path: root/noao/onedspec/Revisions

.help revisions Jun88 noao.onedspec
.nf

splot/eqwidthcp.x
    The 'sg' and 'lg' pointers were allocated as TY_REAL and re-allocated
    as TY_INT (5/4/13)

splot/anshdr.x
    Added a F_FLUSHNL flag to the logfile descriptors to flush the 
    data as it is written (5/12/12, MJF)

splot/spdeblend.x
    Added overplotting of individual components.  (12/5/11, Valdes)

====
2.16
====

hireswcal.cl
    A script I wrote for Simon Schuler to apply the wavelength calibrations
    that he got for hires spectra from Geoff Marcy.  This is put here for
    safe keeping.  (11/16/11, Valdes)

t_deredden.x
    There was a small error in coding the formulae of Cardelli, et al.
    (6/28/11, Valdes)

splot/spdeblend.x
    An integer allocated array was being freed as a real array causing
    an error with deblending on 64-bit systems.  (6/6/11, Valdes)

=======
2.15.1a
=======

======
2.15.1
======

scombine/icgdata.x
    This code uses the SX array to pass in mask values.  The code, taken from
    an old version of imcombine, treats the mask array as integers.  All the
    dereferencing of this mask array were changed to reals.

scombine/t_scombine.x
    The SX pointer (allocated as real) was being used in a Memi, causing
    a segfault on some 64-bit systems (4/3/11/, MJF)

specplot.h
    Fixed improper use of P2R in macro. (3/31/11, MJF)

continuum.par
odcombine.par
sfit.par
splot.par
    The prompt strings which said the "grow" parameter is in pixels were
    changed to remove this.  The grow parameter is in user coordinate units.
    (6/28/10, Valdes)

=======
V2.14.1
=======

t_specplot.x
specplot.par
doc/specplot.hlp
    Added a new "transform" parameter to allow scaling the spectrum pixel
    values.  Currently on "log" is implemented.  (1/5/09, Valdes)

doc/splot.hlp
    The description of the 'e' key incorrectly said a core flux is
    output.  (8/22/08, Valdes)

dispcor.par
    Changed "Conserve flux" to "Conserve total flux" per user request.
    (6/13/08)

rspectext.cl
    Added "addonly" values to all hedit commands.  (4/1/08, Valdes)

odcombine/t_odcombine.x
    Fixed some procedure calls being closed with a ']' instead of a ')'
    (2/17/08, MJF)

ecidentify/eclinelist.x
    A check is made that the second closest match has a match distance
    more that 25% greater than the nearest match.  (12/10/07, Valdes)

t_dopcor.x
    For clarity when the velocity applied is in km/s this is used in the
    log and in DOPCORn keywords.  (12/10/07, Valdes)

=====
V2.14
=====

doc/standard.hlp
    Clarified the equations and formating.  There was an inconsistency
    between the Vega flux and magnitude given in the text.  Either value
    could be changed but the old version of this task was based on the
    specified Vega flux so the magnitude was changes (0.048-0.0336)
    to make the description consistent.  (4/3/07, Valdes)

splot/wrspect.x
    For log sampled data (dc-flag=1) that also has a ltv offset the
    new WCS for the output was wrong.  (12/4/06, Valdes)

smw/shdr.x
    The shdr_rebin procedure was rebinning the target spectrum to its natural
    units while the reference spectrum might be in different units.  The
    correct thing to do is rebin in the reference units.  (10/27/06, Valdes)


=======
V2.12.3
=======

splot/splot.x
    Needed to initialize the aperture so that the aperture selection
    behavior is consistent when calling splot more than once.  (5/16/06, Valdes)

doc/specplot.hlp
    Added a quick example illustrating using batch mode plotting.
    (5/2/05, Valdes)

t_specplot.x
    To support cursor file input without including the x, y, wcs fields
    the call to sp_nearest was modified to avoid floating exceptions.
    (5/2/05, Valdes)

t_sarith.x
    Error handling was improved for onedspec output.  When no pixels
    were selected by using w1/w2 the warning was reported but then
    a segmentation error would occur when trying to close the output
    image.  (3/8/05, Valdes)

t_tweak.x
doc/telluric.hlp
    The normalization used is now printed.  (1/12/05, Valdes)

t_dopcor.x
doc/dopcor.hlp
    A keyword is added to log the operation.  (10/29/04, Valdes)

t_deredden.x
    Adjusted the error reporting to print the warning before closing and
    deleting the output image.  (9/27/04, Valdes)

ecidentify/ecfitdata.x
    The EC_FITPT routine used the wrong pointer for the physical to logical
    coordinate conversion.  (9/10/04, Valdes)

odcombine/		+
doc/odcombine.hlp	+
x_onedspec.x
onedspec.cl
onedspec.men
onedspec.hd
mkpkg
    Added a new task ODCOMBINE which is layered more directly on the
    source for IMCOMBINE.  This version supports bad pixel masks as well
    as most of the new features of IMCOMBINE.  (6/21/04, Valdes)

scombine/x_scombine.x	+
scombine/mkpkg
onedspec.cl
mkpkg
    Packaged SCOMBINE as its own executable noaobin$x_scombine.e.
    (6/21/04, Valdes)

t_standard.x
    In case someone puts the query parameter "star_name" on the command
    line and the calibration cannot be found the task will not only try
    twice before aborting rather than go into an infinite loop.
    (5/21/04, Valdes)

dispcor/t_dispcor.x
dispcor/dispcor.x
dispcor.par
doc/dispcor.hlp
    Added the new parameter "blank" to control the output values when there
    are no input values; i.e. the out of bounds values.  (5/18/04, Valdes)

dispcor/t_discpor.x
    For 2D spectra the "global" option was not working.  (5/14/04, Valdes)

scombine/t_scombine.x
    A check for the input format was added.  If the input is 2D spectra
    then a format error is printed.  (5/14/04, Valdes)

identify/autoid/autoid.x
    Changed "AID_NT(aid) = min (2 * AID_NR(aid), AID_NTF(aid))" to
    "AID_NT(aid) = AID_NTF(aid)".  While this may have speed consequences
    it avoids preselecting target lines.  (4/23/04, Cooke, Valdes)

doc/scombine.hlp
    Made corrections suggested by Francois Schweizer on 2/25/04.
    (3/10/04, valdes with input from schweizer)

t_sarith.x
    When the task was modified 8/3/02 to add something to the WCS the
    wrong pointer was used resulting in a segmentation violation when
    using the "merge" option.  (3/9/04, Valdes)

=======
V2.12.2
=======

identify/autoid/aidlog.x
    Added a test to avoid an arithmetic error if the dispersion
    turns out to be zero.  (1/29/04, Valdes)

identify/autoid/aidautoid.x
identify/autoid/autoid.x
    Two new debugging characters, "nm", were added.
    (1/29/04, Valdes)

aidpars.par
    Default values were changed:
	cddir: "unknown" -> "sign"
	ntarget: 30 -> 100
	ndmax: 20 -> 500
	fmatch: 0.3 -> 0.2
    (1/29/04, Valdes)

identify/autoid/autoid.x
    The parameters that can be specified by header keywords can now be
    either the keyword or the keyword prefixed by '!'.  This was done
    because there are a number of other IRAF tasks that use the '!' prefix
    and users may be confused and use this syntax.  (1/29/04, Valdes)

identify/autoid/autoid.x
aidpars.par
    The number of highest vote potential dispersions checked was
    previously limited to a maximum of three times the number of target lines.
    Now the number may be as large as specified by the "ndmax" parameter.
    The default parameter value was greatly increased to 500.
    (1/29/04, Valdes)

identify/autoid/aidautoid.x
doc/aidpars.hlp
    The algorithm was modified to iterate on the pattern parameter
    "npattern".  After exhausting the search with the initial
    number of lines per pattern the value is reduced successively by
    one down to the minimum of 3.  This makes the algorithm take longer
    but the search is more exhaustive.  Use of larger patterns initially
    allows finding fewer and more likely candidates first to speed
    a solution.  (1/29/04, Valdes)

aidpars.par
identify/autoid/autoid.x
doc/aidpars.hlp
    The "rms" parameter is now specified in units of the "fwidth"
    parameter rather than in pixels.  This is because if fwidth is
    made larger to deal with broad lines (i.e. a wide slit) then the
    expected uncertainties in pixel centroids will be larger.  The
    default value was changed from 0.3 pixels to 0.1 of fwidth.
    (1/29/04, Valdes)

identify/autoid/autoid.x
doc/aidpars.hlp
identify/id_peaks.x
    The selection of target lines was changed from using id_peaks to a
    new routine id_upeaks.  In the former routine the ntarget strongest
    peaks are selected regardless of position in the spectrum.  But this
    can result in no lines being used in some parts of the spectrum if
    the spectrum is dominated by strong lines in just one part of the
    spectrum.  The id_upeaks routine finds lines over the whole spectrum
    by dividing the spectrum into regions and then alternatively selecting
    the brightest line in each region until the desired number of lines
    is obtained.  In this case the number of regions if hardwired at 5.
    (1/29/04, Valdes)

aidpars.par
identify/autoid/autoid.h
doc/aidpars.hlp
identify/autoid/aidinit.x
identify/autoid/aidset.x
identify/autoid/autoid.x
    Added two new parameters to aidpars.  The first is "maxnl" which
    defines the maximum non-linearity to accept after a dispersion function
    fit.  Previously the maximum was hardwired in the code to be 0.5%
    which was too small for many applications.  The default is set at
    2%.  The second new parameter is "crquad" which defines a quadratic
    correction to the pixel positions of detected lines in order to
    "linearize" the pattern of line spacings which are matched against
    the coordinate list.  This was found to not be as important as the
    "maxnl" limitation in handling non-linear dispersion and has a default
    value of zero.  (1/29/04)

sensfunc/sfstds.x
    Added a check for names with a kernel section.  Specifically, names
    that end in ']'.  (12/18/03, Valdes)

smw/shdr.x
    Now if CUNITn is specified in velocity (m/s or km/s) and if
    CTYPEn is VELO (or VELOCITY) then internally the velocity zero point
    reference of 21 centimenters will be automatically added.
    (8/19/03, Valdes)

smw/shdr.x
    Experience has shown that data with no units that users want to
    import is mostly in Angstroms.  So rather than use the old FITS
    standard that units are meters it will now assume Angstroms.
    (8/15/03, Valdes)

smw/smwaxes.x
    A check is made if the physical axis is ra or dec in which case the
    image is considered not to be dispersion corrected.  (8/5/03, Valdes)

smw/smwsctran.x
    If there is an error the physical coordinate system is used instead of
    the world coordinate system.  This is meant to allow coupled WCS
    (particularly celestial WCS) to be used without an error.
    (8/5/03, Valdes)

t_dopcor.x
    Moved the erract before the imunmap/imdelete to produce the correct
    error message.  (7/8/03, Valdes)

dispcor/t_disptrans.x
    When the subroutine dispcor was modified with an extra arguement this
    task was not modified.  The extra argument was added.  (6/4/03, Valdes)

identify/idinit.x
    The restore function was not resetting the shift value at the right
    time.  This had the effect of causing the user shift to be wrong
    in the REIDENTIFY output when refit=no.  (5/27/03, Valdes)

t_tweak.x
    The statistics computation is now relative to neighboring points.
    This change was developed working with the Coude-Feed spectral atlas
    pipeline.  (4/3/03, Valdes)

t_tweak.x
    WHen an error, such as calibration values too low, occurs in twk_fit it
    would exit without closing the graphics.  Now the graphics is opened with
    AW_DEFER and the error action is to first close the graphics before
    returning to the calling routine with the error.  (2/28/03, Valdes)

t_tweak.x (Bug 520)
    The erract was after error cleanup which could cause the incorrect
    error to be reported.  Since the error action is to WARN it makes sense
    to immediate report the error and then do the clean up.  (2/24/03, Valdes)

dispcor/refspectra.x
dispcor/refgspec.x
dispcor/refspectra.com
    The "select" parameter is now included in the common so that
    refgspec does not try to look for the sort and group keyword if
    it is not needed.  (9/5/02, Valdes)

t_sarith.x
    When a new MWCS pointer is created the attribute "sformat" has to
    be added.  (8/3/02, Valdes)

t_fitprofs.x
    Incorrectly used pargi instead of pargb.  (8/2/02, Valdes)

splot/getimage.x
    Gt_setr was being called with an integer argument.  (8/2/02, Valdes)


splot/eqwidthcp.x
    1.  A pargd was used with the real variable cont.
    2.  Variable pi never used.
    (8/2/02, Valdes)

scombine/icscale.x
    Needed to dereference the error string in icgscale.
    (8/2/02, Valdes)

t_dispcor.x
    String argument was incorrectly given as NULL.
    (8/2/02, Valdes)

t_dispcor.x
dispcor/dcio.x
sensfunc/sfoutput.x
    The axis variable is not used and was deleted.  (8/2/02, Valdes)

=======
V2.12.1
=======

scombine/t_scombine.x
doc/scombine.hlp
    The rejection option is ignored when "combine=sum".  The documentation
    did not make this clear and the output log would show whatever was
    set for the rejection parameter.  The help was clarified and the
    code changed to show a rejection of none.  (7/5/02, Valdes)

irsiids/libpkg.a
    Removed a stray libpkg.a link (6/5/02, MJF)

identify/idcenter.x
ecidentify/eccenter.x
    The handling of INDEF values between reals and doubles was not done
    correctly.  (5/28/02, Valdes)

=====
V2.12
=====

t_slist.x
irsiids/t_slist1d.x
    The SMW pointer was being closed without resetting the pointer in
    the SHDR structure.  (5/30/02, Valdes)

identify/iddofit.x
identify/idinit.x
    When deleting features during the fitting the memory allocated for
    the labels was not being updated correctly.  Also in freeing the
    feature memory there was no need to free labels beyond the number
    of features.  (4/30/02, Valdes)

splot/wrspect.x
    Removed possibility of an infinite loop and make error checking
    a little more obvious.  (2/28/02, Valdes, 3/21/02, Valdes))

dispcor/dcio.x
    Removed the special test for dispersion corrected data with no DCLOG
    keyword which prevents re-calibration.  (2/4/02, Valdes)

doc/bplot.hlp
    Fixed a typo in the help page. (1/5/02, MJF)

mkpkg
    Added missing <mach.h> dependency to getcalib.x  (12/13/01, MJF)

sensfunc/mkpkg
    Removed unneeded dependencies for sfimage.x  (12/13/01, MJF)

irsiids/t_flatfit.x
irsiids/t_subsets.x
irsiids/t_sums.x
t_mkspec.x
t_sinterp.x
irsiids/t_bswitch.x
    imgl1r() called with extra arg. (9/20/01, Valdes)

dispcor/t_dispcor.x
    dc_gec() missing arg. (9/20/01, Valdes)

identify/autoid/aidautoid.x
    aid_eval() called with extra arg. (9/20/01, Valdes)

identify/idinit.x
    id_gid() define as a function but should be a subroutine.
    (9/20/01, Valdes)

identify/idshift.x
    id_getid() called as subroutine. (9/20/01, Valdes)

ecidentify/ecffit/ecfshift.x
    ecf_pshift() was incorrect. (9/20/01, Valdes)

identify/iddofit.x
    When removing lines deleted during fitting the labels were not being
    correctly maintained.  (8/2/01, Valdes)

splot/eqwidth.x
    Instead of refusing to compute errors when there is a negative value
    (anywhere in the spectrum and not just in the region) the routine
    simply sets the pixel value in evaluating the sigma.
    (5/16/01, Valdes)

doc/deredden.hlp
    Added information about the range of validity of the extinction
    function.  (4/9/01, Valdes)

t_fitprof.x
    The input image was being unmmaped before the output image which
    can cause problems.  (3/9/01, Valdes)

splot/getimage.x
    The nline parameter was not being set to the current line which could
    cause the ')' and '(' to misbehave.  (2/15/01, Valdes)

doc/sarith.hlp
    The help said incorrectly that flux conservation was used.
    (1/17/01, Valdes)

smw/smwsaveim.x
smw/smwequispec.x
    The APNUM keywords change to AP when the aperture number is greater
    than 999.  (11/8/00, Valdes)

identify/iddb.x
    The user units string is now recorded.  This is to allow velocity
    units to include the reference point.  (5/4/00, Valdes)

splot/eqwidthcp.x
    When x becomes large enough the parabola fitting routine has a
    divide by zero.  The parabola fitting routine was converted to work in
    double, to check the x values for degeneracy, and to avoid the squares
    of large numbers.  The calling routine was also modified to work in
    double.  (2/15/00, Valdes)

smw/smwonedspec.x
smw/smwoldms.x
    Put an error check on imdelf.  The list expansion should have expanded
    to only the header keywords present but for some undiagnosed reason
    sometimes the list expansion returned a non-existant keyword which
    wouod cause an error.  (2/2/00)

dispcor/dispcor.x
    The input linei argument should really be the aperture number needed
    to go back to pixel coordinates.  The change was to use this value only
    to flag if the WCS is 1D (for long slit data) and otherwise to do
    what the procedure did before.  (2/1/00, Valdes)

dispcor/t_dispcor.x
    During changes for 2D dispcor the arguments for a routine changed and
    this was not reflected in the dc_global routines resulting in a
    segmentation violation when the global option is selected.
    (1/27/00, Valdes)

dispcor/dcio.o
    Touched but not changes.  (1/27/00, Valdes)

getcalib.x
t_standard.x
standard.par
lcalib.par
splot.par
doc/standard.hlp
doc/lcalib.hlp
doc/splot.hlp
    1.  The calibration files may now be blackbody curves scaled to a
	specified magnitude.
    2.	STANDARD displays the data and bandpasses in the units of the
	data rather than Angstroms.  The changes were made to allow
	output in other untis but for now this has been disabled.
    (1/24/00, Valdes)

t_sarith.x
    The WCS was set wrong when copying/extracting a region of a
    log-linear spectrum.  (12/14/99, Valdes)

=======
V2.11.3
=======

t_fitprof.x
    Fixed double/int mismatch in a min call.  (11/22/99, Valdes)

smw/mkpkg
ecidentify/ecffit/mkpkg
sensfunc/mkpkg
splot/mkpkg
mkpkg
    Added missing dependencies.  (10/11/99, Valdes)

t_fitprofs.x
doc/fitprofs.hlp
    The background region specification was extended to allow a third
    argument as a scaling factor.  (9/22/99, Valdes)

doc/specwcs.hlp
    Fixed typo defining the variable n.  (9/13/99, Valdes)

t_tweak.x
    1.  Changed wavelength evaluations to double precision.
    2.  The normalization step for the rms calculation was removed for
	the case of sky subtraction.
    (9/8/99, Valdes)

dispcor/t_dispcor.x
dispcor/dcio.x
dispcor/dispcor.x
doc/dispcor.hlp
    Now allows NDSPEC format spectra.  (9/7/99, Valdes)

t_fitprofs.x
doc/fitprofs.hlp
    1.  The background region specification was extended to allow taking
	the average or median or a region.
    2.  The image identification label was incorrect.
    3.  If verbose=no there was an attempt to close a non-existent structure.
    (6/26/99, Valdes)

t_fitprofs.x
    Fixed bug which only allowed the last component to be saved to an
    image.  (8/25/99, Valdes)

=======
V2.11.2
=======

smw/units.x
smw/funits.x
    Stripped trailing whitespace from units label.  (8/5/99, Valdes)

identify/t_reidentify.x
    File date changed but no changes made to the file.  (7/22/99, Valdes)

rspectext.cl
    Added explicit add and del parameters to all the HEDIT calls.
    (7/15/99, Valdes)

t_sarith.x
    The option to have a single second operand to work on a set of
    first operands was not working.  (5/28/99, Valdes)

splot/splot.key
doc/splot.hlp
    Added a reference to :.help and :/help.  (5/12/99, Valdes)

doc/sys/1and2dspec.hlp
doc/sys/Onedspec.hlp
doc/sys/Review.hlp
doc/fitprofs.hlp
doc/reidentify.hlp
doc/sensfunc.hlp
doc/aidpars.hlp
doc/autoidentify.hlp
doc/skytweak.hlp
doc/telluric.hlp
irsiids/doc/powercor.hlp
irsiids/doc/widstape.hlp
    Fixed minor formating problems.  (4/22/99, Valdes)

identify/idlinelist.x
    The call to id_peak was using physical pixels while the subroutine
    expects logical pixels.  A conversion from physical to logical was
    added before calling id_peak.  (3/8/99, Valdes)

scombine/t_scombine.x
    Changed  UT(shin) = imgetr (im, Memc[gain -> snoise]).
    (1/29/99, Valdes)

identify/idgraph.x
    Removed violation of GTOOLS data structure.  (12/18/98, Valdes)

identify/t_reidentify.x
    1.	When interactive=yes and ans is not NO when starting on a new
	image in ri_image, the curfit descriptor was initialized to the
	defaults rather than to reference solution.  This was because
	of a missing ic_copy.
    2.  When the reference and image names are the same the task will now
	skip the reidentify,
    (12/3/98, Valdes)

shdr.x
    Needed to check if data is defined for associated types before trying
    to set flux units.  (11/27/98, Valdes)

shdr.x
    Improved the recognition of CTYPE values.  Most notably WAVELENGTH
    is converted to "waveleng" by MWCS in making the label attribute.
    (11/25/98, Valdes)

doc/splot.hlp
    Fixed help that said the output of a long slit or ND image would be
    a 1D image.  (11/18/98, Valdes)

dispcor/dispcor.x
dispcor/refspectra.x
    The weights when weighting multiple dispersion solutions were only
    being recorded in the WCS attributes to 3 significant digits.  This
    could cause the weights to become unnormalized and cause small
    shifts.  Now whenever the weights are converted to strings the
    format is %.8g.  (11/17/98, Valdes)

t_fitprofs.x
    When the input peak value was INDEF the task would fail with a
    floating overflow value if scale < 1.  This was caused by not checking
    for INDEF before dividing by the scale.  (11/5/98, Valdes)

smw/smwnewcopy.x
    The structure copy was wrong.  (10/28/98, Valdes)

sensfunc/sfimage.x
t_calibraate.x
    The flux calibration gets the wrong sign if dw<0.  (9/25/98, Valdes)

smw/shdr.x
    Default units of Angstroms was added if DC-FLAG is dispersion corrected.
    (9/24/98, Valdes)

dispcor/dcio.x
    The weights are now adjusted to produce weighted average rather than
    weighted sum.  (8/25/98, Valdes)

splot/wrspect.x
    The filling in of data outside of NP1/NP2 was done incorrectly.
    Normally NP1/NP2 cover the entire image line but in echelle data
    it is common for NP2 to be less than the full line.  In this case
    the result of saving an image was loss of the last valid point.
    (7/14/98, Valdes)
 
identify/idinit.x
    When restoring a solution without a dispersion function the shift
    failed to be restored.  This causes a problem with REIDENTIFY when
    working on long slit data with a significant systematic tilt and
    measuring the spatial distortion.  (6/1/98, Valdes)

t_tweak.x
telluric.par
doc/telluric.hlp
    If the calibration is < 0 it is detected but there was an error in
    the error clean up giving a "memory corruption" error.  This error
    was fixed and a new threshold parameter was added to allow the task
    to continue if the calibration data has low values.
    (4/21/98, Valdes)

t_sarith.x
    When rebinning non-linear spectra the dispersion type was not be reset
    to linear resulting in an incorrect spectral WCS.  See buglog 400.
    (4/17/98, Valdes)

doc/scopy.hlp
    Added a note about using epar to set nsum to examples in section III
    as suggested by Ivan King on 4/3/98.  (4/8/98, Valdes)

scombine/t_scombine.x
    If the input spectra are not dispersion corrected and first=no the
    task was incorrectly setting the dispersion correction flag.
    (3/3/98, Valdes)

splot/deblend.x
    The dorefit code was not handling the case of a mixture of profile types.
    (2/12/98, Valdes)

t_tweak.x
    The extra argument in a twk_colon call was removed.
    (2/5/98, Valdes)

dispcor/t_dispcor.x
    Added some errchk declarations.
    (1/26/98, Valdes)

identify/idlinelist.x
doc/autoidentify.hlp
doc/identify.hlp
doc/reidentify.hlp
    When a coordinate list is read it will be sorted and identical
    entries will be eliminated.  Thus, line lists no longer need to
    be sorted.  (1/12/98, Valdes)

