outer
		  midplane.
		- Allow the core boundary condition d^2(up)/dy^2=0 if
		  isupcore=2.
	 	- Add option to turn off radial cross-field drifts only at the
		  private flux and outer wall by setting isybdrywd=1.  This
		  gives only diffusive flux, for regions where matwallo,i=1, 
		  that is everwhere outward if a gradient scale-length option
		  is specified (isnwcono,i=3).
		- Set default for gas flux-limit flalfgy=1e10, which gives no
		  flux limit unless reduced by user; defaults for flalfgx,xy
		  are also 1e10.  All three should be set equal.
 
Version 4.22  Corresponds to CVS archive of 12/10/02.  Changes include:
                - Changed the defaults for background neutral and impurity
                  sources with the exponents now ingb=inzb=2 and
                  ngbackg(ngspmx)=1e14; these changes can have big impact
		  on previous converged solutions.
                - Changed the evaluation of the fqy current to include the
                  points iy=0 and ny; previously omitted because of 2 BC at
                  each wall. Some effect on the potential.
                - Added an option to set a constant value of zeff.  Use
                  iszeffcon as switch and value is zeffcon.
                - Added separate energy transmission factors for the walls,
                  called bceew & bceiw; used if iste,ipfc=4 or iste,iwc=4.
                - Generalized spatially dependent radial diffusion if 
                  isbohmcalc=3 to give, for example,
                  dif_use=(Btmid/Bt)**inbtdif*(Bpmid/Bp)**inbpdif. Also include
                  vy_use=vyconv(iy)*(Btmid/Bt)**inbtdif*(Bpmid/Bp)**inbpdif.
                - Changed the form of the E-dot-J heating term to use E & J
                  explicitly, instead of div(J*phi) form based on div(J)=0.
                  At ix=0 and nx, ex does not include bias-voltage effect.
                - Special case for recycrb,lb < -1 now gives plate neutral BC
                  with ng=nglfix or ngrfix.
 
Version 4.3  Corresponds to CVS archive of 06/30/03.  Changes include:
		- Added a CVS branch for making a PYTHON version of UEDGE;
		  contact authors.
		- Added the capability of making UEDGE with the BASIS MIO
		  software, instead of MMM
                - Joule heating can now be calculated two ways: jhswitch=1
		  gives old method where J.E=-div(J*phi) since div(J)=0; and
		  jhswitch=2 uses J.E directly, with correct BC for biasing.
                - Changed material-wall boundary conditions to apply for
		  matwalli,o > 0, rather than previous matwalli,o = 1.
		- Made electron velocity vey purely diffusive if isybdrywd=1,
		  just as the ion velocity vy has been.
		- Corrected array initialization of pwr_wallz,h with
		  ny+2 --> nx+2 (in pradpltwl)
		- Corrected ixv viewing index in subroutine pradpltwl to work
		  properly for full double nulls with 2 x-points.
		- Corrected re-injected neutral flux to be consistent with
		  pumped flux at iy=0 and iy=ny+1 (iy=0 & 1 inconsistency).
		- Corrected minor bug for kfeix at right boundary - only
		  affects isimpon=5 (Hirshman's reduced-ion impurities).
		- Removed INEL average-ion impurity model (isimpon=3,4).
		- Removed the svrpkg="lsode" solver option; use "vodpk"
		- Corrected flxlimf flux-limiting factor so that ltmax does
		  not appear in denominator, which can lead to divide by 0.
		- Omitted initialization of some arrays (te,iwalli,o; bctype;
		  capx; dphi_iy1; and cfvli) due to IBM SP problem
		- Added variable yloext and switch isyloext for external dyl/dt
		  to test coupling to turbulence
		- Added two user-specified radial velocities, vyte,i_use, 
		  which yield radial energy fluxes vyte,i_use*te,i*ni; used
		  for coupling external energy fluxes (e.g., from BOUT).
		- Moved some calculations from readefit to procefit to support
		  reading MHD data from files other than aeqdsk & neqdsk; 
		  target is MDSPLUS.
		- Removed all calls to allot and instead use gchange; 
		  facilitates PYTHON version
		- Added kypton to the ionization energy tables in fimp.m, so
		  multi-specie kyrpton calculations are possible.
		- Added two routines to calculate the line emissivities of
		  hydrogen and impurities based on scripts from Gary Porter.
		  Uses Isler data.  Subroutine readrates(apidir,fname) reads
		  data in path apidir having name fname.  Subroutine
		  calcrates(a,b,c) calculates emissivity c for density a and
		  electron temperature b.
		- Added a compiled function to calculate line integrals, called
		  lineintegral(arg_rz,rvertex,zvertex), which calculates the
		  integral of the pixel array arg_rz along a path given by the
		  vector (rvertex(1:2),zvertex(1:2)).  Similar in functionality
		  to Gary Porter's script lineintegral, but much faster.  Can
		  now be used in scripts such as write_fslw_H for hydrogen 
		  lines.
		- Changed names of startup files defining various paths and
		  initializations from aph_dir, uedge_path, init.bas to	
		  .aph_dir, .uedge_path, and .init.bas.  Now if absent, UEDGE
		  ignores them and proceeds.  Store in uedge/scripts.
		
