CVS log for ccp4/src/mlphare.f
	Help

Request diff between arbitrary revisions / Display revisions graphically

yorvic -> ysbl

update licence banners

Lattman misspelled

.

move ENDIF statement -bug 523

updates to mlphare.f from EJD - bugs 501,509

add new licence banner

minor format error

changes from Clemens

ejd - error in "occupancy atom" lines for final cycle

incorrect calls to CCPERR

handling of crystal level

increased width of floating point format from 7 to 9 at request of EJD

CV - removal of save and some lines close to 72chars long

removal of SAVE and edit some lines close if not 72 chars long - CV

1. Improved messaging when terminating due to anom occ being zero for a
   site when aocc refinement is requested.
2. Fixed character strings broken across multiple lines (gives warnings on
   OSF1 compiler).

Substantially updated version from EJD:
1. Fewer graphs by default, with improved labelling
2. ISOE/ANOE now taken (correctly) from last REFINEMENT cycle
3. Checks for two input sites being very close (prevents failures after
   RANTAN/SHELX/ACORN if peak search puts a site close to its symmtery
   related neighbour)
4. Check if anom refinement is requested but all AOCC are zero; and if
   OCC is non-zero when only anomalous diffs are used
5. Outputs occupancies with an atom number (useful if checking large no.s
   of sites)
6. Fixed a buglet (last cycle failed to give complete stats for FOMs etc).

Parameterised maximum number of lines available for storing script in
memory, and trapped for exceeding this number.

fixes from ejd.... sorry details were in Dec and now lost....

Fixed bug when converting aniso B-factors for heavy atoms.

EJD: option to turn off phasing for a derivative (sites) is overridden.

Missing comma in format statement.

Format missing from a write statement (Clemens Vonrhein).

failed when trying to refine aniso Bs - ejd fix

x changed to 1x in format statement

Replaced references to non-existant s/r's M3MUL and M3MULT in EXTERNAL
statements with MATMUL and MATMULTRANS (broke on Alphas)

Substantial changes to output formatting but no change to content.

Some parameters moved to mlphare.fh

Removed subroutines M3MUL and M3MULT, replaced with calls to MATMUL and
MATMULTRANS respectively (in modlib).

Fixed broken DCYCLE REFINE ... option - previously no refinement took place
unless PHASE ... option was also specified.

MDW: Changes for possible wavelength in dataset header

Changes for new DCELL record in MTZ header.

removed common block CLEAR and initialised the values the hard way!
this was causing a memory conflict under NT

changes from ejd

EJD: minor change to a warning message.

Harvesting changes added.
Symbols FLAGS and ATSYM renamed to avoid clashes.

Added COORDS keyword to write orthogonal heavy atom coordinates to
pseudo-pdb file.

Standardised resolution label for loggraph

Now also gives sum of CENTRIC and ACENTRIC reflections in table of FOM
vs resolution (EJD).

Correction to previous change.

Changes to copy dataset info to output

Fix from P.Emsley: recognises first 4 letters of BFAC subkeyword of ATOM.

Declare CCPNUN, correct FORMAT string.

Included Maria's message to explain "MATSOL DET LESS THAn 10E-30", from EJD.

Total for FH_a_imag was incorrect.

Typo in LWSYMM arguments.

Changed mlphare to be more like sfall when using ALLIN. Not sure if all
the problems are solved.

EJD: write ATOMi to script

SDSUM not updated properly if no anomalous data.

MDW: APPLY keyword - apply refined scales to derivative data.
     Trap using DFLAG variable wrong input order of keywords.
     MaxNumDerivatives increased to 20 (matches SCALEIT/FHSCAL).
         Means changes to LABIN, ADATA, BDATA.
         MaxNumRefVariables correspondingly increased to 1470.
ACR: NPARS increased to 360 - problem with excessively long LABIN lines.

Was writing out FOMs slightly greater than 1.0
Also problems with output of PRINT RMSF RMSE

.

s/r SFREAD moved to rwbrook.f
EJD: some corrections havent been incorporated..

Increased PARSER LINE to 1200.

Fix from CECCARELLI Chris: refused to include in phasing and refinement cycles any
reflections with negative anomalous differences.

files CLOSED.