=======
V2.11.1
=======

doc/splot.hlp
    Added another paragraph and a correction to the flux calculation
    done by 'e'.  (12/22/97, Valdes)

splot/gfit.x
    The test for computing errors when negative data is detected was
    incorrect and would given an error message even when errors were
    not desired.  (10/23/97, Valdes)

dispcor/dispcor.x
smw/shdr.x
    Changed the maximum distance that the endpoints can be from pixel
    edges before using the pixel values directly instead of integrating
    the interpolator from 0.001 to 0.00001.  (10/7/97, Valdes)

doc/calibrate.hlp
    Added brief discussion about pixels falling outside the wavelength
    range of the sensitivity function.  (9/23/97, Valdes)

===========
V2.11export
===========

identify/iddb.x
ecidentify/ecdb.x
    Increased the number of digits recorded in the database for the fit
    and user values to 9.  (8/22/97, Valdes)

swm/shdr.x
    Added an arbitrary reference for the velocity CTYPE value.
    (8/20/97, Valdes)

smw/smwsaxes.x
    The earlier fix for transposed data was incorrect.  The origin terms
    do not need to be changed but the order of the CD matrix terms
    was incorrect.  (8/15/97, Valdes)

dispcor/refnoextn.x
    Added fit and fits to the possible extensions.  (8/14/97, Valdes)

smw/smwsaxes.x
    When the LTERM is adjusted to correct for a transpose only the matrix
    terms were being corrected.  The origin terms also needed to be
    corrected.  (8/6/97, Valdes)

scombine/t_scombine.x
    Previously an end input pixel had to  completely overlap an output pixel
    otherwise it was flagged as missing data.  This was changed to use
    the end pixels if they overlapped at all.  This change was done to
    allow small dispersion shifts to not affect the end point combining.
    (8/6/97, Valdes)

smw/shdr.x
    No change.  (8/6/97, Valdes)

identify/t_reidentify.x
doc/reidentify.hlp
reidentify.par
imred/*/reidentify.par
twodspec/longslit/reidentify.par
    The shift parameter was restored to it's previous usage.  The automatic
    pattern matching algorithm is not selected by setting the shift to INDEF
    and using the new parameter crsearch.  (7/21/97, Valdes)

identify/idshift.x
identify/t_reidentify.x
    The symbol table of reference solutions was being modified by the shift
    calculation causing the loop over solutions to be wrong.  Now
    ri_image marks and frees the symbol table between calls and loops
    through the symbol table solutions in a way that is not affected by
    new entries in the symbol table.  Also idshift marks and frees the
    symbol table.  Note that marking and freeing is not enough because
    the loop using sthead/stnext will not work.  (7/19/97, Valdes)

identify/autoid/aidshift.x
identify/ididentify.x
identify/idreidentify.x
identify/doc/reidentify.x
identify/reidentify.par
    1.	aid_shift was not using crsearch/cdsearch as expected.
    2.  The call to id_shift in the interactive routines had an incorrect
	argument value.
    3.	The help page for REIDENTIFY was clarified about what the shift
	parameter means.
    4.	The parameter prompt for shift in REIDENTIFY was corrected.
	parameter means.
    (7/17/97, Valdes)

identify/t_reidentify.x
    The nlost parameter now applies when not tracing.  (7/17/97, Valdes)

t_sarith.x
    The power option did not work because the apow routine takes only
    integer powers.  Replace the apow routine with an explicit calculation.
    (7/15/97, Valdes)

=========
V2.11Beta
=========

identify/t_reidentify.x
    The number of features was being used to calculate how many features
    might be lost before it was set.  (6/3/97, Valdes)

doc/disptrans.hlp
doc/onedspec.hlp
doc/splot.hlp
    Added new unit abbreviations.  (5/27/97, Valdes)

splot/splot.x
splot/splotcolon.x
splot/splabel.x		+
splot/splot.key
doc/splot.hlp
    Added colon commands for labeling.  (5/16/97, Valdes)

t_tweak.x
    1. The ? help file was specified as .hlp instead of .key.
    2  Add a divide by zero check.
    (5/14/97, Valdes)

t_scoords.x	+
scoords.par	+
doc/scoords.hlp	+
x_onedspec.x
mkpkg
onedspec.cl
onedspec.hd
onedspec.men
    Added a new task that sets a pixel array spectral coordinate system
    in 1D spectra.  (5/9/97, Valdes)

doc/sapertures.hlp
doc/sinterp.hlp
doc/sflip.hlp
doc/disptrans.hlp
doc/skytweak.hlp
doc/telluric.hlp
doc/sfit.hlp
doc/continuum.hlp
doc/fitprofs.hlp
irsiids/doc/slist1d.hlp
    Changed revision versions.  (4/22/97, Valdes)

t_tweak.x	 +
skytweak.par	 +
telluric.par	 +
doc/skytweak.hlp +
doc/telluric.hlp +
x_onedspec.x
onedspec.cl
onedspec.men
onedspec.hd
    Added tasks for tweaking calibration spectra and applying a sky
    subtraction or telluric correction.  (3/28/97, Valdes)

t_sarith.x
    The wrong INDEF type was used.  (3/17/97, Valdes)

smw/shdr.x
    The shdr_linear needed to transform the requested input range to
    the image WCS units.  (3/13/97, Valdes)

t_calibrate.x
t_standard.x
    Changed to allow input spectra in various units.  (3/12/97, Valdes)

autoidentify.par
identify.par
ecidentify.par
identify/t_autoid.x
identify/t_identify.x
identify/identify.h
identify/idinit.x
identify/idgdata.x
identify/iddofit.x
identify/idmap.x
identify/iddb.x
identify/idlinelist.x
identify/idfitdata.x
ecidentify/ecinit.x
ecidentify/ecgdata.x
ecidentify/t_eciden.x
ecidentify/eclinelist.x
ecidentify/ecidentify.h
ecidentify/ecdb.x
ecidentify/ecfitdata.x
dispcor/t_dispcor.x
dispcor/dcio.x
dispcor/dispcor.h
doc/autoidentify.hlp
doc/identify.hlp
imred/echelle/doc/ecidentify.hlp
imred/irs/identify.par
imred/iids/identify.par
imred/kpnocoude/identify.par
noao/lib/linelists/*
    Changes to allow IDENTIFY/ECIDENTIFY to work in user or line list
    specified units and to have DISPCOR pass on the units.  (3/11/97, Valdes)

splot/usercoords.x
    This routine had a couple of places where it calls smw_c?tran? and
    then takes log sampling in EQUISPEC/NDSPEC explicitly into account.
    Since this is now done by the lower level routines the log conversions
    were removed. (3/3/97, Valdes)

dispcor/dcio.x
doc/dispcor.hlp
    The coordinate transformation between logical and world was changed
    to always produce linear wavelength.  It use to be that for
    equispec and ndspec formats with dc-flag=1 this would produce
    log wavelengths.  This prevented resampling from log back to
    linear.  [The date on dcio.x was touched but no actual change was
    made.] (3/3/97, Valdes)

smw/shdr.x
    The SHDR routines that convert between world and logical were updated
    for the changes in smw_c?tran?.  Previously, these routines explicitly
    applied the log transformation for log sample spectra.  For this
    reason most ONEDSPEC tasks operated correctly.  However, DISPCOR
    does not use SHDR_LW/SHDR_WL so it failed.  Now the log conversions
    are done in smw_c?tran? and not in the SHDR routines.
    (3/3/97, Valdes)

smw/smwsctran.x
smw/smwctran.gx		+
noao/lib/smw.h
    The coordinate transformations in ONEDSPEC tasks assume that dispersion
    coordinates are always in linear dispersion whether or not the
    spectra are stored in log sampling (DC-FLAG=1).  However, this
    was not true for EQUISPEC/NDSPEC format.  Now calls to the smw_c?tran?
    routines will return linear dispersion for all supported ONEDSPEC
    WCS types.  This was needed to fix the problem with DISPCOR and
    log sampled input spectra.

smw/smwsaveim.x
    Deleted unused procedure name in errchk.  No functional change.
    (3/3/97, Valdes)

scombine/t_scombine.x
    Moved the call to smw_openim to "clean up" the WCS from after the
    image size is changed to before because otherwise an error would
    occur trying to access aperture information for any new lines added.
    I no longer recall the purpose of this "clean up" step.
    (2/19/97, Valdes)

splot/sumflux.x
    The conversion to "angstrom" units used the wpc instead of abs(wpc).
    (2/6/97, Valdes)

identify/idpeak.x
identify/ididentify.x
identify/idshift.x
identify/reidentify.x
identify/autoid/autoid.x
    Added a new procedure, id_peaks, that replaces calls to find_peaks.
    The new procedure calls find_peaks and then converts the pixels to
    physical coordinates.  (1/30/97, Valdes)

t_calibrate.x
    The calculation of dw for the flux correction used the index i
    instead of k.  This means that dw was constant which is incorrect
    for spectra with non-linear dispersion.   (1/22/97, Valdes)

doc/splot.hlp
splot/autoexp.x
    The intensity range produced by the 'a', ',', '.', and 'z' keys
    could be wrong if the dispersion function was sufficiently non-linear.
    This now fixed and if the positions for the 'a' key are the same
    it autoscales; i.e. 'a' 'a' is a short cut to autoscale.
    (1/10/97, Valdes)

splot/splot.x
    The whitespace was being removed from the units parameter so that
    any units string that requires whitespace (such as "km/s 4000 ang")
    would fail to be recognized.  (12/3/96, Valdes)

scombine/t_scombine.x
scombine/icscale.x
    The feature of getting scaling, zero, and weight values from the image
    headers did not work because the header values were not cached.
    (11/11/96, Valdes)

t_sarith.x
doc/sarith.hlp
    The noise spectrum type is now only copied unmodified.  This is
    a quick kludge until the noise is properly handled.
    (9/11/96, Valdes)

smw/shdr.x
    Added a spectrum type field to the spectrum structure and a specific
    procedure to decode the spectrum type.
    (9/11/96, Valdes)

splot/gfit.x
splot/spdeblend.x
    Allow error estimates with negative pixels if invgain=0 otherwise
    print a warning.  (7/23/96, Valdes)

doc/standard.hlp
    Added comments about proper use of extinction files.
    (6/26/96, Valdes)

doc/dispcor.hlp
    Update the help file to indicate that the input limits are in
    non-log units even with logarithmic sampling is selected.
    (6/14/96, Valdes)

t_standard.x
    Make minor change to beginning of std_flux to avoid optimizer error
    on Solaris with V4.0 compiler.  (6/10/96, Valdes)

t_sarith.x
    Altered order of opening the output so that any error in reading
    the input data is caught first.  (5/14/96, Valdes)

smw/shdr.x
    Separated call for imgs3r in order to error check for failure to
    get the pixel data (such as occurs with a long pathname to the pixel
    file).  (5/14/96, Valdes)

rspectext.cl
    Removed use of CL variable "list".  (5/6/96, Valdes)

noao$lib/smw.h
    Changed SMW_NSPLIT from 200 to 500.  (4/18/96, Valdes)

smw/smwmerge.x
    1.  When the output format is multispec the code did not open a
	single MWCS but simply opened another split MWCS.
    2.  A pointer rather than the string was incorrectly passed to
	smw_swattrs.
    (4/18/96, Valdes)

smw/shdr.x
    If the units are not defined by an attribute a check is made for
    a CUNITn keyword.  (4/17/96, Valdes)

identify/autoid/autoid.x
    1. Removed useless call to id_log.
    2. Fixed realloc bug.
    3. Fixed bug allowing lines to be found multiple times.
    (4/12/96, Valdes)

identify/autoid/aidshift.x
    Added missing argument in call to aid_init.  (4/11/96, Valdes)

identify/idlinelist.x
    Minor efficiency change that avoids extract calls to id_fitpt.
    (4/11/96, Valdes)

identify/t_autoid.x
    Minor bug fix so that log header is printed.  (4/5/96, Valdes)

t_rstext.x
rspectext.cl
doc/rspectext.hlp
    The task now automatically senses the presence of a header.
    (3/7/96, Valdes)

identify/identify.h
identify/peaks.gx
identify/autoid/autoid.x
    1. The ID_FTYPE entry in the structure was being clobbered by a typo
	in the include file which also mapped ID_LABEL to the same location.
    2.	The peak finding routines were modified so that values of INDEF
	for the threshold and contrast would disable these tests.  This
	is needed when absorption peak data is negated to find the
	absorption peaks which are all negative.
    3.  The autoid.x uses of find_peaks were modified to set the contrast
	and threshold to INDEF instead of zero.
    (2/24/96, Valdes)

splot/getimage.x
    There was a bug in initializing the image section limits such that
    when an image section of the form [n,*] is used and there is not
    display limits (xmin and xmax are INDEF) then the plotted spectrum
    will cover the range of the first axis rather than the second.
    (2/22/96, Valdes)

identify/autoid/*	+
identify/t_autoid.x	+
identify/t_identify.x
identify/t_reidentify.x
identify/idcolon.x
identify/iddb.x
identify/iddoshift.x
identify/idfitdata.x
identify/idgdata.x
identify/ididentify.x
identify/idinit.x
identify/idlinelist.x
identify/idlog.x
identify/idmap.x
identify/idshift.x
identify/idshow.x
identify/peaks.gx
identify/peaks.x
identify/reidentify.x
identify/identify.h
identify/identify.key
identify/mkpkg
doc/aidpars.hlp		+
doc/autoidentify.hlp	+
doc/identify.hlp
doc/reidentify.hlp
x_onedspec.x
aidpars.par		+
autoidentify.par
identify.par
reidentify.par
onedspec.cl
onedspec.par
onedspec.men
onedspec.hd
    Added an automatic line identification algorithm.  This algorithm
    is part of the new task AUTOIDENTIFY and modified versions of
    IDENITFY and REIDENTIFY.  A new pset task AIDPARS contains the
    algorithm parameters.  (2/1/96, Valdes)

onedspec.hd
onedspec.men
    Added linelists$README and onedstds$README as the help topics
    "linelists" and "onedstds".  (1/26/96, Valdes)

smw/shdr.x
    When extracting a wavelength range (without rebinning) and with the
    range flipped there was an error in not checking for the existence
    of the associated spectra causing a segmentation violation.
    (1/22/96, Valdes)

specplot.x
    The scale and offset parameters may now be a constant value, an
    @file containing the values, or a keyword name.  (1/13/96, Valdes)

smw/smwopen.x
    The arrays for the aperture, beam, and limits in equispec format were
    not being initialized to reasonable values which could cause an
    error when doing an ES to MS conversion.  Replaced
    mallocs with callocs. (1/9/96, Valdes)

smw/smwesms.x
    Fixed a typo: smwopn -> smw_open.  (1/9/96, Valdes)

smw/smwsaxes.x
    Uncalibrated long slit (2D) spectra which have been rotated are now
    allowed.  The rotated WCS is reset to pixels.  If the dispersion
    calibration flag is set and the spectra have been rotated then
    an error is reported.  (1/4/96, Valdes)

t_sarith.x
doc/sarith.hlp
doc/scopy.hlp
    Preiously both w1 and w2 had to be specified to select a wavelength
    region to be copied or operated upon.  Now if only one is specified
    the second will default to the appropriate starting or ending
    pixel.  (12/20/95, Valdes)

t_sbands.x
    1.  Converted to work in double precision except the spectrum data
	obtained by shdr_open is only in real.
    2.	Increased the index and eqwidth precision printed from
	%7.4g to %9.6g.
    (12/5/95, Valdes)

identify/idgraph.x
    If the graph x window is outside of the data the x window is now
    autoscaled.  This occurs when a user sets window limits in pixel space
    and then does a fit to wavelength.  The new graph was then plotted in
    the windowed pixel space and no data would be seen.  (12/5/95, Valdes)

t_calibrate.x
    The airmass value computed by get_airm was being ignored causing
    a floating exception (bug log 321).  This was fixed.  (12/4/95, Valdes).

scombine/generic/icpclip.x
    Fixed a bug where a variable was improperly used for two different
    purposes causing the algorithm to fail (bug 316).  (10/19/95, Valdes)

identify/peaks.x
    There was an index bug in is_local_max.  (9/26/95, Valdes)

t_slist.x
    Fixed another case of closing the mwcs pointer without invalidating
    it in the shdr pointer.  (9/26/95, Valdes)

t_fitprofs.x
    Added a check and appropriate error message for a missing positions
    file.  (9/22/95, Valdes)

doc/splot.hlp
    Added explicit equations for the quantities measured by the 'e'
    key in SPLOT.  (9/22/95, Valdes)

identify/ididentify.x
idenitfy/identify.key
doc/identify.hlp
    A new key, 'e', has been added to add features from a line list without
    doing any fits.  This is like the 'l' but without the automatic
    fitting before and after adding new features.  (9/5/95, Valdes)

identify/t_reidentify.x
doc/reidentify.hlp
    If there are no reference features the "addfeatures" option will add
    new features before doing a fit or shift.  (9/5/95, Valdes)

splot/getimage.x
    The change of 5/1/95 allows parsing an image section to determine
    the dispersion line.  However this ignored any range along the
    dispersion.  This change completely parses any image section and
    sets the display range in pixels or wavelength to that of the
    image section along the dispersion.  (8/28/95, Valdes)

identify/t_reidentify.x
identify/iddb.x
    1.  If the reference image does not exist REIDENTIFY would catch the
	error but then attempt to close an unopened database leading to
	a seg vio error rather than the warning.  A check was added for
	the database being open.
    2.  Added a new database procedure that scans a database and saves
	the records.  This allows REIDENTIFY to use a reference database
	even when the reference image doesn't exist.
    (8/23/95, Valdes)

smw/smwdaxis.x
smw/shdr.x
    In the absence of DISPAXIS the software will recognize the FITS
    CTYPE keyword with values of LAMBDA, FREQ, VELO*, WAVELENGTH
    in the units defined in the original FITS paper.  (8/20/95, Valdes)

t_sfit.x
    The logic for checking whether all lines and all bands has been done
    is not as straightforward as indicated in the entry of 4/29/94.
    The checking on bands has been eliminated though a record of
    the bands dones is written to the header.  (8/15/95, Valdes)

smw/smwsaveim.x
scombine/t_scombine.x
    When a new image is opened NEW_COPY it inherits IM_NPHYSDIM and IM_NDIM.
    A routine can change IM_NDIM but not IM_NPHYSDIM.  The routine to
    save an equispec WCS needs to preserve the IM_NPHYSDIM when updating
    an exisiting 2D image which may have been specified as a 1D section.
    In order to tell the routine that a new lower dimensional image
    is desired with a NEW_COPY header the higher level routine can set
    the temporary keyword SMW_NDIM and the routine setting up the WCS
    will use this in prference to the IM_NPHYSDIM.  (8/14/95, Valdes)

identify/idlinelist.x
    The way memory was being allocated for labels was such that not
    all memory would be deallocated at the end.  (8/3/95, Valdes)

identify/iddb.x
    When "adding" features the NALLOC value was not properly updated
    resulting in free uninitialized pointers leading to a segvio.
    (8/3/95, Valdes)

smw/smwdaxis.x
    If the image header dispersion axis is unreasonable a warning is
    printed and the "dispaxis" parameter is used instead.  (8/2/95, Valdes)

sbands.x
    Changed the index and eq width format from 7.2f to 7.4g.
    (7/28/95, Valdes)

splot/voigt.x	+
t_fitprofs.x
splot/splot.x
splot/anshdr.x
splot/eqwidthcp.x
splot/gfit.x
splot/deblend.x
splot/spdeblend.x
splot/splot.key
splot/mkpkg
doc/fitprofs.hlp
doc/splot.hlp
fitprofs.par
splot.par
    Added lorentzian and voigt profile fitting and deblending.  This changed
    the FITPROFS parameters and the input line lists for FITPROFS and
    SPLOT though the old line lists will still work.  A new parameter was
    also added to SPLOT and FITPROFS to set the number of Monte-Carlo
    samples used in the error estimates.
    (7/28/95, Valdes)

splot/splot.x
    Changed when the shdr structure is closed to avoid an error.
    (8/24/95, Valdes)

t_sapertures.x
doc/sapertures.hlp
    Modified to allow aperture ID table to be from an image header
    in the same way as done in the APEXTRACT package.
    (7/24/95, Valdes)

t_specplot.x
specplot.key
doc/specplot.hlp
    Added a new key 'f' to toggle between logical pixels and world
    coordinates.  (7/21/95, Valdes)

dispcor/dcio.x
dispcor/dispcor.h
    The application of a shift now also works with non-linear dispersions
    in the input image.  This is a feature used in the DOFIBERS script
    to align sky lines.  (7/19/95, Valdes)

splot/wrspect.x
    The BANDID keyword was being written with garbage characters
    because a pargstr was used instead of pargi.  (7/14/95, Valdes)

dispcor/dcio.x
    When there is only a shift in the database (a feature added 4/21/94)
    and the image has more than one aperture the weight parameter was being
    clobbered causing incorrect results.  (7/13/95, Valdes)

t_sapertures.x
    Fixed the "dtype" parameter behavior which was not correct.
    (6/30/95, Valdes)

smw/smwonedspec.x
smw/smwsaxes.x
    1.	For the simplest spectra a heuristic to determine DC-FLAG was
	added such that if the wavelength of the first pixel and the
	increment per pixel are both unequal to 1 then the spectrum is
	assumed to be dispersion calibrated.
    2.	The label and units are not overridden if either is present.
	If neither is present but the spectrum is considered to be
	dispersion corrected then it defaults to Wavlength(angstroms).
    (6/30/95, Valdes)

ididentify.x
reidentify.x
    When a line center fails to be found with the 'm' key a message is
    printed pointing to the threshold parameter.  (6/30/95, Valdes)

t_sbands.x
    The allocation scheme was incorrect causing a segmentation violation
    after the first 10 bands.  (6/30/95, Valdes)