Version 4.31  Temporary version with no public release, approximately 
		corresponds to CVS archive of 12/02/03.  Changes include:
		- Physical and chemical sputtering from hydrogen ions incident
 		  on the walls included.
		- The low-energy reduction factor in the Haas chemical
		  sputtering for isch_sput=7 is now an input variable called
		  redf_haas.
		New input variables for wall sputtering:
		  isi_sputw(igsp) # Determines the outer wall sputtering for
		  		  # ions for gas species igsp 
		  isi_sputpf(igsp) # Determines the private flux wall
		  		   # sputtering for ions for gas species igsp
		Options:
		  isi_sputw(igsp) = 0 # No ion sputtering (old case)
		  isi_sputw(igsp) = 1 # Includes physical sputtering from ions
  		  isi_sputw(igsp) = 2 # Adds chemical sputtering from ions
		  		      # using model isch_sput(igsp);  must set
				      # isch_sput to nonzero value (e.g., 7)
		  Normally, igsp=2 for the switches above to simulate carbon
		  in a hydrogen plasma.
		- Setting isch_sput=7 also turns on the chemical sputtering
		  from neutral gas as before; this can be scaled by the factor
		  fchemywo (outer wall) or fchemywi (for private flux region).
		  However the ion-induced chemical sputtering (from
		  isi_sputw=2) does not have a scaling factor.
		- The new sputtering fluxes, multiplied by the cell area, from
		  the ions are stored in the output arrays:
		  sputflxw(ix,igsp) # units of particles/sec (positive) 
		  sputflxpf(ix,igsp) # units of particles/sec (negative)
			where ix is the usual poloidal index.  These
			fluxes are added to the impurity gas flux at the wall,
			in addition to any recycling or pumping specified.
			Note that impurity sputtering is not included yet.
		- Also, the reduction factor in the Haas chemical sputtering
		  model for isch_sput=7 is now a parameter called redf_haas
		  that can be changed from the UEDGE prompt or in an input
		  file.  The default is the previous value of 0.2.

Version 4.32  Corresponds to CVS archive of 12/19/03.  Changes include:
		- All changes noted above in temporary version 4.31
		- A new LINUX branch for the PYTHON version is added, which
		  is accessible in uedge/Pyuedge/LINUX.  Requires MPPL and
		  MAC, but not BASIS; MMM or MIO BASIS builds also work
		- For istabon=3 (DEGAS tables), charge-exchange rate rcx
		  now assumes neutral temp = ion temp (before T_n=const)
		- Corrections, additions to the mesh generating routines:
		  > for inflx & inflx1, replace redundant nyflx, npsi with
		    nym, jdim
		  > other modifications to inflx1 to improve mesh generation
		  > modified grd/grdwrit & grdcomp.m to work for ishalfm=1, 
		    corresponding to a half-mesh
		  > other details from cvs log files flx/flx.v flxread.m 
		    flxwrit.m flxcomp.m and grd/grd.v grdread.m grdwrit.m
		    grdcomp.m
		  > different end-of-mesh test allowed for each plate
		  > made istcvon flag for controlling search directions in
		    mesh construction obsolete; use altsearch flag instead

Version 4.33  Corresponds to CVS archive of 02/20/04.  Changes include:
		- Pyuedge branch has been modified to include MAC (called mac2
		  to avoid confusion) in uedge/Pyuedge/Mac_scripts and
		  the pybasis source in uedge/Pyuedge/Pymac_scripts
		- Pyuedge Makefile cleaned up for LINUX
		- Updated pybasis with new scripts from Dave Grote, including
		  his more extensive pyBasis
		- Removed prefixes flx. from a couple of gallot calls, so
		  that EFIT files can be used with pyuedge
		- Change maximum scale lengths for flux limiting gas density
		  and temp fluxes, lgmax and lgtmax, to be vectors (igsp)

