CVS log for ccp4/src/fhscal.f
	Help

Request diff between arbitrary revisions / Display revisions graphically

update licence banners

add new licence banners

changes from Clemens

report dataset id, rather than sequence number

generalise numbering of csetid
replace LRID by LRIDX (for diagnostics only)

CV- one initialisation

CV - one initialisation

trailing spaces (CV)

increase in the maximum number of derivs to 20

lwarn duplication corrected to allow g77 compilation

Apr 2000 by CCB (Based on SCALEIT mods by PJ)
Added AUTO keyword to allow automatic scaling of related datasets.
Gives warning if inconsistent scaling is used for related datesets.
Extended group of recognised data types to include intensities
(reduced maximum number of datasets to 13 to stay under the 200
column limit currently used in mtzlib)
Additional parameters added.

log output html-ised

Richard Kingston: added F+/- columns to LABIN. "Just as for DANO,
the F+ and F- columns are not used in the scaling process, but the scale
is applied to them if they are specified."
MDW: corrected for new column types

bug fix from Ian T

Updated error msgs to go through CCPERR level 2

IJT: now has an option "CENT" which uses only centrics for scaling,
which may be more accurate in space groups with a lot of centrics
such as tetragonal

LAbels changed from DANO to DPH. Div by zero error trapped when calc. V factor.

Stats. now the same for NaNed and unNaNed data. Also, LIST can be any no.
without crashing.

input wouldn't terminate on EOF

Warning messages printed only once.

Missing number stuff tightened up.

Missing number stuff improved. RESO card added to apply cutoff to output
data.

EJD missing data stuff -- looks incomplete.

If F is specified with no SIG program will terminate.

IJT: I noticed that fhscal was giving slightly different answers when I ran
it with several derivatives at once, compared with individual runs on each
derivative, and looks like I introduced a bug when I made fhscal handle multiple
derivatives.  It doesn't affect the scaling, only the calculation of the
pseudo cell volume for vecref.

fix ccprcs arg

IJT: it seems that the data reduction programs can't be relied on to
produce sensible standard deviations, so I decided to change the default
setting in fhscal so it assumes the worst.

IJT: scale up to 20 datasets

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 END keyword, initialise LIST variable.

version pulled straight from york

orrect o/p lines >132 chars

copyright

IJT: I found some bugs in FHSCAL & VECREF to do with calculation of
occupancies when lattice is centred.  Fortunately effects of bugs exactly
cancel! (which is why I didn't spot it before!)  Therefore provided
corresponding versions of programs are used together, the result will be
the same.

IJT fix (for column label assignments?)

IJT: this had a factor of 0.01 to compensate
for the same factor in fft.  This has now been removed, I see, so the factor
is no longer correct in fhscal.  Also corection for low resolution cutoff.

portability

prior to proper attempt at version control

Initial revision