=======
V2.10.4
=======

t_sarith.x
    The "units_display" WCS attribute is copied if set.  (5/13/95, Valdes)

splot/splot.x
splot/getimage.x
t_specplot.x
    1.  The task "units" parameter value is mapped to "display" if null.
    2.  The units are set with shdr_units.
    (5/13/95, Valdes)

smw/shdr.x
    1.  The spectrum structure is loaded in the image MWCS units ("units").
    2.  The special unit string "display" changes units to the "units_display"
	attribute in shdr_units.
    3.  The special unit string "default" changes units to the image MWCS
	units in shdr_units.
    (5/13/95, Valdes)

doc/sfit.hlp
    Added a description of the "sample" range syntax.  (5/12/95, Valdes)

splot/splot.x
splot/getimage.x
doc/splot.hlp
    Because it can be desirable to use image sections on the input but
    this will cause problems if the user attempts to update the image
    SPLOT was modified to parse the image section for the specified image
    line, column, or band and then map the full image.  (5/1/95, Valdes)

t_sbands.x
doc/t_sbands.x
    Increase the length and changed to g format for the flux so that
    flux calibrated data will print.  (4/12/95, Valdes)

doc/wspectext.hlp
    Fixed typo in example.  (4/12/95, Valdes)

t_sarith.x
    Image extensions are no only stripped for onedspec format output
    images rather than in all image names.  This is necessary to allow
    STF images with explicit extensions not matching the imtype value
    to be specified.  (3/31/95, Valdes)

scombine/icscale.x
doc/scombine.hlp
    The behavior of the weights when using both multiplicative and zero
    point scaling was incorrect; the zero levels have to account for
    the scaling.  (3/27/95, Valdes)

splot/flatten.x
    Removed use of faulty fp_equal test for equality with zero.  This would
    cause continuum normalization to fail for fluxed data.  (2/23/95, Valdes)

sensfunc/sfshift.x
    Deleted points and stars are now ignored in the grey shift calculation.
    (2/22/95, Valdes)

t_sinterp.x
    Updated the image header keywords to give a complete and standard
    linear WCS.  (2/21/95, Valdes)

splot/gfit.x
    If the marked region does not span the profile peak then an pointer
    indexing error occurs when estimating the initial sigma.  Modified
    to estimate the sigma differently in this case.  (2/17/95, Valdes)

t_fitprofs.x
splot/spdeblend.x
splot/gfit.x
    1.  The indexing was incorrect in the Monte-Carlo error estimation.
    2.  Change the number of Monte-Carlo samples from 100 to 50.
    (2/16/95, Valdes)

smw/shdr.x
    If an associated spectrum doesn't exist free any previous spectrum.
    (2/13/95, Valdes)

getcalib.x
    Added missing length argument to strcpy which caused an unaligned
    access error on the Alpha.  (1/27/95, Valdes)

t_dopcor.x
    Fixed typo bug which prevents more than 8 spectra in multispec format
    to work.  This affects primarily echelle data.  (1/18/95, Valdes)

smw/swmctran.x
    The equispec coordinate transformations now include mapping apertures
    and lines.  (1/16/95, Valdes)

smw/smwopenim.x
    Changed unknown coordinate system from a fatal error to a warning.
    (1/14/95, Valdes)

t_standard.x
    Fixed bug in closing sh structure.  (1/3/95, Valdes)

t_standard.x
t_calibrate.x
standard.par
calibrate.par
doc/standard.hlp
doc/calibrate.hlp
    If the exposure time and airmass cannot be determined from the header
    they are queried and updated in the images.  New query parameters
    were added.  (1/2/95, Valdes)

dispcor/refmsgs.x
dispcor/refgspec.x
dispcor/reftable.x
dispcor/refspectra.h
dispcor/refinterp.x
dispcor/reffollow.x
dispcor/refnearest.x
dispcor/refprecede.x
    Added error information if no reference spectrum is found to aid in
    diagnosing the problem.  (12/30/94, Valdes)

dispcor/t_dispcor.x
dispcor/dcio.x
    1.	Improved the error messages again to more clearly pinpoint problems
	with the dispersion database.
    2.	The image extensions are now stripped in REFSPEC keywords.
    (12/30/94, Valdes)

identify/identify.x
identify/reidentify.x
identify/iddofit.x
identify/identify.key
    1.  Added 'v' to change fitting weights.  (12/29/94, Valdes)

identify/t_reidentify.x
doc/reidentify.hlp
    The step parameter for multispec/equispec data is now ignored and
    all apertures are reidentified expect for a value of zero indicates
    don't reidentify anything but the reference aperture.  (11/15/94, Valdes)

onedspec.men
doc/mkspec.hlp
    Highlighted the fact that the MKSPEC task is obsolete.  (11/12/94, Valdes)

doc/identify.hlp
identify/identify.key
identify/idcolon.x
    The help described one of the options for :label to be "coords" when
    it is actually "coord".  Rather than modify the code I modified the
    help.  The colon procedure was modified only in that when it
    reports the current value of the label parameter it shows coord
    and not coords.  (11/8/94, Valdes)

doc/onedspec.hlp
doc/specwcs.hlp
    Added description of dispaxis and nsum package parameters to the package
    description.  (11/1/94, Valdes)

scombine/t_scombine.x
    There was a problem with using SCOMBINE with 2D/3D spectra in that
    it assumed the number of spectra is the second image dimension.
    Changed this to the approriate number of spectra for all spectral
    formats.  (10/27/94, Valdes)

dispcor/dctable.x
    If ignoreaps=yes and there are apertures defined with an aperture table
    or reference image then the defaults for the wavelength scale if
    an undefined aperture is encountered will be that of the first defined
    aperture unless an explicit value has been given with the task parameters.
    This is needed to make the IMRED reductions scripts run as desired.
    (10/12/94, Valdes)

smw/smwonedspec.x
smw/smwoldms.x
    Added a missing call to close the image header keyword template list
    which caused memory to not be freed.  (10/4/94, Valdes)

identify/t_reidentify.x
    Now checks for a zero step and only operates on the specified reference
    line.  (9/15/94, Valdes)

t_sfit.x
doc/sfit.hlp
doc/continuum.hlp
    Extended SFIT and CONTINUUM to work on NDSPEC spectra.  (9/13/94, Valdes)

splot/splot.x
    1.  The 'p' and 'u' now restore the "world" system before setting the
    	dispersion.  Previously if the user switched to "pixel" (with '$')
    	then a units conversion error would occur if the user tried to
	set the dispersion.
    2.  The 'v' key now toggles even if no input units are specified.
    (8/17/94, Valdes)

splot/wrspect.x
    Fixed a bug in which the output units when saving a spectrum were
    incorrectly set to be the current display units rather than the MWCS
    units.
    (8/17/94, Valdes)

splot/wrspect.x
    Fixed a typo in a pointer assignment in the case of overwriting
    an existing 2D image which caused a segmentation violation.
    (8/17/94, Valdes)

doc/splot.hlp
doc/fitprofs.hlp
    Fixed various typos and added suggestions as pointed out by Dave Bell.
    (8/17/94, Valdes)

splot/gfit.x
splot/spdeblend.x
t_fitprofs.x
    Added a check for both sigma0 and invgain being zero.
    (8/17/94, Valdes)
    
t_fitprofs.x
    Failed to treat the scaling of the sigmas properly to avoid overflow 
    problems.
    (8/17/94, Valdes)

onedspec.cl
onedspec.hd
onedspec.men
x_onedspec.x
dispcor/mkpkg
dispcor/t_disptrans.x	+
disptrans.par		+
doc/disptrans.hlp	+
    Added a new task to convert the WCS dispersion relation between units
    and to apply a vacuum/air conversion.  (8/8/94, Valdes)

t_slist.x
    Removed the restriction against N-dim spectra so that this could
    be used with BPLOT to expand a list of apertures.  (7/29/94, Valdes)

splot/spdeblend.x
splot/gfit.x
    1.  The sigmas needed to be scaled to unit mean to avoid possible
	overflow problems during the fitting.
    2.  There was an incorrect calling sequence in gfit for the new
	model parameters.
    (7/26/94, Valdes)

noao/lib/units.h
smw/units.x
splot/splot.key
specplot.key
doc/onedspec.hlp
doc/splot.hlp
    Added nanometers as a unit.  (7/21/94, Valdes)

noao/lib/smw.h
smw/shdr.x
splot/wrspect.x
splot/splot.x
    1.  Added a reddening correction flag to the basic spectrum data structure.
    2.  When writing out a spectrum with WRSPECT also update the calibration
	parameters.
    3.  Restructured WRSPECT to be more general for use with SPECTOOL and
	put an SPLOT specific routine to handle the parameter queries.
    (7/20/94, Valdes)

t_sflip.x		+
sflip.par		+
doc/sflip.hlp		+
x_onedspec.x
mkpkg
onedspec.cl
onedspec.men
onedspec.hd
    Added a new task for flipping spectra.  (7/18/94, Valdes)

splot/wrspect.x
splot/splot.x
smw/smwswattrs.x
    Fixed a rather tricky bug with replacing a spectrum in the current
    image with SPLOT.  (7/13/94, Valdes)

splot/spdeblend.x	+
splot/deblend.x
splot/gfit.x
splot/sumflux.x
splot/eqwidth.x
splot/splot.x
splot.par
t_fitprofs.x
fitprofs.par
doc/splot.hlp
doc/fitprofs.hlp
    1.  Separated the SPLOT specific delending routine from the mathematical
        deblending routines called by the various gaussian fitting routines.
    2.  Replaced deblending code with a version that uses a sigma array
	and subsampling of the pixels.  This version also allows contraining
	the relative line strengths but this feature is not used by
	SPLOT of FITPROFS.
    3.  Added constant noise and inverse gain parameters to SPLOT and FITPROFS.
    4.  If a sigma0 and inverse gain are specified the deblending estimates
	errors in the fit parameters using Monte-Carlo simulation.  The
	errors are recorded in the log and :show output.  This was
	added to both SPLOT and FITPROFS.
    5.  If a sigma0 and inverse gain are specified the centroid, flux, and
	equivalent width estimates (from 'e' key) include error estimates.
	The errors are recorded in the log and :show output.
    (7/12/94, Valdes)

dispcor/t_dispcor.x
dispcor/dcio.x
    1.  Added a check for the existence of both IDENTIFY and ECIDENTIFY
	database files for the same image.
    2.  The recent errcode check addition (5/20) was incorrect in that
	it would not proceed to look for an ECIDENTIFY file if no
	IDENTIFY file was found; i.e. echelle data would fail.  The
	appropriate checking of errors is now done.
    (7/11/94, Valdes)

t_dopcor.x
    The verbose output was enhanced to show the old redshift in the case
    of adding to warn a user.  This only applies to multispec images
    which store the redshift separately.  (7/7/94, Valdes)

t_sbands.x
    Instead of passing a file name to the routine which reads the bandpass
    descriptions a file descriptor is not passed.  This allows the
    calling procedure to use either a file or a string file.
    (6/30/94, Valdes)

doc/sbands.hlp
doc/splot.hlp
    Typo fixes.  (6/30/94, Valdes)

doc/dopcor.hlp
    Made a slight change to description of isvelocity to make as clear as
    possible that velocities are relativistic and not c*z velocities.
    (6/30/94, Valdes)

t_rstext.x	+
rstext.par	+
rspectext.cl
x_onedspec.e
onedspec.cl
mkpkg
    Added a compiled task to reformat the input RSPECTEXT file into the
    formats needed by RTEXTIMAGE and DISPCOR and modified RSPECTEXT
    to use it.  This improves the speed of this script task enormously for
    large input text files since the CL facilities can be slow.
    (6/20/94, Valdes)

splot/wrspect.x
    Failed to initialize a pointer to NULL.  This became a seg vio after the
    changes for the BANDID info.  (6/15/94, Valdes)

scombine/generic/iccclip.x
scombine/generic/icsclip.x
    Found and fixed another typo bug.  (6/7/94, Valdes/Zhang)

scombine/generic/icaclip.x
scombine/generic/iccclip.x
scombine/generic/icpclip.x
scombine/generic/icsclip.x
scombine/generic/icgrow.x
scombine/generic/icmedian.x
    The restoration of deleted pixels to satisfy the nkeep parameter
    was being done inside the iteration loop causing the possiblity
    of a non-terminating loop; i.e. pixels are rejected, they are
    restored, and the number left then does not statisfy the termination
    condition.  The restoration step was moved following the iterative
    rejection.

    There was a bug in how the restored points were added back when
    mclip=no and there are multiple residuals with the same value.
    
    Also updated icgrow and icmedian.  All these files are the same
    as the generic files from IMCOMBINE reduced to only the real datatype.
    (6/13/94, Valdes)

t_sbands.x
    When scanning the bandpass file, if there was an filter file then
    the scanning of the filter file caused the remaining scan of the
    bandpass line to be terminated.  This was fixed by using getline
    instead of fscan in the scanning the bandpass file.  (6/3/94, Valdes)

doc/sbands.hlp
    Fixed a discrepancy in the bandpass file description between the
    description section and the examples.  (6/2/94, Valdes)

splot/splot.x
splot/splotcolon.x
splot/splot.key
splot.par
doc/splot.hlp
    Added an overplot options to permanently toggle overplotting.
    (5/31/94, Valdes)

scombine/icscale.x
    The sigma scaling flag, doscale1, would not be set in the case of
    a mean offset of zero though the scale factors could be different.
    (5/25/94, Valdes/Zhang)

scombine/generic/icsclip.gx
    There was a missing line: l = Memi[mp1].  (5/25/94, Valdes/Zhang)

scombine/generic/icaclip.x
scombine/generic/iccclip.x
scombine/generic/icpclip.x
scombine/generic/icsclip.x
    The reordering step when a central median is used during rejection
    but the final combining is average was incorrect if the number
    of rejected low pixels was greater than the number of pixel
    number of pixels not rejected.  (5/25/94, Valdes)

dispcor/dcio.x
dispcor/t_dispcor.x
    All warning messages were being converted to a single warning which
    was not appropriate in all cases.  Added an errcode check.
    (5/20/94, Valdes)

============================
V2.10.3beta internal release
============================

noao/lib/smw.h
smw/shdr.x
t_fitprofs.x
t_sarith.x
splot/wrspect.x
    The spectrum data structure was modified so that it can contain
    all the associated spectra such as the spectrum, raw spectrum,
    sky, continuum, and sigma.  Also the STYPE field was changed
    to an array of string pointers SID to contain the specturm
    type strings for all the associated spectra.  Except for the
    SID changes (in FITPROFS, SARITH, and SPLOT) the structure
    changes are invisible to any spectral task.  (5/4/94, Valdes)

scombine/icscale.x
scombine/t_scombine.x
    There is now a warning error if the scale, zero, or weight type
    is unknown.  (5/2/94, Valdes)

t_sfit.x
sfit.par
continuum.par
doc/sfit.hlp
doc/continuum.hlp
    1.	The sample regions are now set to the task parameter after each
	fit.  Previously this was only done for the first spectrum and
	after that it was set to "*".
    2.  A straightforward replication of the line selection mechanism
	to allow band selection was added.
    (4/29/94, Valdes)

identify/t_reidentify.x
    The refit=no options would not work if there was not dispersion
    function even though it makes sense to do so.  It was case of
    the if clauses not being defined correctly.  (4/28/94, Valdes)

dispcor/dcio.x
    A possibly very useful and common case is when IDENITFY/REIDENTIFY
    are used on previously dispersion corrected data to get only a
    shift with no dispersion function.  DISPCOR was modified to
    allow this case.  (4/21/94, Valdes)

scombine/iclog.x
    Changed the mean, median, mode, and zero formats from 6g to 7.5g to
    insure 5 significant digits regardless of signs and decimal points.
    (4/13/94, Valdes)

noao/lib/smw.h
smw/shdr.x
    The standard spectrum data structure now includes a pointer for a
    continuum spectrum.  Currently it is unused.  (4/12/94, Valdes)

scombine/icscale.x
    When the combine object is "sum" the task attempts to compute the
    total exposure time.  Since a missing exposure time is represented
    as INDEF this caused an arithmetic error.  The task was modified to
    not compute or output a total exposure time if any of the spectra
    have an undefined exposure time.  (4/11/94, Valdes)

identify/idmark.x
    Changed the mark and mark label color to be the tick label color
    currently in effect.  Eventually the user should have more control
    over the color but this cannot be done without changing GTOOLS or
    IDENTIFY more than is appropriate at the moment.  (4/11/94, Valdes)

doc/identify.hlp
    Fixed a typo in the description of the Legendre polynomial formula.
    (4/11/94, Valdes)

smw/shdr.x
    The case of DC-FLAG=-1 was not being handled by shdr_lw and shdr_wl.
    (4/9/93, Valdes)

smw/shdr.x
    The flux units were not being copied when the spectrum header is
    copied.  (3/31/94, Valdes)

t_sarith.x
    The string used to read in the aperture, band, and beam lists was
    SZ_FNAME which is too short for possible input lines.  Changed
    the lengths to SZ_LINE.  (3/31/94, Valdes)

splot.par
    Changed the mode of line and band to be query so that if SPLOT is run
    from epar line and band queries will still be made.  (3/21/94, Valdes)

scombine/generic/icaclip.x
scombine/generic/iccclip.x
scombine/generic/icsclip.x
    The image sigma was incorrectly computed when an offset scaling is used.
    (3/8/94, Valdes)

smw/shdr.x
    The call to shdr_units can specify "default" to restore the original
    units.  (3/7/94, Valdes)

smw/shdr.x
splot/wrspect.x
t_sarith.x
t_fitprofs.
    Fixed problems when NP1 > 1 due to a IMSHIFT operation that moves
    the first physical pixel higher logical coordinates (or the
    first logical pixel in the image corresponds to a negative
    physical pixel coordinate).  (3/5/94, Valdes)
    
t_deredden.x
    Fixed bug causing memory corruption.  (3/2/94, Valdes)

scombine/icscale.x
scombine/iclog.x
    1.  The exposure time was not being summed when summing spectra.
    2.  The exposure time is now printed whenever the exposure time is used
	even if the times are all equal.
    (2/24/94, Valdes)

t_deredden.x
doc/deredden.hlp
    Overriding a previous correction will apply to the original data
    rather than being incremental.  (2/23/94, Valdes)

smw/shdr.x
noao$lib/smw.h
    Added structure fields for the flux units and shdr_open sets the
    field if possible.  The flux units are determined first by any
    BUNIT keyword, then if the flux calibration flag is set by
    the magnitude of the data.  (2/22/94, Valdes)

smw/funits.x		+
noao$lib/funits.h	+
    Added a flux units package.  (2/21/94, Valdes)

smw/shdr.x
    Added a routine to change the units.  (2/19/94, Valdes)

splot/usercoord.x
    The routine was not correct for input log-linear spectra (dc-flag=1).
    (2/19/94, Valdes)

dispcor/dispcor.x
    Fixed typo (out[1] -> out[i]) which was causing the non-flux conserving
    mode to fail.  (2/18/94, Valdes)

splot.par
specplot.par
doc/splot.hlp
doc/specplot.hlp
    1.  SPLOT will write out the current display units to the WCS attribute
	"units_display".
    2.  The default "units" task parameter now has the null string value
	to allow selecting the units given by "units_display" or the WCS
	units in that order.  (2/18/94, Valdes)

smw/smwsaveim.x
smw/smwesms.x
smw/smwmerge.x
smw/smwndes.x
smw/shdr.x
    1.  A new WCS attribute "units_display" has been defined.  It is now
	stored in the image and transfered when copying WCS if it is defined.
    2.  When a spectrum is opened with shdr_open the user units are set
	to that specified by "units_display" if present.  Otherwise
	the units of the WCS are used.
	(2/18/94, Valdes)

noao$lib/smw.h
shdr.x
    Added a field to the standard spectrum data structure to contain an
    error array.  This array is filled in by shdr_open if a new flag
    value is used.  Since there are no current tasks which use the
    new value this feature is unused in current tasks.  (2/7/94, Valdes)

noao$lib/smw.h
t_fitprofs.x
t_sarith.x
splot/wrspect.x
smw/smwsaveim.x
    Added a field to the standard spectrum data structure to contain the
    type of spectrum; i.e. spectrum, background, sigma.  This type is
    stored in the BANDIDn keywords for multispec format data extracted by
    APEXTRACT.  This information, if present, is now updated on outputing a
    new spectrum.  This is particularly important for SCOPY when the bands
    are adjusted.  (2/4/94, Valdes)

dispcor/t_dispcor.x
    Deleted unused variable, junk, which somehow snuck in.  (2/7/94, Valdes)

t_specplot.x
specplot.key
doc/specplot.hlp
    Extended the :units command to allow specifying individual spectra.
    This is intended to allow multiple spectra to be plotted on a velocity
    scale with different zero points.  (2/4/94, Valdes)

smw/shdr.x
smw/smwmw.x
    Added checks for the aperture number to be outside of the range of
    spectra in N-dimensional spectra.  (1/8/94, Valdes)

splot/splot.x
splot/splotcolon.x
splot/splot.key
doc/splot.hlp
splot.par
    A new options, "flip", has been added to select plotting the spectra
    in decreasing wavelength.  (12/8/93, Valdes)

dispcor/dispcor.x
doc/dispcor.hlp
    When flux=no DISPCOR now computes an average across the output pixel
    rather than interpolating to the pixel center.  This allows
    flux density conservation.  (12/6/93, Valdes)

identify/idinit.x
    Changed aclrr to aclri.  (12/1/93, Valdes)

doc/identify.hlp
    Added a description of the function coefficients. (12/1/93, Valdes)

t_calibrate.x
    Added a warning if the exposure time is not found.  (11/19/93, Valdes)

sensfunc/sfoutput.x
    Instead of using the dispersion range from a single standard star
    the code now uses the maximum range and minimum dispersion.
    (11/15/93, Valdes)

t_sbands.x	+
sbands.par	+
doc/sbands.hlp	+
x_onedspec.x
onedspec.cl
onedspec.men
onedspec.hd
mkpkg
    Added a new task to do bandpass spectrophotometry.  (11/1/93, Valdes)

rspectext.cl
wspectext.cl
doc/rspectext.hlp
doc/wspectext.hlp
onedspec.cl
onedspec.men
onedspec.hd
    Added two script tasks to convert between 1D image spectra and
    ascii text spectra.  (10/22/93, Valdes)

splot/splot.x
splot/getimage.x
splot/splotfun.x
doc/splot.hlp
    If a wavelength scale is set with 'p' or 'u' then all subsequent
    spectra which are not dispersion calibrated will use that wavelength
    scale.  (9/2/93, Valdes)

t_sapertures.x
    The negative beam number warning is only issued if verbose = yes.
    (9/1/93, Valdes)

dispcor/t_dispcor.x
smw/smwesms.x
    The aperture IDs were not being properly propagated.  (9/1/93, Valdes)

t_fitprofs.x
fitprofs.par
doc/fitprofs.hlp
    1.  Fixed bug with close MWCS
    2.  Add a bands parameter for 3D images.
    (8/31/93, Valdes)

t_deredden.x
    There was an error in freeing the sh pointer causing a segmentation
    violation after the spectra are successfully dereddened.  (8/13/93, Valdes)

splot/splot.x
doc/splot.hlp
    The '(' and ')' keys will now cycle in bands if there is only one line.
    (8/10/93, Valdes)

t_sapertures.x
    Modified to ignore attempts to set a negative beam number.
    (8/9/93, Valdes)

splot/wrspect.x
    Added check against an error opening an output image in shdr_open.
    (8/4/93, Valdes)

splot/fudgex.x
    Added check against a divide by zero if the cursor is not moved.
    (8/4/93, Valdes)

splot/splotcolon.x
    The call to ans_hdr in the COMMENT case was missing the key argument.
    (8/3/93, Valdes)

smw/smwonedspec.x
    For spectra which are dispersion corrected (DC-FLAG set) but have no
    units the code was setting the "label" rather than "units" to
    "angstroms".  (8/3/93, Valdes)