Version 4.34  Corresponds to CVS archive of 07/02/04.  Changes include:
		- Added user-input fluxes fniyos_use, feeyosn_use, and
		  feiyosn_use; to fix the fluxes fniy etc to these values,
		  added evolving vy_cft etc; for BOUT coupling
		- Added fixed convective components for feey and feiy called
		  vyte_use and vyti_use, added to kye,i_use for BOUT coupling
		- Added impurity-impurity collisions (CX & elastic) for large
		  impurity fractions.  New cross-section sigcxms and elastic
		  rate keligii; impact neutral diffusion & ion up and ti.
		  New collision frequencies nucxi, nueli, and nuelg.
		- New impurity-impurity friction turned on with cfupimpg and
		  ion temp (ti) cooling turned on with cftiimpg.
		- Modified subroutines writing DEGAS namelist input files to
		  account for extra poloidal x-point cells via nxxpt>0 and
		  different nxcore inside and outside.
		- Mapping of mesh from UEDGE to DEGAS now utilizes (rm,zm), 
		  not (cmeshx,cmeshy). Also, wall DEGAS construction now done
		  in subroutine degasgrid.

Version 4.35  Corresponds to CVS archive of 09/14/04.  Changes include:
		- Included three more options for hydrogen atomic rates from
		  D. Stotler (Aug. 2004):
		    -- istabon=11 for optically thin - same as istabon=10
		    -- istabon=12 for optically thick Lyman_Alpha
		    -- istabon=13 for optically thick to all lines
		- The new hydrogen rates also include diagnostic data for
		  coupling coefficients of n=4-9 hydrogen lines to the ground
		  state (called pneX1, where X=n) and to the continuum (pneX2).
		  Previous only n=2-3 line data was available.
		- Provided an option to make the toroidal magnetic field a 
		  constant at bcentrg; set isbphicon=1; used for approximating
		  stellartors.

Version 4.36  Corresponds to CVS archive of 10/11/04.  Changes include:
		- Included the escape-factor radiation transport model for
		  hydrogen Lyman-alpha radiation developed by H. Scott and M.
		  Adams for UEDGE.  Hydrogen ionization/recombination rates 
		  and energy loss are now a function of the minimum local 
		  optical depth for Lyman-alpha  to a boundary surface
		  [calculated as rtau(ix,iy)].  These use a new hydrogen table
		  from H. Scott called ehrtau.dat with rates a function of te, 
		  ne, and rtau. To access the model, set
		    -- rtauxfac=1 (or positive), and either:
		    -- istabon=14 gives linear intepolation in rtau, or
		    -- istabon=15 gives log interpolation in rtau
		  Note that these models may cause convergence problems
		  since they give nonlocal dependences for rtau, and no attempt
		  has yet been made to provide a "good" Jacobian for the 
  		  preconditioner.
		- A user-specified radial convective velocity, vyup_use(ix,iy)
		  has been included to convect parallel momentum energy density
		  radially.  In anology with vyte_use for Te, values can come
		  from BOUT in the UEDGE/BOUT coupling.

Version 4.37  Corresponds to CVS archive of 04/03/05.  Changes include:
		- Changed flags controlling plate temperature BC; ifluxl,r is
		  now replaced by ibctepl,r and ibctipl,r for electrons and 
 		  ions, with l,r denoting left and right plates; =0 for fixed
		  te=tepltl,r; =1 for sheath trans; =2 for dte,i/dx=0.
		- Added core density BC that allows user to specify radial
		  gradients if isnicore=5; lynicore is scale length and
		  ncoremin is minimum allowed density
		- For the core density BC isnicore = 3, added the neutral 
		  current through the core bdry to curcore, i.e., the net ion
		  current is now curcore-recycc*sum(fngy) over the bdry; 
		  NOTE: presently, this just applies to hydrogen (species 1).
		- Added variable core recycling coeff. recycc (=1 default)
		- NOTE: with Ver_4.34, small added cx scattering included for
		  impurity ions through sigcxms, but not initialized until 
		  Ver_4.35-6; thus, may need to set sigcxms=0. to agree with
		  older cases.