CCPNUN not declared. Caused problems with AIX.

Prevented H characters being continued on  next line. Stops Alpha compiler
infos. Also, variables were not initalised correctly in mail loop.

Magic number stuff tightened. Removed some unused variables. Will now produce
NaNs and carry over all refls.

WARNING: can now produce NaNs York changes. Changed some knobby code.
Tidied up.

Add #!/bin/sh to top of silly script, remove leading space on
here-terminator; comments about risks.  (This stuff should be done
shell tools...)

Back out change to 1.54 which can put NaNs in o/p and screw other things.

Avoid aliasing problem in call to BTOANG.

EJD: didnt set Missing value properly for FOM PHIB

EJD Mod SFREAD to cope with comments in atomsf.lib.

EJD: remove debugging (?) write.

Changed output description of real and recip. matrices.

EJD missing data changes safety (?)

EJD: make sure Angstrom coordinates are in PDB format - Orthogonal code 1.
    ( Incidently changed IPRINF to ISYSW - output a mess - some ISYSW (=6)
      Some WRITE(6, .....)

Avoid NaN lossage.

Variable names have changed completely. Magic number routine added. Resolution
in Angstroms added to tables. SCALE FP option removed. Alteration in double
precision variables. SCRIPT keyword added. Other minor changes.

Phasing power was 0.0 when only had PRINT AVE. Now corrected.

TQL23 incorrectly called.

increase parser buffer for big labins (PRE)

Errors in unassigned Sigma labels even clearer.

Error trap for 'ALLIN' incase some input labels are same as program labels.
There are no problems if the columns are expressed explicitly.

HL, FHn and FHAn have been properly initialised; following problems with
reflections were FP=0 and FPHn=0.

'ALLIN' should now work with new MTZLIB.

Declaration standardised, tidied

correct last change

use TITLE

replace general eigenroutine by specialised 3x3 one

comments only

.

EJD: Bad bug for MAD data.  It assumed it could estimate the centric
phase from the the first FH.  This DOES NOT WORK if your first FH = 0.0.

Add/change history info writing for a bunch of programs.  Mostly call the
new LWHSTL.  Several previously called incorrectly LWHIST with scalar
instead of character array.

Add history info to hklout

EJD: Minor corrections - Aniso Eigen vectors not printed.

increase MCYCLE from PRE

o/p cosmetic

remove PHARE title

From Doug Ohlendorf <OHLEN@dccc.med.umn.edu> via EJD to handle 20 derivatives.

PRE: increase MCYCLE to 30

EJD: Tidying.  Make sure all reflecions written out even if not phased.
     Labels wrong for 5th 6th deriv..     Extend format for Cullr etc.

version pulled straight from york

.

EJD tidy up o/p

EJD: long-standing error in figure of merit table

EJD: Now it doesnt need
a special treatment for FP /FPHi labels

EJD various small corrections

copyright

EJD change (divide-by-zero fix?)

small change from EJD

format fixes (missing leading space)

EJD has changed all sqrt() to sqrt(abs()) after getting a sqrt of -ve
number error.

EJD:      Stop mlphare applying changes to coordinates on last cycle.
 Wrong thing to do..

*** empty log message ***

*** empty log message ***

new version from york:
*  Dave Schuller complains about documentation for using prior
  phase info in MLPHARE.
  1) Note WC not used at all. oops..
*  Considerable changes to MLPHARE ouput.

EJD: Serious error in applying scales to derivatives.
Applies SCALES to FP as well as FPH - change order for parsing.

typo

ccprcs call was wrong

version from york with continuation cards limited to 19.
not sure exactly what's supposed to have changed.

fix ccprcs call

*** empty log message ***

correct calc. of Hendrikson--Lattman coeffs for centric terms in hlcoeff

*** empty log message ***

something went wrong at r1.5 -- lost LMB changes.

uninitialised variable

stabilise 1.4.2 branch

previous patch was wrong -- back out and do differently (basically portability)

DL portability patches

fetched from LMB

EJD error correction

from EJD: anomalous parameter correction.  Also now outputs zero FPs.

`fix for Venki' from EJD

prior to proper attempt at version control

Roughly correposnds to first `official' release of MTZ version for un*x
and VMS around 14--15/5/92.

Initial revision