============
V2.10.3 beta
============

splot.par
splot/smooth.x
doc/splot.hlp
    1.  The parameter file parameter prompt for the smoothing box size was
	modified to request an odd number.
    2.  If an even number is given, a warning is printed.
    3.  The help for the parameter boxsize indicates the the value must
	be odd.
    (6/28/93, Valdes)

scombine/icscale.x
    The result of reading an @file for the zero or weight parameter was
    being placed in the scales array.  This has been fixed.  This
    affected only one IRAFX users.  (6/28/93, Valdes)

specplot.key
    Added missing :redshift and :velocity commands in the summary.  Also
    sorted and cleaned up the multicolumn lists.  (6/15/93, Valdes)

t_dopcor.x
dopcor.par
doc/dopcor.hlp
    An new parameter has been added to allow combining sequential
    corrections in "multispec" format spectra.  (6/15/93, Valdes)

usercoord.x
wrspect.x
t_dopcor.x
    When smw_swattrs is called it is possible that the smw pointer will be
    changes (promoting an equispec format to multispec).  If this happens
    and the pointer is part of an open shdr structure then the routine
    must invalidate the mwcs stuff and possibly open or update the shdr
    structure.  (6/14/93, Valdes)

bplot.cl
doc/bplot.hlp
    The query parameters from SPLOT were added as hidden parameters in
    BPLOT to allow such things as writing output spectra without generating
    queries.  (6/8/93, Valdes)

identify/ididentify.x
    Added newlines when printing to the status line.  This is needed when
    redirecting the output to a file in the IMRED scripts.  (6/4/93, Valdes)
   
identify/iddelete.x
    The label pointers needed to be updated when deleting a feature.
    (6/4/93, Valdes)

t_specplot.x
    Modified the log output format to include the aperture number.
    (5/25/93, Valdes)

t_sarith.x
t_fitprofs.x
wrspect.x
    The conversion from logical to physical coordinates was incorrect in
    that it truncated the physical coordinates.  This could cause a subtle
    error in the coordinate system.  (5/20/93, Valdes)

identify/idmap.x
    The user specified vector axis is interpreted as a logical axis rather
    than a physical axis.  This is only significant for transposed images.
    (5/14/93, Valdes)

smw/smwsaxes.x
smw/smwsaveim.x
    Transposed NDSPEC images are now allowed.  (5/11/93, Valdes)

getcalib.x
    Added a search for alternate standard names in a file <caldir>names.men
    if that file is present.  (5/4/93, Valdes)

splot/splot.x
splot/anshdr.x
splot/avgsnr.x
    Added logging of the 'm' key output.  (5/4/93, Valdes)

splot/splot.x
splot/splotfun.x
    1.  fun_do was not initializing the pointers passed to getimage.
	This proves to be a problem if an error occurs in getting the
	second image data, such as due to a mistype, so that the
	next time the routine is called an invalid pointer is found
	and a segmentation error occurs.
    2.  Added a time delay on an error message in fun_do followed by the
	function mode prompt.
    (3/2/93, Valdes)

sensfunc/sfstds.x
    1.  Eliminated input stars/apertures that have no data.
    2.  Eliminated input flux points outside the range of the
	star/aperture wavelength range.
    3.  Improved the iterative fitting to drop back to a polynomial
	function if the lowest order spline does not fit.
    (2/12/93, Valdes)

identify/idgraph.x
    Because these procedures used the SX array as temporary storage it
    caused the initialize option to fail.  (2/3/92, Valdes)

onedspec.men
    Removed reference to dispaxis.  (1/21/93, Valdes)

scombine/generic/icaclip.x
scombine/generic/iccclip.x
scombine/generic/icpclip.x
scombine/generic/icsclip.x
    When using mclip=yes and when more pixels are rejected than allowed by
    the nkeep parameter there was a subtle bug in how the pixels are added
    back which can result in a segmentation violation.
	if (nh == n2)  ==>  if (nh == n[i])
    (1/20/93, Valdes)

sensfunc/sensfunc.h
sensfunc/sfgraph.x
sensfunc/sfginit.x
sensfunc/sfimage.x
sensfunc/sfcgraph.x
sensfunc/sfextinct.x
sensfunc/sfcolors.x
sensfunc/sfcolon.x
sensfunc/sfmove.x
sensfunc/sfundelete.x
sensfunc/sfdelete.x
sensfunc/sfadd.x
sensfunc/mkpkg
sensfunc/sensfunc.key
sensfunc.par
doc/sensfunc.hlp
    Added color support. (12/17/92, Valdes)

splot/gfit.x
splot/eqwidthcp.x
splot/deblend.x
splot/splot.x
identify/idmark.x
    Added color support.  (12/8/92, Valdes)

splot/sumflux.x
    1.  There was no check of whether esum was INDEF (a possible value) before
        multiplying by wpc.  A check was added.
    2.  Because of a change to fp_equalr which occured on (10/18) the
        equivalent widths of flux calibrated data would be INDEF.  To
	compensate the test is made on scaled data.
	(12/7/92, Valdes)

units.h
    The conversion factors for millimeter and centimeter were off by a
    factor of 10.  (12/4/92, Valdes)

dispcor/dcio.x
    The wrong axis was selected in computing the logical NW.  (11/24/92, Valdes)

splot/splot.x
splot/usercoord.x
splot/splot.key
splot/mkpkg
doc/splot.hlp
splot.par
    Changed the 'u' and 'p' keys to include additional ways to adjust the
    dispersion scale.  In particular a doppler and zeropoint adjustment can
    be made using the cursor and entering a coordinate.  Note that these
    two adjustments apply to all coordinate systems and units and do not
    require assuming a linear dispersion.  In effect these are interactive,
    cursor marking versions of DOPCOR (without the flux correction) and
    SPECSHIFT.  The coordinates are specified in the current displayed
    units.  The code that does the adjustment is now well integrated with
    the MWCS rather than fudging the W0 and WP entries.  The output of a
    new spectrum with 'i' will properly handle the adjusted coordinate
    system.  (11/20/92, Valdes)

bplot.cl
irsiids/bplot.cl
doc/bplot.hlp
gcurval -> gcurval.dat
    Changed the name of the default cursor file to avoid stripping.
    (11/20/92, Valdes)

splot/wrspect.x
    Fixed typo affecting 3D images:  PNDIM(out) --> PNDIM(sh2).
    (11/19/92, Valdes)

splot/wrspect.x
    A spectrum was being written using the W0, WPC of the current units
    rather than Angstroms as it should be.  A call to un_ctran to convert
    to the MWCS units was added.  (11/17/92, Valdes)

t_specplot.x
specplot.h
doc/specplot.hlp
specplot.key
    Added a color parameter for specifying the color of each spectrum
    on color graphics terminals.  (10/30/92, Valdes)

t_sarith.x
t_fitprofs.x
splot/wrspect.x
    1.	The doppler correction was still not properly handled.  Instead of
	dividing by (1 - z) it should multiple by (1 + z) in order to
	be symmetric with the WCS driver.
    2.  To avoid roundoff with multispec format W0 and W1 (which are real)
	are not used when recalculating the w1, dw attribute values.  Instead
	shdr_lw is called to get the double precision values.
	(10/16/92, Valdes)

dispcor/t_dispcor.x
dispcor/dcio.x
doc/dispcor.hlp
    DISPCOR will now allow multiple uses of IDENTIFY dispersion solutions
    in a simple way with but with continuing protection against accidental
    multiple uses of the same dispersion solutions.  When a spectrum is
    first dispersion corrected using one or more reference spectra keywords
    the dispersion flag is set and the reference spectra keywords are moved to
    DCLOGn keywords.  If DISPCOR is called again without setting new
    reference spectra keywords then the spectra are resampled (rebinned)
    using the current coordinate system.  If new reference spectra are set
    then DISPCOR will apply these new dispersion functions.  Thus the user
    now explicitly enables multiple dispersion functions by adding
    reference spectra keywords and DISPCOR eliminates accidental multiple
    uses of the same dispersion function by renaming the reference
    spectra.  The renamed keywords also provide a history.

    Some additional log and verbose output was added to better inform the
    user about what is done.
    (10/15/92, Valdes)

t_specshift.x		+
specshift.par		+
doc/specshift.hlp	+
x_onedspec.x
mkpkg
onedspec.cl
onedspec.men
onedspec.hd
imred$argus/argus.cl
imred$ctioslit/ctioslit.cl
imred$echelle/echelle.cl
imred$hydra/hydra.cl
imred$iids/iids.cl
imred$irs/irs.cl
imred$kpnocoude/kpnocoude.cl
imred$kpnoslit/kpnoslit.cl
imred$specred/specred.cl
imred$argus/argus.men
imred$ctioslit/ctioslit.men
imred$echelle/echelle.men
imred$hydra/hydra.men
imred$iids/iids.men
imred$irs/irs.men
imred$kpnocoude/kpnocoude.men
imred$kpnoslit/kpnoslit.men
imred$specred/specred.men
    The new task SPECSHIFT applies a coordinate system shift to selected
    spectra.  For linear coordinate systems this is done by changing
    the wavelength of the first physical pixel.  For nonlinear systems
    the existing shift coefficient is adjusted.
    (10/14/92, Valdes)

dispcor/dcio.x
    Added step to update the linear part of the nonlinear WCS.
    This is mostly cosmetic.
    (10/14/92, Valdes)

dispcor/idmap.x
    Changed the way the image is opened to avoid updating the WCS.
    (10/14/92, Valdes)

*doc/onedspec.hlp
smw.x
    1.  Spectra in a single image which all have the same linear dispersion
	are now stored with linear axis types.  This gives a simpler header
	structure than the multispec axis type for this common case.  This
	modification applies to 1, 2, and 3 dimensional images.
    2.  Extensions were added to allow importing spectra which use
	a different WCS driver than multispec or linear.
    (10/13/92, Valdes)

doc/onedspec.hlp
    First an error in a font switch causing part of the text to all be in
    standout.  (10/9/92, Valdes)

scombine/t_scombine.x
scombine/icombine.h
scombine/icombine.com
scombine/icombine.x
scombine/icscale.x
scombine/iclog.x
scombine/generic/iccclip.x
scombine/generic/icsclip.x
scombine/generic/icpclip.x
scombine/generic/icaclip.x
scombine/generic/icgrow.x
scombine.par
doc/scombine.hlp
    The weighting was changed from using the square root of the exposure time
    or spectrum statistics to using the values directly.  This corresponds
    to variance weighting.  Other options for specifying the scaling and
    weighting factors were added; namely from a file or from a different
    image header keyword.  The \fInkeep\fR parameter was added to allow
    controling the maximum number of pixels to be rejected by the clipping
    algorithms.  The \fIsnoise\fR parameter was added to include a sensitivity
    or scale noise component to the noise model.
    (10/2/92, Valdes)

splot/usercoords.x
    This routine no longer puts a default value in the wavelength parameters.
    This will allow using SPLOT to noninteractively set wavelengths.
    (9/17/92, Valdes)

identify/idfitdata.x
identify/idmark.x
identify/idgdata.x
identify/idcenter.x
    IDENITFY/REIDENTIFY use the standard SHDR interface which eliminates
    data with negative physical coordinates.  This occurs because NP1 is
    then computed to be positive.  The case where this can occur is using
    IMSHIFT with a positive shift though explicit use of NP1 could also do
    it.  However, the above routines use the MWCS logical-physical and
    physical-logical conversions without accounting for NP1.  This results
    in incorrect results.  The routines were fixed to apply NP1.  (9/16/92,
    Valdes)

splot/splot.x
splot/getimage.x
    Modified getimage to also allow specification of the aperture.  This
    is needed in order for the scrolling through lines, the '(' and ')'
    keys, to work correctly by indicating that the aperture number is
    to be ignored.  (9/8/92, Valdes)

dispcor/dcio.x
    The computation of the aperture center was not prepared to deal with
    INDEF aperture limits.  (9/3/92, Valdes)

smw.x
    There was a type mismatch when setting aplow and aphigh to INDEF.
    Changed to set them to INDEFD.  This bug caused the APLOW and APHIGH
    keywords to appear in the image header unexpectedly with IDENTIFY
    on the VaxStation port.  (8/31/92, Valdes)

ecidentify/ecgetim.x
identify/idnoextn.x
    The algorithm for stripping the image extension could get confused
    with the name such as ec025.john.ec --> ec025n.ec.  The routines
    were modified to use xt_imroot which does a better job.  (8/31/92, Valdes)

t_sarith.x
smw.x
    Added provision to save multispec title in MSTITLE keyword when
    separating out multispec spectra or converting to simple 1D format and
    to restore the title when combining 1D spectra into a multispec
    spectrum.  (8/24/92, Valdes)

sensfunc/sfsvstats.x
    A real variable was used where a double should have been giving round
    off errors in the computation of the standard deviation.  (8/13/92, Valdes)

t_sfit.x
    Output images are of type real regardless of the input type.
    (8/11/92, Valdes)

scombine/icscale.x
    The zero level offsets were being incorrectly scaled twice.
    (8/10/92, Valdes)

dispcor/refgspec.x
    Arguments incompatible with intrinsic function:
	sortval = mod (sortval + 24. - timewrap, 24.)
    Changed second 24. to 24.0D0.  (8/10/92, Valdes)

splot/fixx.x
    Arguments incompatible with intrinsic function:
	z1 = max (0.5, min (double (SN(sh)+.499), shdr_wl(sh, z1)))
	z2 = max (0.5, min (double (SN(sh)+.499), shdr_wl(sh, z2)))
    The 0.5 should be double.  (8/10/92, Valdes)

shdr.x
    Arguments incompatible with intrinsic function:  on lines 268-269,
    319-320, need to real the image limits.  (8/10/92, Valdes)

units.x
onedspec.hlp
    The velocity label was changed to "cz velocity" to show that it
    is c*z and not a true velocity.  (7/30/92, Valdes)

dispcor/t_dispcor.x
    Changed WCSDIM to be 3 in the case of a 3D image.  (7/27/92, Valdes)

splot/splot.x
    Getttng a new image always forces the data to be read even if the
    same image is given.  (7/20/92, Valdes)

smw.x
    Altered the way in which old APNUM keywords are deleted to avoid
    a problem with the limit on the number of keywords that can be
    mapped with imofnl in the imio$db package.  (7/17/92, Valdes)

splot/replot.x
    Replaced gascale with gt_ascale to do the autoscaling only within
    the GTOOLS window.  (7/16/92, Valdes)

t_sapertures.x
sapertures.par
doc/sapertures.hlp
    Modified this task to allow resetting the WCS to pixels and changing
    any of the WCS fields.  (7/2/92, Valdes)

splot/wrspect.x
    Harmless typo fix mwopen -> mw_open.  (7/1/92, Valdes)

t_sarith.x
    Modified to properly handle 3D images.  (7/1/92, Valdes)

t_sarith.x
    Onedspec output format now splits out the bands as well.
    (7/1/92, Valdes)

=======
V2.10.2
=======

t_dopcor.x
doc/dopcor.hlp
    1.  The conversion from velocity to z was incorrect.
    2.  Checks were added for reasonable velocities and redshifts.
    3.  A negative sign for a header parameter changes the sense of
	a redshift if the parameter is a redshift.

=======
V2.10.1
=======

t_deredden.x
    The declaration for decode_ranges was incorrect.  Changed from bool to int.
    (7/21/92, Valdes)

shdr.x
    1.  An earlier fix left the aaxis parameter undefined for longslit images.
    This meant that references to IM_LEN(im,aaxis) yield the dimension
    of the image rather than the axis length.
    2.  Discovered that image sections don't automatically reset the lengths
    of the higher dimensions to 1 as assumed in several tasks.  SHDR now
    resets these.  (7/20/92, Valdes)

=======
V2.10.0
=======

irsiids/batchred.cl
    The parameter "recformat" in STANDARD and CALIBRATE and "apertures" in
    CALIBRATE are no longer present.  The BATCHRED task was modified to not
    add these parameters to the PROCESS script.  (7/6/92, Valdes)

shdr.x
    The resampling in shdr_linear and shdr_rebin is now an average rather
    than a sum.  (6/23/92, Valdes)

splot/wrspect.x
    New output spectra are created type real.  (6/22/92, Valdes)

scombine/icscale.x
scombine/t_scombine.x
    The exposure time is only required now if scaling or weighting by
    the exposure time.  (6/22/92, Valdes)
    
mwcs$wfmspec.x
    The inverse coordinate transform could fail in some cases.  An extra
    check was added to avoid this.  (6/17/92, Valdes)

smw.x
    Added special case to convert a 2D image which has a second dimension
    length of 1 to a 1D image.  Note this is different than a 1D section
    of a 2D image.  (6/17/92, Valdes)

shdr.x
    Added additional check for a 2D image with the dispersion axis
    along a dimension of length 1; for example [800,1] with dispaxis=2.
    This will also give an warning and then choose the appropriate
    axis.  (6/17/92, Valdes)

t_sarith.x
t_fitprofs.x
splot/wrspect.x
    The doppler correction was not properly handled when creating a new
    output spectrum.  (6/17/92, Valdes)

shdr.x
    The change to catch an inappropriate dispersion axis for TWODSPEC
    images was not complete.  I'm not fully sure anymore what should be
    done but I made the checking better. (6/3/92, Valdes)

t_sinterp.x
    Change the roundoff when computing the number of pixels to nearest
    integer.  (6/3/92, Valdes)

scombine/t_scombine:
    There was a bug in which the j loop index was redefined in the loop
    when checkin the MINMAX rejection limits.  (6/1/92, Valdes)

t_sarith.hlp
    Needed to allocate the coeff pointer in sa_1d.  Attempting to copy
    a long slit spectrum to onedspec format caused a segmentation violation.
    (5/27/92, Valdes)

doc/scopy.hlp
doc/sarith.hlp
    The examples incorrectly showed nsum to be a task parameter.
    (5/21/92, Valdes)

bplot.cl
    The error when a nonexistent image was specified was not properly
    handled.  (5/18/92, Valdes)

splot/splot.key
    Clarified 'o' key description.  (5/14/92, Valdes)

smw.x
scombine/t_scombine.x
    1. Added additional commands to delete keywords which should not be
       present.
    2. When mapping the output image a copy of the input image header is
       made.  This header may contain WCS keywords which are invalid.
       A call is now made to smw_openim() which has the effect of cleaning
       up the header.
    (5/14/92, Valdes)