Version 4.38  Corresponds to CVS archive of 05/24/05.  Changes include:
		- Corrected a bug for reinjecting wall fluxes at remote
		  locations when iscpli,o = 1 (default=0); previously albedo 
		  sources were double-counted when remote reinjection used.
		- Generalized lynix from a 2D array to 3D with the third
		  index being for density species; used if isulynix=1
		  (default=0).
		- Added a third dimension to arrays (wsveh, wsveh0, welms1,
		  welms2) in subroutine readeh2 of aphread.m. For this routine,
		  the third index does not vary from unity, but now corrects 
		  some compiler complaints.
		- Update the dce/Package file to avoid "for" loops with	
		  RPC_GENERATED, which can cause compiler problems
		- Corrected a possible overflow problem on some machines
		  introduced by changes to bbb/boundary.m, Version 4.37,
                  concerning recoding the plate temperature boundary 
		  conditions.
 		- The pyuedge make system has been improved by L. LoDestro 
		  and is included here.  Only of interest for PYTHON users
		  going down the uedge/Pyuedge branch in the uedge tree; see
		  uedge/README_Pyuedge for details of how to use.
		- Added a volume source for diffusive neutrals.  Current is
		  ivolcurg(igsp); shape uses r,z0ng positions & r,zwng widths.

Version 4.39  Corresponds to CVS archive of 03/17/06.  Changes include:
		- Corrected a bug in the reinjection of neutral gas from one
		  "source/pumping" wall region to another.
		- Added a diagnostic section to jac_calc to stop UEDGE at
                  a specific Jacobian elements & compare yldot_pert and
		  yldot_unpt.
		- Included collisionality factors coll_fe,i for grad_B
		  drifts, reducing them by 
		  cfnus_e,i*nu_star_e,i/(1 + cfnus_e,i*nu_star_e,i).
		- Generalized connection length variable, lcon, to be 2D and
		  changed name to lconi,e; near separatrix, limit by banana
		  width defined through temperatures tibsep and tebsep.
		- Added option to use the neutral pressure, pg=ng*tg, as the
		  differenced variable for the neutral continuity equation.
		  This option turned on by ineudif=2 (the option to use 
		  log(ng) as the variable is now moved to ineudif=3.
		- Implemented substantial improvements to the Jacobian 
		  calculation that now captures all elements for orthogonal
		  meshes and most elements for nonorthogonal; set yinc=2 
		  as default "box" y-range for the Jacobian calculation. 
		- Changed to use tg instead of ti in the calculation
		  of one component of the nonothogonal gas flux (fngxy);
		  if istgcon=0 (default), tg=ti, and it makes no difference,
		  but otherwise it will.
		- Corrected upstream boundary condition to include iy=iysptrx
		  for the seldom-used isfixlb=1 case that forces the ix=0 BC
		  to be defined by nibprof, etc. profiles.
  		- Included potential boundary conditions for the guard cells
		  that along the limiter with islbcp if islimon=1.
		- Implemented neutral gas albedo on the core boundary; 
		  activated for isngcore=0 and pumped flux is
		  0.25*vtg_bar*ng*(1-albedoc), with albedoc user input.
		- Implemented a neutral "recycling" model at the core boundary
		  to account for the fraction of neutral flux into the core
		  that comes back as ions (should be unity).  This neutral
		  recycling coefficient, recycc, is activated if isnicore=3.
		  Note that all other recycling coefficients describe the
		  opposite process where an ion flux at the wall is returned
		  as an inward neutral flux.
		- Added an mhdgeo=2 option for the circular annulus geometry 
		  that is used for efficient BOUT/UEDGE coupling runs; these 
		  core-only cases treat ix=0 and ix=nx+1 cells as dummies and 
		  do not make the usual assignment of ex(0,)=ex(1,) nor 
		  ex(nx,)=ex(nx-1,).
		- Added two new variables to control reading the atomic physics
		  data for hydrogen (newaph=1 is the default) and newapi for 
		  impurities; prevents problem of too many files open for
		  time-dependent runs under PYTHON.
		- Added a switch (itrap_negni) to trap negative ni condition
		  with an error message rather the "yuck".
		- Added a coefficient (alftng) to turn on the neutral thermal-
		  force coefficient (Helander's value is 0.24, but default=0).
		- Added user-specified volume gas source, psgov_use [1/m**3 s].
		- Modified the BASIS-replacement routine basfilex in 
		  com/dummy_py.f to return the input argument instead of 
		  nothing, which fixes a problem with reading atomic physics 
		  data in PYTHON version.
		- For the PYTHON version, users must specify the directory name
		  aphdir in their input file with its full path.
		- Added variables rs_com and zs_com to flx.v to pass output of
		  subr. findstrike via common for python (argument list also
		  works for BASIS version).