=====
V2.10
=====

doc/*.hlp
doc/sys/onedv210.ms +
    Make documentation changes to allow all revisions to be obtained with
    "help onedspec.* sec=rev".  The package revisions summary was prepared
    and installed.  (5/6/92, Valdes)

splot/splot.x
splot/splotcolon.x
splot/splot.key
splot.par
doc/splot.hlp
    1.	Added the option "wreset" to have the graph limits automatically
	restored to the initial values for each new spectrum.
    2.  Added colon commands to change the options interactively.
    (5/6/92, Valdes)

shdr.x
    A 1D image section of a 2D (not multispec) image which is not along
    the specified dispersion axis will now print a warning and use the
    specified axis rather than aborting.  (5/6/92, Valdes)

smw.x
shdr.x
    Added checks in the case of log-linear dispersion (DC-FLAG=1) that
    the coordinates make sense.  Otherwise a linear dispersion is used.
    This comes up when DC-FLAG is set to 1 but the other coordinate
    information is incorrect or missing resulting in pixel coordinates.
    Without this check there would be an attempt to take the dex of
    a pixel coordinate causing a floating overflow error.
    (5/5/92, Valdes)

identify/t_reidentify.x
    Added call to strip whitespace from the reference image name
    accidentally entered by the user.  Extra whitespace  caused a
    mysterious behavior in finding a database entry which was hard
    to track down.
    (5/1/92, Valdes)

identify/idinit.x
    Added check to not unmap the database if it was never openned.
    This would cause a segmentation error if a database was never
    accessed.
    (5/1/92, Valdes)

identify/iddb.x
identify/t_reidentify.x
identify/identify.h
    The database interaction was poorly done resulting in repeatedly
    opening and reading the database file.  If there are many entries this
    becomes very slow.  The DTTEXT routines were modified to add a remap
    routine allowing a database file to remain open but automatically
    closing and opening a new database if the database name changes.  It
    also allows changing access modes by closing and opening the file but
    leaving the rest of the data structure alone. This avoids the need to
    rescan the file each time the access mode changes and allows existence
    checks for entries (from the original scan) while still in APPEND mode
    without having to switch file access modes.  The identify structure was
    extended to include the database pointer so that id_dbread and
    id_dbwrite could use the remap routine without closing the database
    between calls.  Thus, repeated calls to id_dbread and id_dbwrite for
    the same image are much more efficient and the database is only scanned
    once in the first read.  There is still a slight inefficiency in that
    switching between reading and writing requires reopening the file.  For
    the purposes of simple checking for existing entries without needing to
    read the entry and change modes a new routine id_dbcheck was added.
    Finally, the logic in REIDENTIFY was modified so that repeated mode
    switches between reading and writing are avoided.  The id_dbcheck
    routine is used when override checking is enabled.  REIDENTIFY
    is now much faster when dealing with large numbers of spectra in
    images (long slit with a fine step size or multifiber spectra with
    many fibers).  (4/30/92, Valdes)

smw.x
    An axis map is set for 1D multispec images.  (4/27/92, Valdes)

shdr.x
    Shdr_system was changing the wrong pointers causing later calls to
    shdr_open to produce an invalid coordinate system.  (2/18/92, Valdes)

scombine/t_scombine.x
scombine/iclog.x
scombine/icscale.x
scombine/icombine.x
scombine.par
doc/scombine.hlp
    1.  The gain and read noise must be read when the image is open and
    	are stored in the RA and DEC spectrum structure parameters.
    2.  NCOMBINE is not used on input.
    3.  The exposure time is taken from the spectrum structure and the
        keyword name is no longer a parameter.
    (2/12/92, Valdes)

scombine/icscale.x
    Changed action for negative scaling, etc. to a warning.
    (2/10/92, Valdes)

calibrate.par
sensfunc.par
standard.par
onedspec.par
doc/calibrate.hlp
doc/sensfunc.hlp
doc/standard.hlp
doc/package.hlp
    1. Redirected observatory parameter to package parameter
    2. Added observatory package parameter
    (2/6/92, Valdes)

ecidentify/ecdofit.x
ecidentify/ecffit/ecfcolon.x
    1.  The rejected points were not being reset between fits resulting in
	misleading RMS values.
    2.  Expanded the :show in fit mode.
    (2/6/92, Valdes)

t_standard.x
standard.key
    1.  The abbreviation of N or Y for NO or YES is now allowed.
    2.  The key file was moved from noaolib$scr to onedspec$
    (2/6/92, Valdes)

t_calibrate.x
t_standard.x
irsiids/t_bswitch.x
    Converted from obsimcheck to obsimopen.  (2/4/92, Valdes)

identify/*
doc/identify.hlp
    Added feature labels.  (1/30/92, Valdes)

refspectra.par
doc/refspectra.hlp
dispcor/ref*
    1.  Added group parameter
    2.  Sort parameter is now used as a double
    3.  If group or sort keywords are specified but not found it is a fatal
	error.
    (1/29/92, Valdes)

t_sfit.x
sfit.par
continuum.par
eccontinuum.par
doc/sfit.hlp
doc/continuum.hlp
    Added the new "markrej" parameter used in ICFIT to control whether
    rejected points are marked.  (1/21/92, Valdes)

getcalib.x
    The standard star parameter query will now print the file "standards.men"
    in the calibration directory if the user supplied name does not match an
    available file. (1/20/92, Valdes)

irsiids/t_widstape.x
    Modified the widstape task to support the new mag tape name syntax.
    (1/7/92, Davis)

identify/t_reidentify.x
    If there is no dispersion function then no shift will now be computed.
    (11/18/91, Valdes)

ecidentify/ecffit/ecffit.x
    Removed the progress print statements because they mess up the screen
    clear on XTERM.  Someday it might be desirable to put them back again.
    (11/11/91, Valdes)

doc/bswitch.hlp
    Fixed minor typo where the keyword BEAM-NUM was refered to as BEAM.
    (6/19/91, Valdes)

t_combine.x
    1. The final coord scale must have WPC > 0.  Needed to add an abs(WPC)
       in case an input spectrum had negative WPC.  (5/3/91, Valdes)

getnimage.x
t_bswitch.x
    Moved procedure add_spec from getnimage.x to t_bswitch.x (4/25/91, Valdes)

t_calibrate.x
    MWCS modifications.  Aperture selection option removed.  (4/24/91, Valdes)

splot/getimage.x
splot/splotfun.x
splot/splot.x
splot/replot.x
splot/autoexp.x
    Modified to use separate coordinate array. (3/29/91, Valdes)

iwewcs.x
gmwcs.x
wfinit.x
wfmspec.x
mwopenim1.x
mkpkg
idsm_keywrds.x
load_hdr.x
    Initial WCS modifications (3/28/91, Valdes)

====
V3.1
====

t_calibrate.x
    Moved calibration messages outside of loop over bands.
    (3/26/91, Valdes)

ecidentify/ecidentify.x
ecidentify.par
    Added autowrite parameter which is similar to that of IDENTIFY.
    (3/21/91, Valdes)

ecidentify/ecffit/ecfsolve.x
    The residual vector was not correctly set by ecf_solve.  (3/18/91, Valdes)

t_scopy.x
    1.  If no beam number is found for ONEDSPEC images it defaults to 1.
    2.  The image titles are converted to APID for ONEDSPEC images
	going to MULTISPEC if the title differs from the main MULTISPEC
	title.
    3.  Added checking for repeated aperture numbers in ONEDSPEC to
	MULTISPEC.
    (3/13/91, Valdes)

identify/t_reidentify.x
reidentify.par
    Interactive parameter is now four valued to allow better control
    of reidentification queries such as in the IMRED scripts.

dispcor/ecio.x
dispcor/ecdispcor.x
    1. Fixed datatype error when reading the low and high values from the
       APNUM keywords.
    2. Added REFSHFT capability for use with the FOE package.
    3. Added support for third dimension produced by APEXTRACT.
    (1/31/91, Valdes)

t_scopy.x
    Fixed bugs in renumber option.  It was renumbering before checking the
    aperture list rather than after.
    (1/31/91, Valdes)

ecdispcor.par
    The parameter override needed to be changed to the parameter rebin.
    (1/16/91, Valdes)

identify/iddb.x
    REIDENTIFY checked if an entry in the database was absent by checking for
    an error return from id_dbread.  The error return was made without
    first closing the database file.  When reidentifying a large number
    of images/apertures the task would run out of file descriptors.
    The fix was to put a database close statement before the error
    call.  (1/7/91, Valdes)

splot/getimage.x
load_hdr.x
    1.  Added error checking for aperture out of bounds in multispec format.
    2.  Added automatic limit on band specification in multispec format.
    3.  Added missing nband=0 for case of 1D image section.
    (1/7/91, Valdes)

identify/idlinelist.x
    The 'l' did not find lines because the first pass to finding MAXFEATURES
    did not discriminate against finding the same line with different
    user coordinates.  This locked out weaker features during the finding.
    Then when the features were added to the feature least the MINSEP
    parameter eliminated the duplicates resulting in fewer than MAXFEATURES
    features.  (12/19/90, Valdes)

splot/stshelp.key
splot/getimage.x
splot/anshdr.x
splot/mktitle.x
splot/mkpkg
splot/splotfun.x
splot/splot.x
splot/splot.key
splot.par
    1.  Added support for bands in 3D images.  This involved adding a
	band task parameter and a '%' key.
    2.  The 'o' overplot key is now a toggle for the next graph.  It
	does not query for the image.  The user follows 'o' with 'g',
	'#', or '%'.
    (12/19/90, Valdes)

splot/deblend.x
splot/gfit.x +
splot/stsfit.x
splot/stsfit.key +
splot/splot.x
splot/mkpkg
noao$lib/scr/splot.key -
splot/splot.key +
    1.  The background was not subtracted in the initial amplitude estimate.
    2.  The tau parameter in the call the hfti was too large.  Changed
	from .001 to 1E-10.
    3.  Added new gaussian fitting function, key 'G'.
    4.  Changed line help to use a file rather than coding the print
	statements.
    5.  Moved key file to source directory.
    (12/19/90, Valdes)

t_scopy.x
    1.  Added a renumber option.
    2.  For an input list of 1D images without onedspec extensions one can
	uses a null aperture list to pack them into a single multispec
	image.
	(12/13/90, Valdes)

splot/deblend.x
    1.  Fixed bug that was scaling twice in computing the initial peak values.
        This was also fixed in NEWIMRED.
    2.  Last deblending prompt was not erased.  Replaced with exiting
	deblending message.
    (12/4/90, Valdes)

t_sfit.x
    Fixed logfile prefix string from STFONTINUUM to SFIT.
    (11/20/90, Valdes and Seaman)

t_bswitch.x
t_calibrate.x
t_standard.x
sensfunc/sfimage.x
bswitch.par
calibrate.par
standard.par
sensfunc.par
doc/bswitch.hlp
doc/calibrate.hlp
doc/standard.hlp
doc/sensfunc.hlp
    Converted to using observatory database.  (11/19/90, Valdes)

t_fitprofs.x
onedspec.hd
doc/fitprofs.hlp +
    1. Modified to write output model even if there is a fitting error to
       avoid output images with not pixel file.
    2. The image title was not dereferenced when generating the log title
       string with onedspec format.
    3. Added help page.
    (11/2/90, Valdes)

identify/iddoshift.x
    Added image label shift info.
    (10/29/90, Valdes)

indentify/t_reidentify.x
    1. The entrance into the interactive mode was not initializing such things
    as the feature type and width.  It now initializes using parameters from
    IDENTIFY if needed.
    2. When not in verbose mode but when entering the interactive IDENTIFY
    it did not print the revised statistics line.  This has been fixed.
    (10/22/90, Valdes)

ecidentify.par
noao$imred/echelle/doc/ecidentify.hlp
ecidentify/ ecidentify/ecffit
noao$lib/scr/ecidentify.key --> ecidentify/ecidentify.key
noao$lib/scr/ecffit.key --> ecidentify/ecffit/ecffit.key
    1.  Moved key files to source directory.
    2.  Made changes allowing iterative rejection in the echelle dispersion
	tasks.  This adds three parameters to the ECIDENTIFY parameter
	file, the database files (backwards compatible), and colon
	commands in fitting mode.  The feature lists printed and in the
	database now include an additional column to indicated rejected
	lines. (10/15/90, Valdes)

splot/splot.x
    Changed the temporary spool file to be in tmp$.  (10/3/90, Valdes)

doc/dispcor.hlp
    Added notes warning that flux conservation will change the units of the
    flux.  (10/3/90, Valdes)

splot/splot.x
doc/splot.hlp
noao$lib/scr/splot.key
    Added :log and :nolog commands to toggle logging of measurements.
    (10/3/90, Valdes)

load_hdr.x
    Header keyword datatype conversion errors are now a warning.
    (10/3/90, Valdes)

identify/idcolon.x
    Unrecognized or ambiguous colon commands are now noted.  (10/2/90, Valdes)

dispcor.par (also in imred.iids and imred.irs)
dispcor/dispcor.x
dispcor/dcio.x
dispcor/ranges.x
doc/dispcor.hlp
    1.  is_in_range not considers INDEF to be equivalent to MAX_INT.  This
	has the effect that if no range is specified, "", then INDEF is in
	the range while is some specific range which is not open ended will
	not include INDEF in the list.
    2.  Added new verbose parameter and modified program to print messages
	when spectra are skipped.
    3.  Ignoreaps now only applies to the global wavelength determination.
    (10/2/90, Valdes)

ecidentify/ecffit/ecfgraph.x
    Put a check to avoid trying to plot points outside the defined window.
    Plotting very deviant points outside the rescaled window causes a
    gio floating overflow error.  This fix is a workaround before the
    real bug gets fixed.  (9/20/90, Valdes)

identify/idgdata.x
identify/idmap.x
    Make changes to allow working with 3D multispec images.
    (9/14/90, Valdes)

calibrate.x
sensfunc/sfgimage.x
    Make simple changes to allow working with 3D multispec images.
    (9/12/90, Valdes)

splot/splot.x
splot/fudgex.x
doc/splot.hlp
    1.  Changed the 'x' key to use only the x cursor values and connect the
	nearest pixels.  (8/31/90, Valdes)
    2.  Added a new option, xydraw, to select drawing between x-y points
	instead of using nearest pixel values.  (9/5/90, Valdes)

bplot.cl
doc/bplot.hlp
    BPLOT revised to use new SLIST.  This is a much simpler and better
    script.  It selects on aperture numbers.
    (8/24/90, Valdes)

t_slist.x
doc/slist.hlp
    SLIST now has a format parameter.  In multispec mode more approriate
    output is obtained.  The multispec mode allows selection by aperture.
    The short header listing is good for making lists for scripts to scan.
    (8/24/90, Valdes)

================================
V3 of ONEDSPEC installed 8/23/90
================================

fortran/polft1.f
    Fixed bug in which reference was made to a part of some work arrays
    not used by the program.  This caused an arithmetic error on the MIPS.
    (7/20/90, Valdes)

onedspec.cl
onedspec.men
onedspec.cl
bplot.cl
doc/msdispcor.hlp +
doc/bplot.hlp
    1.  Added MSDISPCOR to the package.
    2.  Replaced the old BPLOT with the code from MSBPLOT.  This program
	also uses change to SPLOT which selects by aperture number.

load_hdr.x
splot/splot.x
splot/mktitle.x
splot/deblend.x
splot/eqwidth.x
splot/eqwidthcp.x
splot/anshdr.x +
splot/anssave.x -
splot/mkpkg
doc/splot.hlp
noao$lib/scr/splot.key
t_standard.x
    1.  Added mapping of APID keyword, if present, to the iids structure
	LABEL field.
    2.  SPLOT, STANDARD  modified to use LABEL field instead of IM_TITLE.
    3.  SPLOT modified to use different line type during overplotting.
    4.  Removed maximum number limit for deblending.
    5.  SPLOT now uses aperture number if the image is multispec/echelle.
    6.  Added a new key, "#", to get new aperture without query about
	image.

t_specplot.x
load_hdr.x
idsm_keywrds.x
dispcor/dcio.x
dispcor/dispcor.x
dispcor/msdispcor.x
dispcor/ecdispcor.x
sensfunc/sfoutput.x
    Added CD1_1 as allowed substitute for WPC and CDELT1

onedspec.hd
    The revisions help is now a sys option.

t_scopy.x +
t_sapertures.x +
scopy.par +
sapertures.par +
doc/scopy.hlp +
doc/sapertures.hlp +
mkpkg
x_onedspec.x
t_msselect.x -
    1.  New task SCOPY added to handle copying and extraction apertures
	between different formats
    2.  New task SAPERTURES added to modify APNUM and APID info using
	a text file.
    3.	Removed MSSELECT/ECSELECT as they are replaced by SCOPY.

onedspec.cl
onedspec.hd
onedspec.men
x_onedspec.x
t_sfit.x +
sfit.par +
continuum.par
t_ecctm.x -
continuum.cl -
mkpkg
doc/sfit.hlp
doc/continuum.hlp
    1.	New task SFIT added.  This is a modification of Rob Seamans ECCONTINUUM
	task.
    2.  A new output option was added to output the data with any rejected
	points replaced by fitting values.  This replacement also may be used
	with the difference and ratio output types.
    3. 	ECCONTINUUM is just a different name for SFIT.
    4. 	CONTINUUM is just a different name for SFIT.  The script version
	based on FIT1D has been removed.

onedspec.par
    Incremented version number to V3.

t_specplot.x
specplot.par
specplot.key +
doc/specplot.hlp
noao$lib/scr/specplot.key -
    1.  Added apertures and logfile parameters.
    2.	Moved key file to source directory
    3.  Added to save sp_vshow parameters in logfile.
    4.  Added option to undelete last deleted spectrum.
    5.  Extended to also plot anything in third dimension.
    6.  Added sysid parameter.
    7.  Added ability to set line type to histogram

dispcor/dispcor.x
    1.	Added aperture position information to APNUM keyword.

msdispcor.par +
dispcor/msdispcor.x
dispcor/msio.x
dispcor/msdispcor.com
dispcor/mkpkg
    1.  Added logfile.  This is particularly for logging reference
	shift interpolation information.
    2.  Added support for 3D format
    3.  Added aperture position info for spatial interpolation.  The positions
	are read for the object from the APNUM keyword, propagated as
	needed, and read from the database for the dispersion functions.
    3a.	A reference shift spectrum may be specified.
    4.  Communicate aperture number through ms_seteval call and then do a
	lookup for all other parameters.
    5.  Propagate independent beam number.
    6.  The number of apertures in the reference spectrum need not be
	the same as the object spectrum though all object spectra must
	have a reference dispersion function.
    7.  Everything is now done by aperture number.  This allows line
	numbers to change, particularly between the dispersion reference
	image and the data image.
    8.  Dependence of msdispcor.x on msdispcor.com removed.
    9.  Fixed rounding problem in wavelengths.

dispcor/ecdispcor.x
dispcor/ecio.x
dispcor/ecdispcor.com
dispcor/mkpkg
    1.  Everything is done by aperture number using a call to ec_seteval.
	This removes dependence on ecdispcor.com
    2.  Aperture limit info is propagated

identify/identify.x --> identify/t_identify.x
identify/identify.h
identify/linelist.x + *.x
identify/iddoshift.x
identify/iddb.x
identify.par
mkpkg
    1.	Added an autowrite parameter to IDENTIFY.
    2.  Simplified linelist package by passing id pointer.  This affects
	calling sequence of a number of procedures.
    3.  Zero weight points are ignored and the number of valid features
	used in the shift is printed.
    4.  New id structure made some minor changes in main task.
    5.  Dependence on center1d.h removed by including emmission/absorption
	definitions in identify.h and new field in id structure.

identify/iddb.x
identify/identify.h
identify/idgdata.x
identify/ididentify.x
identify/idinit.x
identify/identify.key +
identify/idmap.x +
identify/idnoextn.x +
identify/idgetim.x -
idreplot.x -
mkpkg
    Added support for multispec format.
    1.  The database name string includes aperture number.
    2.  Image remains open for efficient movement through 2D image.
    3.  A number of new fields are part of the id structure including
	the image pointer, spectrum format, image axis, line number,
	aperture info, and structure for saving copies of id structure.
    4.  Added j, k, o keys to scroll through apertures.
    5.  Changes are saved internally for multiple apertures until done
	with the image.

identify/idreidentify.x --> identify/t_reidentify.x
reidentify.par
reidentify.x -
    1.  REIDENTIFY completely rewritten for efficiency, support for
	multiaperture data, and for additional features and algorithms.
    2.  The same number or order of apertures is not required.
    3.  REIDENTIFY parameters changed to include interactive, track,
	override, addfeatures, coordlist, match, maxfeatures, minsep,
	graphics, cursor, and answer.

------------------------------------------------------------------------------

load_hdr.x
splot/getimage.x
    1.	Fixed bug that was setting NP1 to 1 instead of zero.
    2.  Now load_hdr adjusts W0 to first good pixel.
    3.  SPLOT no longer adjust W0 to first good pixel since it is done
	by load_hdr.
    (7/11/90, Valdes)

onedspec$ecidentify/ecidentify.x
onedspec$ecidentify/t_ecreid.x
onedspec$ecidentify/ecdofit.x
onedspec$ecidentify/ecffit/ecffit.x
onedspec$ecidentify/ecffit/ecfsolve.x
    1.	Added a fixed order fitting option so that ECREIDENTIFY will refit
	with order fixed.  This is mostly just passing a parameter down
	to ecf_solve.  (6/12/90, Valdes)

onedspec$dispcor/msio.x
    If an aperture identify entry was missing from the database the task
    would quit with not error message.  This is fixed now though the
    new version to be installed soon will not have this approach to
    mapping the dispersion solutions anyway.  (6/4/90, Valdes)

onedspec$doc/splot.hlp
    Included query parameters since a user was asking about them.
    (6/1/90, Valdes)

====
V2.9
====

onedspec$t_sums.x
onedspec$sums.par
    If an image already exists a new query parameter will be used to get
    a new image name.  (3/29/90, Valdes)

onedspec$batchred.cl
    Turned on extinction correction in calibrate for the case the spectra
    are not already extinction calibrated.  (3/29/90, Valdes)

onedspec$ecidentify/ecdofit.x
    When INDEF valued lines were used and features were deleted during
    fitting the resorting of the feature list would get messed up.
    This is a very rare condition which has now been fixed.
    (3/16/90, Valdes)

onedspec$identify/idgdata.x
onedspec$dispcor/dcio.x
onedspec$dispcor/ecdispcor.x
onedspec$dispcor/dispcor.x
onedspec$dispcor/msdispcor.x
onedspec$dispcor/disptable.r
onedspec$sensfunc/sfoutput.x
onedspec$load_hdr.x
onedspec$idsm_keywrds.x
onedspec$t_specplot.x
    Added CDn_n to the set of keywords which may be used for the dispersion.
    (2/8/90, Valdes)

onedspec$splot/eqwidths.x
onedspec$splot/sumflux.x
    The equivalent width is now computed using the ratio of the spectrum
    to the continuum.  The previous approximation is printed in the log
    file for comparison.
    (3/5/90, Valdes)

onedspec$splot/splot.x
onedspec$splot/mktitle.x
    1. For :show added test for existence of spool file and an appropriate
       message if it does not exist.
    2. Increase length of plotted title to SZ_LINE from 32.
       (3/2/90, Valdes)

onedspec$identify/iddofit.x
    When INDEF valued lines were used and features were deleted during
    fitting the resorting of the feature list would get messed up.
    This is a very rare condition which has now been fixed.
    (1/17/90, Valdes)

onedspec$dispcor/ecdispcor.x
    The sum option was actually the same as the average option!
    (1/15/90, Valdes)

199c199
< 	call calloc (spec, nw, TY_REAL) 
---
> 	call malloc (spec, nw, TY_REAL) 
208a209
> 	    call aclrr (Memr[spec], nw)
212c213,218
< 	    case SUM, AVERAGE:
---
> 	    case SUM:
> 		do j = 1, nw
> 		    if (Memr[spec+j-1] != 0.)
> 		        Memr[outdata+j-1] = Memr[outdata+j-1] +
> 			    Memr[spec+j-1]
> 	    case AVERAGE:

onedspec$load_hdr.x
    Add limit checks for NP1 and NP2.  (11/8/89, Valdes)

onedspec$sensfunc/sfstds.x
    The data for apertures which are in the aperture list when the ignoreaps
    flag is set was not being read unless the aperture list included
    aperture 1.  This has been fixed. (11/8/89, Valdes)

onedspec$load_hdr.x
onedspec$t_specplot.x
onedspec$splot/splot.x
onedspec$splot/mktitle.x
	1. The new APID titles for multispec format spectra is now mapped into
	the unused LABEL element of the IDS structure.  For other formats
	or if the keyword is missing then the image title is substituted.
	2. SPLOT now labels with the LABEL string rather the the image
	title to allow individual titles for multispec spectra.
	3. SPECPLOT uses the APID titles if present.
	(10/27/89, Valdes)

onedspec$identify/iddofit.x
    The order of evaluation in complex if statements is not necessarily
    left to right as I'd thought.  This caused a bus error on the
    Convex.  The particular change is as follows:

    old:
	if (rejpts != NULL && Memi[rejpts+k-1] == YES)
	    WTS(id,j) = 0.
	else
	    WTS(id,j) = Memd[wts+k-1]
    new:
	WTS(id,j) = Memd[wts+k-1]
	if (rejpts != NULL)
	    if (Memi[rejpts+k-1] == YES)
		WTS(id,j) = 0.

onedspec$load_hdr.x
    Modified header access to use imaccf to check if header parameter exists
    rather than rely on an error return.  On a Sun3x the error checking
    results in an exception.  (9/28/89, Valdes)

onedspec$t_calibrate.x
    The data outside of calibration range message was changed to print how many
    pixels are outside of the calibration range is printed once.
    (8/8/89, Valdes)

====
V2.8
====

onedspec$idsmtn.h
onedspec$t_subsets.x
onedspec$t_standard.x
onedspec$t_slist.x
onedspec$t_shedit.x
onedspec$t_flatdiv.x
onedspec$t_calibrate.x
onedspec$t_bswitch.x
onedspec$t_addsets.x
onedspec$load_hdr.x
onedspec$idsm_keywrds.x
onedspec$sensfunc/sfimage.x
onedspec$splot/mktitle.x
onedspec$shparams.par
    The exposure time is used as a real rather than an integer (7/11/89, Valdes)

onedspec$t_specplot.x
    The wavelengths were off by one pixel because CRPIX was uninitialized
    and so defaulting to zero instead of 1.  (6/6/89, Valdes)

onedspec$sensfunc/sfstds.x
    Previously added check for INDEF exposure time extended to also check
    for zero exposure time.  (6/1/89, Valdes)

onedspec$dispcor/msio.x
    Because of a recent change in IDENTIFY in which 2D images with a
    second dimension of 1 are treated as 1D images a related change
    was required to allow multispec format spectra to be dispersion
    corrected if there is only one spectrum.  (5/15/89, Valdes)

onedspec$load_hdr.x
    Airmass values less than 1 are mapped in INDEF to force an airmass
    computation.  (5/8/89, Valdes)

onedspec$splot/getimage.x
    If the spectrum has only 1 line (even if it is two dimensional) there
    is no query for the line number.  Also the line number given by the
    user for 2D images is limited to the range of image lines to avoid
    an out of bounds error.  (5/6/89, Valdes)

onedspec$dispcor/dispcor.x
onedspec$dispcor/dcio.x
onedspec$doc/dispcor.hlp
    1. The output spectrum will be of real datatype if the input spectrum
       is short datatype.
    2. The last dispersion function defined for a 2D image is used for
       all lines of a 2D image.
       (5/6/89, Valdes)

onedspec$doc/dispcor.hlp
    Fixed mistake in description of the ignoreaps parameter.  (5/6/89, Valdes)

onedspec$identify/identify.h
onedspec$identify/*.x
    1.  Added weights to the IDENTIFY data structure.
    2.  Modified files to use the weights parameter.
    3.  The weights are currently used to flag iteratively rejected points
	during fitting of the dispersion function.
    4.  Reidentify now prints the RMS of only those lines used in the fit
	and shows the number of points fit.
    5.  The database files now include a column for the weights.
    (5/5/89, Valdes)

onedspec$t_standard.x
onedspec$standard.par
    1.  A warning message is printed if the exposure time is not found.
    2.  Removed ennumerated value in parameter file.
    (4/10/89, Valdes)

onedspec$sensfunc/sfstds.x
    1.	Standard values with negative counts are ignored thus avoiding
	arithmetic problems.
    2.	Warning message is printed if the exposure time in not defined and
    	a value of 1 is used.  (4/10/89, Valdes)

onedspec$dispcor/msio.x +
onedspec$dispcor/msdispcor.com +
onedspec$dispcor/msdispcor.x +
onedspec$t_msselect.x +
onedspec$dispcor/dispcor.x
onedspec$dispcor/mkpkg
onedspec$mkpkg
onedspec$x_onedspec.x
    1.  New task MSDISPCOR to make dispersion correction in related
	spectra in "multispec" format.  This is a cross between
	ECDISPCOR and DISPCOR.
    2.  New tasks MSSELECT and ECSELECT to extract subsets of spectra
	from echelle and multispec format.  ECSELECT is simply an
	alternate task name for MSSELECT.
    3.  These new tasks use the procedures in the ONEDSPEC object
	library but appear as logical tasks in the new MSRED package
	and in the ECHELLE package.
	(3/29/89, Valdes)

onedspec$dispcor/dispcor.x
    When not flux conserving the procedure asieval was being called
    with a double value instead of a real giving completely incorrect
    results.  (3/22/89, Valdes)

onedspec$dispcor/refmatch.x
    There was a bug in the matching option in which the object image was
    begin substituted for the reference image.  (3/14/89, Valdes)

onedspec$t_specplot.x
onedspec$splot.par
onedspec$splot/wrspect.x
onedspec$load_hdr.x
onedspec$identify/iddb.x
    1.  Modified SPECPLOT to accept "multispec" and "echelle" formats.
    2.  Modified SPLOT to accept "multispec" format for output.  This is
	only cosmetic since it is the same as "echelle" format.
    3.  Modified ONEDSPEC header reader to accept "multispec" format.
	This is only cosmetic since it is the same as "echelle" format.
    4.  Modified IDENTIFY to not include the image section in the REFSPEC
	parameter for use with "multispec" format.
    (3/8/89, Valdes)

onedspec$dispcor/dispcor.x
onedspec$doc/dispcor.hlp
    Simple modification to allow task to operate on all lines in a 2D
    image.  This is how the old program also worked.  (3/8/89, Valdes)

onedspec$t_calibrate.x
    1.	CALIBRATE did not take the differing lengths of the echelle orders
	into account and so gave many warnings about spectrum extends outside
	of flux calibration limits.
    2.	The warning is now only printed once per spectrum/order rather than
	for each pixel.
    (2/27/89, Valdes)

onedspec$t_specplot.x
    Made CRPIX1 a real valued parameter. (2/27/89, Valdes)

onedspec$t_widstape.x
	The function mtfile is now used to determine if the input file is
	a mag tape.  Previously, the code was checking that the first two
	letters of the input file were 'mt', which fails for remote tape
	drives.  (2/22/89 ShJ)

onedspec$doc/refspectra.hlp
    A new help page for the refspectra task has been installed.
    (2/27/88, Davis)

onedspec$doc/continuum.hlp
    Added a warning about near zero divisions.  (2/14/89, Valdes)

onedspec$identify/idlinelist.x
onedspec$ecidentify/eclinelist.x
    Setting the coordinate line list to null no longer issues a warning.
    (2/13/89, Valdes)

onedspec$specplot.x
onedspec$doc/specplot.hlp
noao$lib/scr/specplot.key
    1. Added vertical shifts in scale.
    2. Added horizontal shifts in velocity.
    3. Added velocity and redshift colon commands.
    (2/8/89, Valdes)

onedspec$splot/splot.x
    The default key now prints the spectrum value at the x coordinate in
    addition to the cursor x, y coordinates.  (2/7/89, Valdes)

onedspec$dispcor/dispcor.x
onedspec$dispcor/ecdispcor.x
onedspec$dispcor.par
onedspec$ecdispcor.par
imred$coude/dispcor.par
imred$echelle/ecdispcor.par
imred$iids/dispcor.par
imred$irs/dispcor.par
imred$specphot/dispcor.par
onedspec$doc/dispcor.hlp
imred$echelle/doc/ecdispcor.hlp
    Changed "override" parameter to "rebin".  Also rebin=no acts only
    on nondispersion corrected spectra while rebin=yes acts only on
    dispersion corrected spectra.  (2/2/89, Valdes)

onedspec$dispcor/refaverage.x
onedspec$dispcor/reffollow.x
onedspec$dispcor/refgspec.x
onedspec$dispcor/refinterp.x
onedspec$dispcor/refmatch.x
onedspec$dispcor/refnearest.x
onedspec$dispcor/refprecede.x
onedspec$refspectra.par
onedspec$doc/refspectra.hlp
imred$coude/refspectra.par
imred$echelle/refspectra.par
imred$iids/refspectra.par
imred$irs/refspectra.par
imred$specphot/refspectra.par
    Added timewrap parameter and reorganized calling sequences so the
    sortval is set only in refgspec.  (2/2/89, Valdes)

onedspec$reidentify.x
    Stripped the image extension from the reference spectrum.
    (1/31/89, Valdes)

noao$lib/scr/ecidentify.key
    Fixed minor typo "j Go to next order" --> "k Go to next order".
    (1/26/89, Valdes)

onedspec$dcio.x
    An erroneous sfree in dc_gspec was removed. (1/26/89, Valdes)

onedspec$idsm_keywrds.x
onedspec$load_hdr.x
onedspec$dispcor/dispcor.x
onedspec$dispcor/ecdispcor.x
    Changed CRPIX usage to real.  (1/26/89, Valdes)

onedspec$names.par
imred$coude/names.par
imred$iids/names.par
imred$irs/names.par
    Made the "input" parameter prompt indicate it is a list rather than
    a single file.  (1/24/89, Valdes)

onedspec$splot.par
imred$coude/splot.par
imred$echelle/splot.par
imred$iids/splot.par
imred$irs/splot.par
imred$specphot/splot.par
    Made the minimum line number be 1 instead of 0.  (1/24/89, Valdes)

onedspec$splot/splot.x
    The 'w' window option in SPLOT now only redraws automatically in
    "auto" mode.  (1/24/89, Valdes)

onedspec$ecidentify/ecffit/ecffit.x
    The 'o' key now accepts the default order for fitting; i.e. a
    carriage return for the prompt.  Also the message about fitting
    now also includes the order offset being used.  (1/24/89, Valdes)

onedspec$idgdata.x
    Now allow 2D images with a second dimension of 1. (1/24/89, Valdes)

onedspec$dispcor/refinterp.x
    When interpolating on a parameter that is the same for a set of arcs
    and an object one wants two arcs to be identified; i.e. the one before
    and after.  This did not happen until this bug fix.  (1/20/89 Valdes)

onedspec$sensfunc.par
imred$echelle/sensfunc.par
imred$iids/sensfunc.par
imred$irs/sensfunc.par
imred$specphot/sensfunc.par
onedspec$standard.par
imred$echelle/standard.par
imred$iids/standard.par
imred$irs/standard.par
imred$specphot/standard.par
    Fixed missing default value for answer parameter.  (1/20/89, Valdes)

onedspec$splot/pixind.x
    Removed use of AINT function which was misbehaving on Sun386i.
    (12/16/88 Valdes)

onedspec$identify/reidentfy.x
onedspec$identify/idreidentfy.x
onedspec$identify/idreplot.x	+
onedspec$doc/reidentfy.hlp
onedspec$reidentfy.par
twodspec$longslit/reidentfy.par
imred$coude/reidentfy.par
imred$iids/reidentfy.par
imred$irs/reidentfy.par
imred$specplot/reidentfy.par
    Added plotfile for residuals.  (12/16/88 Valdes)

onedspec$dispcor/dcio.x
    If a reference spectrum is an image section its database entry will
    be the file with the section stripped.  Since the database entry
    is written by IDENTIFY I copied the database access code that
    strips the image section.  (12/8/88 Valdes)

onedspec$dispcor/dispcor.x
    The use of some real variables in the flux conservation calculation
    resulted in incorrect results when the resolution was very high.
    The code was carefully rewritten to do all possible calculations in
    double precision. (12/8/88 Valdes)

onedspec$t_specplot.x +
onedspec$specplot.par +
onedspec$specplot.h +
onedspec$doc/specplot.hlp +
noao$lib/scr/specplot.key +
onedspec$x_onedspec.x
onedspec$onedspec.cl
onedspec$onedspec.men
onedspec$onedspec.hd
onedspec$mkpkg
    New task added (12/7/88 Valdes)

onedspec$t_standard.x
    Fixed minor bug: missing parg in eprintf when dispersion solution
    missing.  (11/4/88 & 11/17/88)

onedspec$identify/ididentify.x
onedspec$identify/idfitdata.x
    The nonmonotonic error message was being lost because it is flushed
    immediately to the screen and then the screen is cleared to redraw
    the graph.  This has now been fixed by checking for an error just
    before the cursor read.  (11/2/88)

onedspec$identify/identify.x
onedspec$identify/iddb.x
onedspec$identify.par
onedspec$doc/identify.hlp
    1.  Added the additional icfit parameters (except naverage) to IDENTIFY
	so the user can set the default fitting parameters more fully.
    2.  All the ICFIT fitting parameters are now written to the database and
	read back.  This allows IDENTIFY and REIDENTIFY to start with exactly
	the same fitting parameters as previously used.  (11/2/88)

onedspec$t_bswitch.x
    Added a test for the extinction correction request before trying to compute
    the airmass.  (11/1/88)

onedspec$ecidentify/eccolon.x
    1.	When the label parameter was initially set to user all the labels
    were being printed not just those for the current aperture. The bug has
    been fixed. (9/9/88)

onedspec$dispcor/dispcor.x
    1.	A bug was fixed in the log+ option of dispcor and ecdispcor. The
	problem was that the end points of the wavelength region were in
	linear wavelength units but the w1 and dw parameters were in log units,
	causing an erroneous computation of the index for the first pixel.
	This bug has been fixed.  (9/9/88)

onedspec$dispcor/refspectra.x
onedspec$onedspec.cl
onedspec$onedspec.men
onedspec$batchred.cl +
onedspec$batchred.par +
onedspec$bswitch.par +
onedspec$coefs.par -
onedspec$standard.par
onedspec$sensfunc.par
    1.  BATCHRED and BSWITCH were put back into this package.
    2.  COEFS was removed from this package.
    3.  Enumerated strings were added to SENSFUNC and STANDARD parameter
	files to prevent the tasks from dying on a bad value (i.e. clgwrd
	was causing an error).  By putting the allowed values in the parameter
	file the CL will wait for an allowed value.
    4.  REFSPECTRA does not change the value of the confirm parameter now.
	(7/29/88 Valdes)


onedspec$splot/deblend.x
onedspec$doc/splot.hlp
    1.  The fitting parameter initialization was being done even before the
	'q'.  Thus, the '-' subtraction did not use the fit but the initial
	parameters.
    2.  Modified the initial sigma to be 1/4 of the range divided by the number
	of lines.  The 1/2 was too large.  (7/26/88 Valdes)

onedspec$splot.par
onedspec$splot/deblend.x
onedspec$splot/scr_help.x -
onedspec$doc/splot.hlp
    1.  Removed unused parameters inblend, fixsep, difference, subtract from
	parameter file.
    2.  Fixed bug with '-' in deblending (continuum was not being subtracted).
    3.  Removed unused source file.
    4.  Update the help page.  (7/19/88 Valdes)

onedspec$splot/deblend.x
    Fixed bug introduced below.  (7/12/88 Valdes)

onedspec$splot/deblend.x
onedspec$onedspec.hd
noao$lib/scr/deblend.key
onedspec$doc/splot.hlp
    1.  After moving the parameter initialization to within the options loop the
        initializations were being done wrong.
    2.  The 'd' option was not doing what it was supposed to.
    3.  Added a print newline to clear the status line if four lines were
	entered since this does not go through the 'q' case which was
	doing the clear.
    4.  The n sigma cases had the wrong mneumonics in the help.
    5.  The src definitions in the help table were pointing to wrong files
	since the names and directories for the files have been changed
	(7/1/88 Valdes)


onedspec$t_names.x
onedspec$mkpkg
    Modified this task to use the ODR package.  This also strips the image
    extension allowing the append option to work.  (6/28/88 Valdes)

onedspec$coincor.x
    When doing both coincidence and power law corrections failed to put the
    output of the coicidence correction as the input to the power law
    correction.	(6/23/88 Valdes)

onedspec$identify/idgdata.x
    Added an error check to IMMAP.  Failure to do this gave a segmentation
    violation on the SUNS.	(6/23/88 Valdes)

onedspec$continuum.cl
    1. Added a parameter to allow a cursor list text file to be passed to the
    normcontinuum task.

onedspec$ecidentify/ecgdata.x
onedspec$ecidentify/ecffit/ecfcolon.x
onedspec$ecidentify/ecffit/ecfset.x
onedspec$ecidentify/ecffit/ecfsolve.x
onedspec$ecidentify/ecffit/ecfrms.x +
onedspec$ecidentify/ecffit/mkpkg
noao$lib/scr/ecidentify.key
noao$lib/scr/ecffit.key +
    The following was fixed.				(5/20/88 Valdes)
    1.	Error in graph title string.
    2.	Missing cursor key help.
    3.	Error in ":function" command in fitting mode.
    4.	Rms calculated with deleted points.

onedspec$dispcor/dispcor.x
onedspec$dispcor/ecdispcor.x
    1.	Failed to initialize the output spectrum to zero so that points
	outsided the input data range are zero.  (5/17/88 Valdes)

onedspec$dispcor/refaverage.x
    1.  Instead of checking the reference spectra for aperture and reference
	flag it was test the input image.  This was changed. (5/17/88 Valdes)

onedspec$load_hdr.x
onedspec$splot/deblend.x
    1. The deblending was fitting a function without the factor of 2 in
       the Gaussian sigma definition.  This caused the printed Gaussian
       parameters to be off by a factor of sqrt(2).
    2. Slight change to not have the header loading change the specified
       input line.  It is up to the calling code to determine if this is
       a valid line.  (5/17/88 Valdes)

onedspec$identify/idreidentify.x
    1. Added check for nonmonotonic dispersion solution.  (4/30/88)

onedspec$onedspec.cl
onedspec$onedspec.men
onedspec$onedspec.hd
    1. Task EXTINCT was removed.  The script and help page remain in case
    they are desired.  Later they will also disappear.  The function of
    this script is replaced by CALIBRATE.  (4/26/88 Valdes)
    2. Task BATCHRED was removed to the IMRED packages.  (4/27/88 Valdes)

onedspec$splot/splot.x
onedspec$splot/deblend.x
onedspec$t_flatfit.x
onedspec$identify/ididentify.x
onedspec$ecidentify/ecidentify.x
onedspec$ecidentify/ecffit/ecffit.x
noao$lib/scr/splot.key
noao$lib/scr/identify.key
noao$lib/scr/ecidentify.key
noao$lib/scr/deblend.key
noao$lib/scr/ecffit.key
noao$lib/scr/flatfit.key
    Added 'I' interrupt key. (4/20/88 Valdes)

onedspec$identify/identify.h
    Variables defined as integers instead of real (ID_MATCH, ID_MINSEP)
    (4/18/88 Valdes)

onedspec$sensfunc/t_sensfunc.x
onedspec$sensfunc/sfsensfunc.x
onedspec$sensfunc/sfstds.x
onedspec$sensfunc/sfginit.x
onedspec$sensfunc/sfoutput.x
onedspec$sensfunc.par
onedspec$doc/sensfunc.hlp
noao$lib/scr/sensfunc.key
    1. Added beam number to output sensitivity image header.
    2. Added 'I' interrupt key.
    3. Added aperture number selection.
    4. Added interactive query.  (4/15/88 Valdes)

onedspec$splot/getimage.x
    Modified to recognize echelle format spectra on input.  (4/8/88 Valdes)

onedspec$load_hdr.x
    Modified to recognize echelle format spectra on input.  (4/8/88 Valdes)

onedspec$mkpkg
onedspec$splot/mkpkg
onedspec$identify/mkpkg
onedspec$fortran/mkpkg
onedspec$onedutil.cl
onedspec$onedspec.cl
onedspec$onedspec.par
onedspec$onedspec.men
onedspec$onedspec.hd
onedspec$identify/*
onedspec$t_flatdiv.x
onedspec$t_coefs.x
onedspec$t_combine.x
onedspec$dispcor.par
onedspec$identify/identify.par --> onedspec$identify.par
onedspec$identify/reidentify.par --> onedspec$reidentify.par
onedspec$doc/dispcor.hlp

onedspec$dispcor/* +
onedspec$ecidentify/* +
onedspec$x_onedspec.x +
onedspec$refspectra.par +
onedspec$dispcor1.par +
onedspec$ecidentify.par +
onedspec$ecreidentify.par +
onedspec$doc/refspectra.hlp +

onedspec$x_wavecal.x -
onedspec$x_fluxcal.x -
onedspec$x_onedutil.x -
onedspec$identify/x_identify.x -
onedspec$identify/libpkg.a -
onedspec$dbx/ -
onedspec$dbxio.h -
onedspec$userstd/ -
onedspec$t_dispcor.x -
onedspec$fudge.x -
onedspec$rlsq.x -
onedspec$userstdc.x -
onedspec$readstd.x -
onedspec$qsortra.x -
onedspec$statfile.x -
onedspec$ascrcomp.x -
onedspec$identify/icghelp.x -
onedspec$splot/spflip.x -
    The ONEDSPEC package has been completely reorganized by combining
    executables, eliminating obsolete procedures, and adding new
    versions of IDENTIFY and DISPCOR as well and new tasks for echelle
    format data.  (4/7/88 Valdes)