Version 4.40  Corresponds to CVS archive of 08/05/06.  Changes include:
		- Removed remanent test xlinc<20 and yinc<4  for Jacobian
		  calculation; can interfere with Jacobian testing, but
		  had no effect on normal running.
		- Include the effect of neutral-neutral collisions in the 
		  definition of the neutral particle diffusivity by using 
		  rnn2cx in the definition of nucx (previously, rnn2cx term 
		  only included in thermal and viscous coeff).
		- Include exponential power factors flgamtg and flgamvg in 
		  the flux limit expression for the thermal and viscous 
		  diffusion coefficients.
		- Added the scaling factor cfloxiplt that multiples the 
  		  neutral convective energy component at the divertor plates 
		  through floxi.  Since neutrals move away from the plate, 
		  the energy contribution from them into the plasma should
		  be zero, and cfloxiplt=0 produces this effect (but initial 
		  default=1 for backward compatibility).
		- Added two indice arrays iym1a and iyp1a to carry the 
		  (trivial) values of iy-1 and iy+1 over the 2D domain.  
		  Used as a variable for the MDSPlus data-tree only.
		- Added a floor-density of 0.1*ngbackg for the isngcore=2 
		  option, which sets the neutral density on the core 
		  boundary using the sqrt(l_ioniz*l_cx) criterion.  The 
		  addition prevents negative boundary neutral densities.
		- Include the factor csfacti as a multiplier of ti in the 
		  definition of the plate Bohm speed, cs.
		- Changed the if-test on isphion to a test on  
		  isphion+isphiofft so as to calculate the poloidal 
		  electric field when the potential is temporarily off.  
		  This fixes a problem with the Jaacobian when trying to 
		  run with fixed potential.
		- Included MDSPlus configuration file linux_mds for MIO 
		  in directory uedge/builder/std; also copy packages_mds
		  to packages file to include MDSPlus in UEDGE build.
		- Introduced the spatially-dependent eqn-on-off arrays 
		  isnionxy(ix,iy,ifld), isuponxy, isteonxy, istionxy, 
		  isphionxy, and isngonxy. This has resulted in many small 
		  changes to if-tests for switching equations on and off.
		  User-specified arrays isnioffxy(ix,iy,ifld), isupoffxy,
		  isteoffxy, istioffxy, isphioffxy, and isngoffx together 
		  with the original eqn-on-off arrays isnion, etc. to set 
		  isnionxy, etc.  That is, isnion still functions as before, 
		  and the user can set isnioffxy(ix,iy,ifld) to unity to
		  turn of ni in specific spatial regions.

Version 4.41  Corresponds to CVS archive of 10/18/06.  Changes include:
		- Corrected some read out-of-bounds problems with help of
		  M. Rensink associated with the 4.40 version (isnioffxy) 
		  and more long-standing, e.g., the spline routine for the 
		  mesh generator.
                - Corrected another potential read out-of-bounds in 
		  subroutines nphygeo and nonorthg reported by UCSD where 
 	 	  ixlb(2) was used in an if-test if(nxpt=2 .and. ixlb(2)); 
		  the INTEL compiled code searches for ixlb(2) even when 
		  nxpt=1 whereas Portland compiled code tests the first 
		  condition (false) and doesn't cause a problem; use 
		  ixlb(nxpt) fixes this INTEL problem.
		- Documented more fully that the parameter MXMISO that sets
		  the maximum number of impurity isotopes must be set 
		  consistently in BOTH uedge/api/api.v AND uedge/com/com.v.
		  Presently, the default is MXMISO=5.
		- Added MIO configure files to uedge/builder/std provided and
		  used by UCSD for their Intel compiler.
		- Made adjustment relative to diffusive neutral model for
		  hydrogen (rarely used); corrected dimension of cngmom & 
		  cmwall from ngspmx to nispmx, and use cmwall(ifld) instead 
		  of cmwall(1) for loss term of radial momentum, but only for 
		  diffusive hydrogen neutrals; only cngmom(1) and cmwall(1)
		  should ever be non-zero.
		- For spatial variation of diffusion coefficients using
		  isbohmcalc=3, change exponent inbtdif (D,chi~1/Bt**inbtdif)
		  from an integer to a real, allowing incremental variation.
		- Added the factor cftaud for scaling the hydrogen-impurity
		  drag time for the impurity force-balance equation.


Version 4.41_mds_shl  Identical to Version 4.41 above but has packages mds  and shl added.