-------------------------------------------------------------------------------

onedspec$mkpkg
onedspec$splot/deblend.x
    Fixed bugs related to initial guesses for width and peak and scaling.
    Replaced Gauss-Jordan routine by Householder transformation routine
    for stability.  (4/6/88 Valdes)

onedspec$load_hdr.x
    The test for wavelengths in meters per second was W0 < 0.001.
    Now the test is abs(W0) < 0.001.  (3/10/88 Valdes)

onedspec$identify/ididentify.x
    The 't' was calling fit_to_pix with the real valued cursor position
    while the procedure expects a double.  Added a double coercion to fix
    the bug.  (2/18/88 Valdes)

onedspec$splot/anssave.x
onedspec$splot/mktitle.x
onedspec$splot/getimage.x
onedspec$splot/splot.x
onedspec$splot/splotfun.x
    1.  Titles (on the graph and in the log file) for two dimensional images
	now contain the line number given as an image section.
    2.  The log file title now includes a time stamp. (1/29/88 Valdes)

onedspec$identify/ididentify.x
    When recentering all the features the fitted coordinates are now
    updated and the tick marks moved to the new center position.
    (1/4/87 Valdes)

onedspec$identify/iddb.x
    DBGETR was declared as real for the new shift parameter causing a wrong
    wavelength scale to appear. (12/22/87 Valdes)

onedspec$doc/identify.hlp
    Fixed minor typo. (12/7/87 Valdes)

onedspec$sextract.cl +
onedspec$doc/sextract.cl +
onedspec$onedutil.cl
onedspec$onedutil.par
onedspec$onedutil.men
onedspec$onedutil.hd
    Added a new task, SEXTRACT, to extract subspectra. (11/19/87)

onedspec$t_dispcor.x
    The default starting wavelength and wavelength interval are now printed
    in g format so that the user sees the full value.  (11/9/87)

onedspec$identify/identify.x
onedspec$identify/reidentify.x
onedspec$identify/idgraph.x
    1.  The XTOOLS change to XT_MK1D now permits the sections "column 51"
	and "column 051" to be recognized identically.
    2.  REIDENTIFY now aborts with a useful error message if their is not
	database record for the reference image instead of later causing
	a segmentation error.
    3.  IDENTIFY can now plot in point mode using the GTOOLS commands if
	desired.
    (11/9/97 Valdes)

noao$onedspec$sensfunc/sfextinct.x
noao$onedspec$sensfunc/sfsensfunc.x
noao$onedspec$sensfunc/sfreset.x
noao$onedspec$sensfunc/sfmarks.x
noao$onedspec$sensfunc/sfadd.x
noao$onedspec$sensfunc/sfdelete.x
noao$onedspec$sensfunc/sfundelete.x
noao$onedspec$sensfunc/sfmove.x
noao$onedspec$sensfunc/sfgraph.x
noao$onedspec$sensfunc/sfginit.x
noao$onedspec$sensfunc/sfcomposite.x
noao$onedspec$sensfunc/sfcolon.x
noao$onedspec$sensfunc/sfshift.x
noao$onedspec$sensfunc/sensfunc.h
noao$onedspec$doc/sensfunc.hlp
noao$lib/scr/sensfunc.key
    A number of changes were made based on user comments.
    1.  A bug was fixed which caused the ":order" command to crash the task.
	The integer valued order was being passed as a char in the colon
	decoding task.
    2.  The shift key 's' now toggles allowing a shift to be undone
	without initializing all the data.  Also a message is printed
	to indicate what has been done.
    3.  The composite key 'c' now toggles allowing a composite to be undone
	without initializing all the data.  Also a message is printed
	to indicate what has been done.  A deleted composite point deletes
	the original data at that wavelength when toggling back.
    4.  The extinction key 'e' now toggles allowing an extinction
	correction to be undone without initializing all the data.
	Also a message is printed to indicate what has been done.
    5.  A different symbol may be used to indicated added points.
    6.  Changing the function or order does not automatically do a
	new fit.
    7.  A new key 'g' was added to do a fit and redraw the graph(s).
	The existing 'f' key does a fit and overplots as before.
    (11/6/87 Valdes)

onedspec$splot/replot.x
onedspec$splot/splot.x
onedspec$splot/autoexp.x
    Modified REPLOT to use GTOOLS task GTVPLOT.  This allows the user to
    select point mode.  The calling sequence for REPLOT has a new argument
    to allow calling this procedure for overplotting.  (11/5/87 Valdes)

onedspec$identify/*
onedspec$identify/iddoshift.x +
    Added shift options to IDENTIFY and a refit option to REIDENTIFY.
    This allows maintaining the same coordinate function with an additive
    shift. (11/3/87 Valdes)

onedspec$sensfunc/sfgraphs.x -
onedspec$sensfunc/mkpkg.x
    A zero length file, possibly confused with sfgraph.x was deleted and
    deleted from the mkpkg. (10/26/87 Valdes)

onedspec$splot/deblend.x
onedspec$splot/sumflux.x
    1. The input data to the deblending routine are now scaled to values
    near unity.  Also the fitting is iterated three times to make the
    results more consistent.
    2. When computing the line center with 'e' the data is scaled to
    avoid underflows in summing residuals to the 1.5 power.
    (See bug report 16) (10/22/87 Valdes)

onedspec$sensfunc/sfsensfunc.x
onedspec$sensfunc/sfextinct.x
    1. Aperture number for new aperture in title was undefined in the first
    graph.  Set title after determining aperture number.
    2. In a rare case it was possible for a square root of zero to occur
    in the extinction significance calculation which is fatal on VMS.
    Added check of argument before square root call. (Valdes)

onedspec$splot.par
    Changed all interactive query parameters from auto mode to query
    mode to force a query even when run in menu mode and with :go.
    (9/15/87 Valdes)

onedspec$t_standard.x
onedspec$t_lcalib.x
onedspec$splot/plotstd.x
onedspec$standard.par
onedspec$lcalib.par
onedspec$splot.par
onedspec$doc/standard.hlp
onedspec$doc/lcalib.hlp
onedspec$doc/splot.hlp
    The magnitude to absolute flux conversion constant has been made a
    user changable parameter in the three tasks dealing with the flux
    calibration tables.  (9/3/87 Valdes)

onedspec$t_sensfunc.x -
onedspec$sensfunc/* +
noao$lib/scr/sensfunc.key +
onedspec$sensfunc.par
onedspec$doc/sensfunc.hlp
onedspec$t_standard.x
onedspec$doc/standard.hlp
onedspec$bswitch.par
onedspec$getextn.x
    SENSFUNC has been completely rewritten.  It now allows determination
    of extinction, display of flux calibrated spectra, and many nice
    features for displaying and manipulating the data.  For full details
    read the new help page.

    The new sensfunc required some modifications to STANDARD in the
    format of the output file produced by standard.  The parameters for
    BSWITCH no longer have the grey scale parameter add_const or the
    (never implemented) revised extinction file rev_ext_file which are
    not produced by SENSFUNC any more.

    The extinction loading procedure was modified to allow a null
    extinction file to correspond to no extinction and to eliminate the
    procedure get_new_ext and fix_ext which were used for the old grey
    constant and never implemented revised extinction file. (9/3/87
    Valdes)

onedspec$splot/mkpkg
onedspec$splot/splot.x
onedspec$splot/splotfun.x
    Errors getting a spectrum in function mode were ignored and the spectrum
    was replotted.  Changed to return the error as a warning and not redraw
    the plot.

onedspec$t_dispcor.x (routine reinterp)
    The reinterpolation now has additional tests:
    1)  When the interpolation point is within a minimum distance of
	an input pixel (0.001) it does not interpolate.  This was
	done because the interpolation grid is sometimes meant to be
	identical with the input but the computation of the output grid
	is very slightly off (this was observed in COMBINE).
    2)	If one of the points to be interpolated between has a value of
	0.0 (used to mark missing data in ONEDSPEC) then the rebinned
	point is set to 0.0 in order to propagate the missing point.
	This is important for combining spectra with COMBINE. (8/5/87 Valdes)

====
V2.5
====

onedspec$t_sinterp.x
    Valdes, June 22, 1987
    1.  Removed a warning message to allow comments in the input table.

onedspec$splot/avgsnr.x
    Valdes, June 19, 1987
    1.  A possible type of data is Fnu calibrated data with values in the
	range 1e-25.  Attempting to determine an average, rms, and
	signal-to-noise ratio with SPLOT caused a divide by zero error
	due to underflowing the sum of squares.  This has been modified
	to shift and scale the data before computing the sum of squares.

onedspec$t_standard.x
    Valdes, June 12, 1987
    1.  There was an uninitialized memory problem with the space allocated
	for adding points.  This bug was introduced with the May 15th
	modifications to the structure of the calibration files.

onedspec$load_hdr.x
onedspec$idsm_keywrds.x
onedspec$t_calibrate.x
    Valdes, June 9, 1987
    1.  Added EXPTIME as a recognized exposure time keyword.
    2.  Added check against INDEF or 0 exposure time in CALIBRATE.

onedspec$bplot.cl
    Valdes, June 4, 1987
    1.  The BPLOT script is now back the way it was earlier because the
	earlier bug with the CL and list files seems to have gone away
	while the new script relies on writing to parameter files which
	doesn't work in the background.

onedspec$onedspec.cl
onedspec$onedspec.hd
onedspec$onedspec.men
onedspec$powercor.cl +
onedspec$powercor.par +
onedspec$getcalib.x
onedspec$doc/powercor.hlp +
    Valdes, June 1, 1987
    1.  Added task POWERCOR from IIDS.
    2.  Added an error check for a bad extinction file.

onedspec$splot/deblend.x
    Valdes, May 19, 1987
    1.  A bug that was introduced into deblending during the last set of
	changes was fixed.

onedspec$onedutil.par
onedspec$lcalib.par
onedspec$t_lcalib.x
    Valdes, May 19, 1987
    1.  Make the default for the calibration parameters in LCALIB to
	be package parameters of the same name in keeping with the way these
	parameters are used in the other ONEDSPEC tasks.
    2.  Added the calibration parameters to the ONEDUTIL package and
	the default is to refer to the parameters of the package that loaded
	it.  This will be either ONEDSPEC or one of the IMRED packages.
    3.  Modified LCALIB to not require the extinction file when reading
	star calibration info.

onedspec$mkpkg
onedspec$bswitch.par
onedspec$lcalib.par
onedspec$onedspec.par
onedspec$splot.par
onedspec$standard.par
onedspec$t_lcalib.x
onedspec$t_standard.x
onedspec$x_fluxcal.x
onedspec$x_onedutil.x
onedspec$getcalib.x
onedspec$getextn.x
onedspec$plotstd.x -
onedspec$splot/mkpkg
onedspec$splot/plotstd.x +
onedspec$doc/standard.hlp
onedspec$doc/lcalib.hlp
onedspec$doc/onedspec.hlp
onedspec$doc/splot.hlp
onedspec$doc/bswitch.hlp
noao$imred/echelle (par files)
noao$imred/iids (par files)
noao$imred/irs (par files)
noao$imred/specphot (par files)
noao$lib/onedstds (data files)
    Valdes, May 15, 1987
    1.  The major change was to change the format of the calibration data
	from the very constrained old format to a more flexible format.
	This also involved adding a new parameter "extinction" and changing
	"calib_file" to "caldir".
    2.  The calibration data files were converted to the new format in a
	number of subdirectories.
    3.  The parameter files in the IMRED directories were also updated.
    4.  Moved plotstd.x to splot directory.  It is an splot routine and should
	be with the other splot source.
    5.  Moved LCALIB from the FLUXCAL executable to the ONEDUTIL executable.

onedspec$splot/usercoord.x
    Valdes, May 8, 1987
    1.  When setting a wavelength scale using the 'u' key on data lacking
	any wavelength information (W0 and WPC == INDEF) there was a bug
	causing a message of the form "cursor not moved".

onedspec$splot/deblend.x
onedspec$splot/splot.x
onedspec$splot/eqwidthcp.x
    Valdes, April 30, 1987
    1.  I missed a couple of places where READ_WRITE access was used
	in SPLOT (see March 13, 1987).  These have been removed.
    2.  There was a bug in the 'k' and 'v' type equivalent width
	procedures which produced wrong results unless the cursor was
	very near the center.
    3.  When applying deblending to a single line the starting position
	is now the minimum or maximum point of the continuum subtracted
	profile rather than the center of the continuum limits.

onedspec$splot/deblend.x
onedspec$splot/splot.x
onedspec$splot/anssave.x
onedspec$splot/eqwidthcp.x
onedspec$splot/eqwidth.x
onedspec$doc/splot.hlp
noao$lib/scr/splot.key
    Valdes, April 28, 1987
    1.  SPLOT now prints only one line of output on the graphics status line
	when doing deblending or equivalent width measurments.  The full
	output is saved in the log file and also internally.  These changes
	were made to allow reasonable behavior in terminals which cannot
	display text and graphics simultaneously (PC emulators, VT240's).
    2.  To get the full output of previous measurements during the course of
	the task execution a new command ":show" has been added.
    3.  It was possible for deblending to yeild negative sigmas.  This has been
	fixed as well.

onedspec$doc/names.hlp
    Valdes, April 27, 1987
    1.  A bug note was added to the task help stating that the append option
	is intended only for image sections.  Appending any other string
	produces names not acceptable to ONEDSPEC.

onedspec$identify/identify.x
onedspec$identify/ididentify.x
onedspec$identify/idlinelist.x
onedspec$identify/idnewfeature.x
    Valdes, April 15, 1987
    1.  Added bell if feature not found with 'm'.
    2.  When automatically identifying lines, 'l' it now requires a new line
	to be within the matching distance relative to the current fit and
	if two centers are withing "minsep" then the closest match to the
	user coordinate is selected.
    3.  Default initial graph for fitting is residuals vs. wavelength.

onedspec$t_standard.x
    Davis, April 13, 1987
    1. At Frank's suggestion I added a test in STANDARD to make sure that
    the exposure time is never less than 1 second.

onedspec$t_standard.x
    Davis, April 10, 1987
    1. In order to check for an INDEF valued exposure time STANDARD on VMS/IRAF
    was testing a boolean compared to a fp 0.0. The test was always coming up
    true if the exposure keyword was defined; and exposure time was being set
    to 1. If no exposure keyword was present INDEFI was being used for the
    exposure time. I changed the test to test for an integer INDEF and
    every thing seemed ok. Lyra, IRAF and IRAFX were updated.

onedspec$t_standard.x
    Valdes, April 3, 1987
    1.  STANDARD was using INDEF if there was no exposure time in the
	header rather than the intended 1.0 as described in the
	documentation.  It now uses 1 for the exposure time if there
	is no exposure time in the header.

onedspec$coincor.x
    Valdes, March 23, 1987
    1.  In the power correction the value of the output when the input
	was negative was undefined.  Now it is the input value.

onedspec$splot/getimage.x
onedspec$splot/wrspect.x
onedspec$splot/deblend.x
onedspec$splot/eqwidth.x
onedspec$splot/eqwidthcp.x
    Valdes, March 13, 1987
    1.  SPLOT no longer opens the image READ_WRITE.  This was unnecessary
	and would prevent someone from examining data for which they don't
	have write permission.
    2.  Modified the deblend and eqivalent width options to deactivate the
	workstation since they produce multiline output.

onedspec$t_dispcor.x
onedspec$dispcor.par
onedspec$doc/dispcor.par
    Valdes, March 5, 1987
    1.  It is now a fatal error if the dispersion solution (from IDENTIFY)
	is nonmonotonic.
    2.  The starting wavelength and wavelength intervals are now list
	structured parameters to allow files containing the values to
	be used.  With no file the user is queried and a carriage
	return or nonnumeric value will use the default value.
    3.  The way wavelength information is printed out has been improved.
    4.  A missing carriage return was added to the error message when
	an image is not found.
    5.  The order of the parameters, some default values, some of the
	prompts, and their modes have been changed to be more consistent
	with other tasks and more easily useable with command line arguments.
    6.  The help page was modified to reflect these changes.

onedspec$identify/ididentify.x
onedspec$identify/idreidentify.x
onedspec$identify/idfitdata.x
onedspec$identify/idcolon.x
    Valdes, March 5, 1987
    1.  IDENTIFY now prints a warning about a nonmonotonic coordinate
	solution.
    2.  Changes were made to not print the current feature when error
	messages are printed thus giving the user a change to read them.
    3.  When attempting to change images to a nonexistant image
	the immap was improperly error checked.  This could result in
	fatal errors (particularly on VMS).

onedspec$dispcor.par
    Valdes, February 27, 1987
    1.  Prompt was changed from
	"File containing ..." to "Database containing ..."

onedspec$userstd/nearpt.x
onedspec$oned.h
    Valdes, February 25, 1987
    1.  Changed nearest point algorithm to use NDC coordinates. This required
	adding the GIO pointer to the arguments.
    2.  Change all procedures calling near_pt to include GIO pointer
	argument.
    3.	Changed maximum distance to 0.05 (NDC)

onedspec$splot/splot.x
    Valdes, February 25, 1987
    1.  When exiting from the 'f' function mode in SPLOT the function
	status line is now erased.

noao$onedspec
    Valdes, February 19, 1987
    1.  Made required GIO modifications.  The tasks affected are SPLOT,
	STANDARD, FLATFIT, SENSFUNC, and IDENTIFY.  Please report any
	bugs.

onedspec$coincor.x
onedspec$t_coincor.x
onedspec$t_flatdiv.x
onedspec$t_flatfit.x
onedspec$doc/coincor.hlp
    Valdes, February 9, 1987
    1.  A number of interface errors were fixed.
    2.  The coincidence correction procedure now takes an input and output
	array.  Previously it modified the given array.
    3.  The basic IIDS correction is now checked for values which would
	cause the log function to give an exception or instruction error.
    4.  The major change in COINCOR is that if the output root image name
	is null then the operation is done in place.  When dealing with
	~1000 images this saves on disk space and directory manipulations.
    5.  The help page for COINCOR was appropriately updated.

onedspec$fortran/polft1.f
onedspec$getextn.x
onedspec$t_calibrate.x
onedspec$t_sensfunc.x
    Valdes, February 5, 1987
    1.  The following errors reported by Skip Schaller (Steward Obs, AOS port)
	were fixed.
	polft1.f:	Minus sign out of place in expression
	getextn.x:	Remove declaration for max(), min(), log10()
	t_calibrate.x:	Remove declaration for min()
	t_sensfunc.x:	Remove declaration for log10()

onedspec$oned.h
    Valdes, January 30, 1987
    1.  The maximum number of beams the package can handle has been
	increased from 50 to 100.

onedspec$t_combine.x
onedspec$combine.par
onedspec$doc/combine.hlp
    Valdes, January 30, 1987
    1.  An new parameter called "combine" was added which specifies the type
	of combining (either average or sum).  The help documentation was
	updated.

onedspec$identify/idcolon.x
onedspec$identify/ididentify.x
    Valdes, January 16, 1987
    1.  Colon command dictionary and switch rewritten to use macro definitions.
    2.  ? help facility rewritten to use system paging facility instead of ad
	hoc menu facility.

onedspec$gcurval
    Valdes, January 12, 1987
    1.  Changed "0 0 0 q" to "0 0 1 q" since this was detected as an error
	in V2.5.  This file is used by BPLOT.

onedspec$batchred.cl
noao$imred/iids/batchred.cl
noao$imred/irs/batchred.cl
    Valdes, December 29, 1986
    1.  This script creates the user script "process.cl".  It was creating
	it with an out-of-date syntax which no longer worked.  Modified
	BATCHRED to create a valid script.

onedspec$lcalib.par
    Valdes, December 18, 1986
    1.  The default for the calibration file in task LCALIB is now that
	for the task STANDARD.

onedspec$identify/idreidentify.x
    Valdes, December 3, 1986
    1.  REIDENTIFY was not correctly tracking when there was no fit.
	75: FIT(j) = FIT(i)	==>	FIT(j) = fit

onedspec$t_flatfit.x
onedspec$t_flatdiv.x
onedspec$flatfit.par
onedspec$flatdiv.par
onedspec$doc/flatfit.hlp
onedspec$doc/flatdiv.hlp
    Valdes, December 2, 1986
    1.  The tasks FLATFIT and FLATDIV may optionally apply coincidence
	corrections.  They were not updated to include the IIDS nonlinear
	correction made earlier.  They have now been updated.

onedspec$t_bswitch.x
onedspec$t_flatfit.x
onedspec$t_sums.x
    Valdes, December 1, 1986
    1.  The tasks BSWITCH, FLATFIT, and SUMS created new images with only
	the standard ONEDSPEC header information and without any other
	user parameters.  These tasks worked this way because they may
	sum many spectra for each beam and the connection between the
	input image header and output image header was not obvious.  They
	have been modified to use the last input image for each beam as
	the image header template for the output image of that beam.
	When there is no summing then the output image header will be
	a copy of the input image header with updated ONEDSPEC parameters.

onedspec$identify/idlinelist.x
    Valdes, November 25, 1986
    1.  It used to be that if there were no coordinate list then the
	default user coordinate was the pixel coordinate.  This changed
	at some point.  This has been fixed.

onedspec$identify/identify.x
    Valdes, November 21, 1986
    1.  The common variable labels is now initialized every time the
	task runs.

onedspec$load_hdr.x
onedspec$splot/splot.x
onedspec$splot/usercoord.x
    Valdes, November 17, 1986
    1.  Since people insist on using W0 and WPC to define the wavelength
	coordinates and are then confused because CRVAL1 and CDELT1 are
	used I changed the default precedence.  The ONEDSPEC package now
	looks for W0 and WPC first and then resorts to the FITS coordinate
	keywords.  Also if the coordinate values are less the 0.001
	it assumes that the units are meters and converts to Angstroms.
	This arises when a strict interpretation of the FITS coordinates
	(units of meters) is used for optical spectral data.
    2.  The key 'p' in SPLOT has been modified to query for the starting
	and ending wavelength.  The default values are those last defined.
	Thus, this key may be used at any time to set the wavelength scale.
	To return to wavelength scale after '$' the user simply types
	carriage return to accept the defaults.
    3.  The key 'u' in SPLOT has been modified to work in all cases.
	Previously it only worked if the plot was in pixel coordinates.
	If run in wavelength coordinates funny results would be obtained.
	Now the user may mark two points even in wavelength coordinates.

onedspec$coincor.x
    Valdes, November 13, 1986
    1.  The power law correction is applied only to positive data.
	Negative data is not changed.

onedspec$splot/eqwidth.x
onedspec$splot/deblend.x
onedspec$splot/eqwidthcp.x
    Valdes, November 3, 1986
    1.  Changed print format statements to keep columns from running together
	for flux calibrated data.

onedspec$splot/*.x
onedspec$splot/mkpkg
onedspec$splot/idsmtn.h -
onedspec$splot/oned.h -
    Valdes, October 28, 1986
    1.  Changed include references to point to include files in the main
	package directory ("idsmtn.h" -> "../idsmtn.h" and
	"oned.h" -> "../oned.h").
    2.  Deleted the copies of the include file in this directory.

onedspec$t_coincor.x
onedspec$coincor.x
onedspec$coincor.par
onedspec$doc/coincor.hlp
onedspec$oned.h
onedspec$onedspec.par
onedspec$onedspec.men
    Valdes, October 21, 1986
    1.  Modified COINCOR to include a power law correction as part of the
	IIDS correction.
    2.  A new paramter was added to COINCOR and ONEDSPEC, called "power",
	for the IIDS power law correction.
    3.  The help page for COINCOR was revised.

onedspec$splot/splot.x
onedspec$splot/getimage.x
onedspec$splot/wrspect.x
    Valdes, October 20, 1986
    1.  Added ability to write modified spectrum to the current image in
	SPLOT.
    2.  There were several errors in the code which were fixed.  These
	included modifying an IMIO buffer and extra arguments.

onedspec$splot/splot.x
onedspec$splot/eqwidth.x
onedspec$splot/eqwidthcp.x
onedspec$splot/deblend.x
onedspec$splot/saveans.x
onedspec$doc/splot.hlp
    Valdes, October 15, 1986
    1.  The routines for the keys 'd', 'e', 'h', 'k', and 'v' now print
	information in a same format.  They all have a header line and
	a line containing the values.  There reason for this is that,
	with the additional information now included, it requires two
	lines for "quantity: value" format anyway.  They also print the
	information which is common to all methods in the same order.
    2.  The deblending routine 'd' now includes the continuum, equivalent
	width, and sigma of the Gaussian fits.  It also plots the continuum
	slope as is done with the 'e' key.
    3.  The equivalent width routine 'e' now includes the continuum.
    4.  The 'h', 'k', and 'v' routines now include flux and FWHM.
    5.  The 'h', 'k', and 'v' routines now work on emission lines as well
	as absorption lines.
    6.  The 'h', 'k', and 'v' routines define the gaussian profile in the
	same way as the deblend routine; i.e. exp (-0.5 * (dw/sigma)**2)
    7.  Help revised.

onedspec$splot/splot.x
onedspec$splot/autoexp.x
    Valdes, October 14, 1986
    1.	The SPLOT windowing keys 'a', 'z', ',', and '.' were not compatible
	with the GTOOLS windowing.  AUTOEXP.X was rewritten to use the
	GTOOLS structure while operating as before.

onedspec$splot/splot.x
onedspec$splot/eqwidthcp.x
onedspec$splot/scrhelp.x
onedspec$splot/stshelp.x
onedspec$doc/splot.hlp
    Valdes, October 8, 1986
    1.  There are two methods of measuring equivalent widths using a simple
	Gaussian line model.  The original method which requires a unit
	continuum has been restored as the 'k' key. (See the revision
	of September 18, 1986).
    2.  The second method recently added which uses the y cursor to mark
	the continuum and uses the half flux level for determining the
	line width is available with the last available key; the 'v' key.
    3.  The 'h' key for one sided measurements still requires a second key
	but now in addition to defining which side of the line to use
	it also defines which method to used.
    4.  The help page has been updated to reflect the changes.

onedspec$doc/rebin.hlp
    Valdes, October 7, 1986
    1.  Typo in V2.3 documentation fixed: "set to know" -> "set to no".

onedspec$t_shedit.x +
onedspec$shedit.par +
onedspec$shparams.par +
onedspec$doc/shedit.hlp +
onedspec$onedspec.cl
onedspec$onedspec.men
onedspec$onedspec.hd
    Valdes, September 29, 1986
    1.  A onedspec header editor called SHEDIT has been added.  It uses
	EPARAM as the editor.
    2.  A help page is available.

onedspec$identify/reidentify.x
    Valdes, September 25, 1986
    1.  REIDENTIFY was passing a constant 0. to ID_REIDENTIFY which expects
	a double.  Replaced 0. with "double (0.)" as the argument.
	This caused a failure in the AOS IRAF.

onedspec$splot/eqwidthcp.x
onedspec$splot/doc/splot.hlp
    Valdes, September 18, 1986
    1.  The 'k' key used to determine equivalent widths by fitting a Gaussian
	profile based only on the depth of the core, the line width at some
	point, and the continuum had several problems.  First, people failed
	to realize that the continuum had to be 1.  Second, the y cursor
	position was used for measuring the width of the line.  Third, if
	the y cursor position was not within the line then square root and
	logarithm exceptions occured.  These problems have been fixed as
	follows:
	    1.  The y cursor is now used to mark the continuum.  This
	    has been made very clear in the documentation.
	    2.  This allows equivalent widths to be measured for any
	    absorption line even when the continuum is not 1!
	    3.  The level at which the width of the line is measured is
	    now the point half way between the continuum and the minimum
	    point in the line.  Previously this point was set by the
	    y cursor position.
	    4.  If the y cursor position is below the line minimum or
	    the left and right edges of the line are not found at the half
	    flux point an informative error is printed and the equivalent
	    width is not evaluated.
	    5.  The search for the left and right edges was previously
	    limited to +- 9 pixels.  This limit has been removed.  The
	    search now extends to the limits of the spectrum if necessary.
	    6.  The information printed includes the gaussian parameters
	    as well as the equivalent width.
	    7.  The gaussian model is plotted over the spectrum in order
	    to judge the reasonableness of the equivalent width measurement.

onedspec$splot.par
onedspec$doc/splot.hlp
    Valdes, September 11, 1986
    1.  Added ? to boolean prompts.  The prompt
	    Fix separation of lines:
	was confusing a user who tried to give the value of the separation.
	The new prompt is
	    Fix separation of lines?:
    2.  This parameter was also not in the documentation!

onedspec$t_dispcor.x
    Valdes, September 11, 1986
    1.  DISPCOR requires reference spectra to exist as well as the identify
	database entry.  The error message was misleading.  The error message
	is now more specific.

onedspec$splot/splot.x
onedspec$splot/anssave.x
    Valdes, September 8, 1986
    1.  Modified SPLOT to append to the answer file each time an aswer is
	written rather than opening the answer file at the beginning and
	closing it at the end.  This eliminates the annoying creation of
	a file everytime SPLOT is used.

onedspec$t_dispcor.x
    Valdes, September 8, 1986
    1.  Procedure dcorrect was defined as a function but used as a subroutine.
	This was found and corrected during the Alliant port.

onedspec$identify/xtpage.x +
onedspec$identify/xtmenu.x +
onedspec$identify/ididentify.x
    Valdes, September 5, 1986
    1.  Added paging and menu features to '?' help.

onedspec$bplot.cl
    Valdes, August 26, 1986
    1.  The BPLOT script has been rewritten.  Rather than calling SPLOT
	in a loop, once for each image, a cursor command file is created
	containing cursor commands for all the images and then SPLOT is
	called with a list of images.  This fixes an undiagnosed bug and
	is more efficient.

onedspec$identify/ididentify.x
onedspec$identify/iddofit.x
onedspec$identify/idgdata.x
onedspec$identify/idfitdata.x
    Valdes, August 22, 1986
    1.  ICFIT no longer inherits the window from IDENTIFY.  Entering ICFIT
	will do autoscaling.
    2.  IDENTIFY now uses the image header coordinate information if there
	is no database dispersion solution.  The parameters used are
	CRPIX, CRVAL, and CDELT.  This allows IDENTIFY to be used with
	linearized spectra in the ONEDSPEC related packages.

onedspec$identify/identify.com
onedspec$identify/identify.x
onedspec$identify/idcenter.x
onedspec$identify/idcolon.x
onedspec$identify/idshow.x
onedspec$identify/reidentify.x
onedspec$identify/identify.par
onedspec$identify/reidentify.par
    Valdes, August 18, 1986
    1.  IDENTIFY and REIDENTIFY modified to include a detection threshold
	parameter for feature centering.
    2.  The help pages were updated.

====================================
Version 2.3 Release, August 18, 1986
====================================

onedspec$splot/wrspect.hlp:  Valdes, August 14, 1986
    1.  The test for whether a new image will overwrite an existing image
	used ACCESS which is for nonimage files only.  This caused a problem
	with recognizing the automatic image extensions.  The modification
	uses IMMAP and IFERR to check if the new image would overwrite an
	existing image.

onedspec$doc/setdisp.hlp:  Valdes, August 8, 1986
    1.  The wording defining the meaning of "dispaxis" was changed because
	of user confusion.

onedspec$identify/idmark.x: Valdes, August 8, 1986
    1.  The optional labels have been adjusted to be half size and
	to have a path of up.  Note that on a vt640 the default text
	quality uses hardware generation so this change will not be
	visible unless you reset the text quality to high.
    2.  The size of the ticks and the gaps have changed slightly.

onedspec$t_dispcor.x: Davis, July 28, 1986
    1. DISPCOR was failing with a bus error on class2. It turned out that
        the get_feature1 routine was trying to get the flex_par parameter out
        of the image header after the image had been closed. I moved the
        imunmap call to the end of the routine.

onedspec$t_dispcor.x: Valdes, July 7, 1986
    1.  DISPCOR was opening comparison images when collecting dispersion
	solutions from the database and failing to close them.  In one
	particular large usage 509 images were opened before
	an out of memory failure!

onedspec$splot: Valdes, July 7, 1986
    1.  In SPLOT the 'w' key has been redefined to 'i' (create a new image).
	Key 'w' now windows the graph.
    2.  The help page and menus updated.

onedspec$identify/: Valdes, July 7, 1986
    1.  Redefined the 'r' key to be 't' so that 'r' can be the standard
	redraw key.
    2.  Help page and '?' menu updated.

onedspec$doc/standard.hlp, lcalib.hlp, sinterp.hlp:  Valdes, July 7, 1986
    1.  Help pages updated to reflect name changes in the standard
	calibration files.

onedspec$identify/: Valdes, July 3, 1986
    1.  Modified package to use new ICFIT package.
    2.  Changed coordinate list parameter to onedstds$henear.dat.
    3.  Updated help page for IDENTIFY to refect new default coordlist.

onedspec$identify/identify.x,reidentify.x,idgetim.x:  Valdes, July 1, 1986
    1.  Replaced calls to imtgetim with idgetim.  Idgetim calls
	imtgetim to get next image name but it then removes any
	image extension.  This is necessary to prevent having two
	different names by which an image may be identified in the
	database.

=====================================
STScI Pre-release and SUN 2.3 Release
=====================================
	
ondespec$getnimage.x:  Valdes, June 19, 1986
    1.  Changed BOOLS in common to INTS for safety's sake.

onedspec$(t_sensfunc.x,fudge.x,userstd.x): Valdes, June 19, 1986
    1.  SENSFUNC was not correctly accumulating grey constant corrections
	between different apertures.  This was fixed by rewriting the
	RLSQ procedures (moved into a file of their own, rlsq.x) and
	making appropriate changes in the rest of the code.
    2.  The grey constant was being computed incorrectly.

onedspec$t_flatfit.x: Valdes, June 18, 1986
    1.  FLATFIT aborted when an error is made specifying a nonexistant
	image.  It now prints an error message and goes on to the
	next spectrum.

onedspec$t_coefs.x: Valdes, June 16, 1986
    1.  Task was calling the wrong database package.  This produced
	totally wrong code since one package returns a structure
	pointer and the other returns FIO channel number.
	This error was probably introduced in May.

onedspec$t_standard.x: Valdes, June 12, 1986
    1.  Minor bug in STANDARD introduced when fixing problem with
	wavelengths (May 19).  Title was no longer being written to
	the STD file.

onedspec$t_dispcor.x: Valdes, June 12, 1986
    1.  DISPCOR had a fixed limit of 100 comparison spectra for all
	apertures in the database.  If this limit was exceeded memory
	would be corrupted (i.e. no check for exceeding the end of the
	array).  This has been changed to use dynamic memory allocation
	so that there is no limit on the number of comparison spectra.

onedspec$identify/ididentify.x: Valdes, June 11, 1986
    1.  Windowing key 'w' added.
    2.  Help page updated to reflect the 'w' and 'y' keys.

onedspec$splot.x: Valdes, June 10, 1986
    1.  Now sets dispersion correction flag when the user defines
	a wavelength scale and writes a new image.

onedspec$identify/splot.x: Valdes, June 9, 1986
    1.  Added check on the validity of the imio pointer when attempting
	to unmap the image.  This occured with a next image failed to
	access the specified image.

onedspec$identify/ididentify.x: Valdes, June 9, 1986
    1.  Changed Memr to Memd in 'y' option.

onedspec$identify/reidentify.x: Valdes, June 2, 1986
    1.  Changed from file template to image template.

onedspec$t_sensfunc.x: Valdes, June 2, 1986
    1.  Added check for square root of zero which is a fatal error on VMS.

onedspec$t_standard.x,t_sensfunc.x: Valdes, May 19, 1986
    1.  The output of STANDARD gave the wavelengths of the left edge of the
	first pixel and the right edge of the last pixel instead of the
	centers.  This causes slight errors downstream in SENSFUNC.
	This has been changed to give the actual W0 and WEND.
        I tried to check that all wavelengths were being calculated and
	used correctly.
    2.  SENSFUNC was not correctly using the output of STANDARD.  In some
	cases it assumed the starting and ending wavelengths were at
	the edges of the pixel and in other cases it assumed they were
	at the centers of the pixels.  The errors largely canceled out
	except that the W0 in the header for the SENSITIVITY image was
	wrong but WPC and the number of points was correct.  Again, I tried
	to check that everything is now consistent.
    3.  SENSFUNC was extrapolating observations when forming the composite
	sensitivity curve.  This leads to significant errors when some
	observations do not extend as far as others in wavelength.  This
	was noticed as a large increase in the RMS relative to the original
	RMS based only on the observations.  Now extrapolations are not
	allowed and only observations covering a given range of wavelengths
	are used in forming the composite curve.  Note that interpolations
	are still used if an observation does not contain a point at a
	particular wavelength.
    4.  The help page for SENSFUNC was modified to explain the difference
	between the RMS of the input points and the RMS of the composite
	points.
    5.  INTRP.F had to be modified because it considered a wavelength
	equal to the first wavelength in the table as an extrapolation.

onedspec$bswitch.par: Valdes, May 19, 1986
    1.  The BSWITCH parameter "add_const" has been changed to use the value
	from SENSFUNC of the same name.  The help page was also modified

onedspec$t_sensfunc.x: Valdes, May 16, 1986
    1.  SENSFUNC was not writing a complete header needed by LONGSLIT.
	Now it goes through the standard ONEDSPEC header package to create
	the senstivity images.

onedspec$t_bswitch.x: Valdes, May 14, 1986
    1.  BSWITCH was not reinitializing properly when not using IDSMODE.
	The effect was to give extraneous output.
    2.  All occurances of "== INDEFI" where changed to use the "IS_INDEF"
	macro.

onedspec$t_rebin.x: Valdes, May 14, 1986
    1.  If the image has not been dispersion corrected then an error is
	printed and the next image is processed.

onedspec$bplot.cl: Valdes, May 13, 1986
    1.  BPLOT has been modified to call SPLOT separately for each input
	image.  This has the effect of repeating the cursor file for each
	image.

onedspec$t_coefs.x: Valdes, May 12, 1986
    1.  COEFS was not writing a correct IDENTIFY database entry.

onedspec$t_rebin.x: Valdes, May 10, 1986
    1.  Rebinning into logarithmic intervals was not working.  This has
	been fixed.  A number of logical changes were required.
    2.  Rather than use an interative method for determining the coordinate
	transformation the transformation can be determined explicitly since
	both the input and output coordinates are linear.
    3.  The logarithm flag was previously ignored if a primary spectrum was
	used.  This prevented making the input and primary spectrum
	the same and then specifying either log or linear output.  This
	is a common way to use this task for converting to log intervals.
    4.  The primary spectrum was not being unmapped.

onedutil$bplot.cl: Valdes, May 9, 1986
    1.  BPLOT has been modified to use the new SPLOT.  The script is now
	a simple one line call to splot.
    2.  The input is now a image list instead of a file containing
	image names.  Note that to use a file containing image names
	the syntax is now "@file".
    3.  The cursor input file is now a parameter of the task allowing
	users to define their own set of commands.
    4.  The graphics device parameter is now standardized with other
	graphics tasks.
    5.  A modified help page is available.

onedspec$splot.x: Valdes, May 9, 1986
    1.  SPLOT now accepts a list of input spectra and processes them
	sequentially.  The parameter name has been changed from "image"
	to "images".
    2.  New SPLOT parameters XMIN, XMAX, YMIN, YMAX allow the user to
	set the limits of the initial plot.  These values may be modified
	interactively with :/xwindow and :/ywindow.
    3.  A modified help page is available.

onedspec$identify/reidentify.x: Valdes, May 8, 1986
    1.  Set log output to be flushed with every line written instead of
	being buffered.

onedspec$sflip.x:  Valdes, May 8, 1986
    1.  A new task has been added to the ONEDUTIL package call SFLIP.
	It flips the dispersion direction of spectra while maintaining
	the proper dispersion image header parameters.
    2.  A help page has been added for the task SFLIP.

onedspec$splot:  Valdes, May 7, 1986
    1.  Changed interpretation of W0 in logarithmic binning to be the
	logarithm of the wavelength of the first pixel.

onedspec$t_dispcor.x, t_rebin.x, t_combine.x:  Valdes, May 7, 1986
    1.  Changed meaning of w0 in logarithmic coordinates to be consistent
	with usual linear formula.  That is with a logarithmic wavelength
	interval the zero point is the logarithm of the starting wavelength.
    2.  Assumed increasing wavelengths in both the output spectra
	and the input spectra.  This restriction has been lifted.
    3.  Default output bins are in increasing wavelength with increasing pixel
	coordinate even when the input dispersion relation has the opposite
	sense.
    4.  The logic in REBIN for col_out = 0 was modified appropriately.
    5.  The help page for DISPCOR has been modified to indicate the new
	ability to have arbitrary input and output dispersion directions.

onedspec$userstd:  Valdes, May 6, 1986
    1.  Previously no graph of the errors would be made if the residuals
	were all the same.
    2.  Warning message was removed.
    3.  Boxes now drawn in NDC with standard size and do not depend on the
	range of the data or the size of the graph.

onedspec$userstdc.x: Valdes, May 6, 1986
    1.  Code incorrectly limited highest order for fit to one less than the
	number of points.  The order is now limited to the number of points.
    2.  Previously no graph of the errors would be made if the residuals
	were all the same.
    3.  Warning messages were removed.
    4.  Boxes now drawn in NDC with standard size and do not depend on the
	range of the data or the size of the graph.

onedspec$identify/idlog.x: Valdes, May 1, 1986
    1.  Column headings were adjusted.

onedspec$onedspec.cl:  Valdes, May 1, 1986
    1.  Removed loading of list and plot packages in ONEDSPEC package script.
	These packages are loaded with the NOAO package.

onedspec:  Valdes, April 27, 1986
    1.  Package pathname "noao.onedspec.onedutil" added to help pages for
	ONEDUTIL package tasks.

onedspec:  Valdes, April 7, 1986
    1.  OBSERVATORY task from IMRED package loaded with ONEDSPEC.
    2.  Latitude parameter removed from the ONEDSPEC package parameters.
    3.  DISPCOR, STANDARD, and BSWITCH latitude parameters changed to
	reference OBSERVATORY parameters.
    4.  The help pages for these tasks were revised.

onedspec$t_flatfit.x:  Valdes, April 7, 1986
    1.  Fixed minor bug.

onedspec$t_sinterp.x:  Valdes, April 6, 1986
    1.  Fixed bug in SINTERP.  It was using CURFIT with a pointer argument
	for the weights instead of a real array.  CURFIT is used only if
	the interpolation mode is one of the CURFIT types.  Obviously
	this option was never tested.
    2.  Entry points removed for portability.
    3.  The interpolation wavelengths when generating a curve were strongly
	subject to accumulated roundoff error; x = x + dx.  This was modified
	to use the construct, x = x1 + (i - 1) * dx, which may still have
	a precision limitation but not an accumulated roundoff error.

onedspec:  Valdes, April 5, 1986
    1.  Found very bad error in numerous places.  The arguments to CLGCUR
	were too few and of those that were there one was of the wrong
	datatype!!!  This was not a problem on the VAXes but very
	bad and hard to find on the SUN.
    2.  Fixed SUN bugs in SENSFUNC due to the statement:
	    call amovks (1, Mems[flags], npts)
	Apparently numeric constants are integer sized which causes problems
	on the SUN which has high order bytes first.  Watch out for this
	construct!

onedspec$getnimage:  Valdes, April 4, 1986
    1.  The entry points in this procedure caused tasks to fail on the
	SUN.  USE OF ENTRY POINTS IS HAZARDOUS TO THE HEALTH OF PORTABLE
	PROGRAMS.  I shall have to see if there are any more entry points
	in ONEDSPEC.

onedspec$getairm:  Valdes, April 4, 1986
    1.  Fixed minor bug in determining HA from ST and RA.  ST was still
	assumed to be in seconds which is not the case any more.

onedspec:  Valdes, March 28, 1986
    1.  ADDSETS would fail if an image was missing.  I modified it
	to detect missing files and continue on.

onedspec:  Valdes, March 27, 1986
    1.  The header parameters CRPIXn, CRVALn, CDELTn have been added to
	the image headers.  They replace W0 and WPC though W0 and WPC
	are still recorded in the header (for now).
    2.  A new task, SETDISP, has been added to set the dispersion axis
	(must be 1 for ONEDSPEC), the dispersion type, and the dispersion
	unit.  These are currently only used for labeling in IDENTIFY
	and thus the task is optional for the moment.
    3.  SPLOT modified to label the wavelength axis using CTYPE1 and CUNIT1.

onedspec$splot/deblend.x: Valdes, March 27, 1986
    1.  Moved deblend.x and eqwidthcp.x to splot directory.
    2.  There was a typo(?) in deblend.x of SPLOT which converted
	sigma to FWHM as FWHM = 2.345 * sigma.  This has been corrected
	to FWHM = 2.355 * sigma.
    3.  The help page for SPLOT was updated.

onedspec$identify: Valdes, March 26, 1986
    1.  Fixed bug in IDENTIFY which failed to add new lines with the 'l'
	command when the initial wavelength axis was pixels.

onedspec$identify: Valdes, March 24, 1986
    1.  Fixed minor bug in REIDENTIFY.  It was calling IC_FREE instead
	of IC_FREED (the new double precision version) while the rest
	of the package was in double precision.

onedspec: Valdes, March 21-22, 1986
    1.  Continued changes in the ONEDSPEC header parameters.  All the
	internal ONEDSPEC header parameters are initialized.  Those not
	in the image header are initialized to INDEF if no other default
	makes sense.  Then when a new image is created only the parameters
	which are not INDEF are written to the new image header.  Hopefully
	there isn't a obscure use in the package that assumes the default
	value of a parameter is zero (this was the previous default default).
	A bug of this sort occurred in SPLOT which assumed that W0 and WPC
	are zero if the image has not been dispersion corrected.  This was
	changed.
    2.  SLIST now prints INDEF for the parameters which are indefinite.
    3.  UT and ST are now stored internally as real values like all the other
	time and angle parameters.  Previously the were stored as integer
	seconds.
    4.  UT, ST, RA, DEC, and HA are written to new images as sexigesimal
	strings instead of real values.  This is contrary to the FITS standard
	but this is the way its been done previously.
    5.  Comments for parameters which are updated by ONEDSPEC are deleted
	when a new image is created.  This is because the database interface
	does not allow comments and when entering a new value the comment
	could be partially overwritten resulting in a nonsensical FITS cards.
	Parameters which ONEDSPEC does not use are not touched.

onedspec$fortran/intrp.f: Valdes, March 20, 1986
    1.  Converted entry points into separate procedures.  Entry points,
	while legal FORTRAN, tend to cause problems except in the very
	best compilers.  The change was sparked by the failure of the
	SUN optimizer.  It is not 100% certain that this caused the
	failure but it works now.
	
onedspec: Valdes, March 19, 1986
    1.  All double precision variables have been change to single
	precision.  The double precision is an anachronism.  There were
	numerous type mismatches with calling procedures using double
	precision and the called procedure expecting single precision.
	These problems were only found recently on the SUN workstation
	which has a reversed order to the bytes.  On the VAX this error
	is not caught.
    2.  The header parameters are accessed through the image database
	interface rather than directly.  This cleans things up alot and
	will make the transition to a real database easier.
	It does, however, mean that comments and sexigesimal notation are
	no longer used.
    3.  Most tasks creating an output image now make a copy of the relevant
	input image header.  This allows header parameters which are not
	recognized by ONEDSPEC to be propagated to the new images.

onedspec$t_combine.x: Valdes, March 19, 1986
    1.  Rebinning did not work after fix to DISPCOR (Feb 14) because macro codes
	were wrong.

onedspec$t_rebin.x: Valdes, March 19, 1986
    1.  Did not work after fix to DISPCOR (Feb 14) because macro codes
	were wrong.

onedspec$identify: Valdes, March 14, 1986
    1.  Modified IDENTIFY to store the line list internally instead of
	scanning the line list file every time.

onedspec:  Valdes, March 14, 1986
    1.  Fixed a bug in LOAD_HDR.X which caused a roundoff error in the UT
	and ST values.  This was a problem when creating a new image since
	it inherited slightly different values than the original image.
    2.  A double precision airmass variable was being passed to GET_AIRM which
	expected a single precisions variable.  This bug became apparent
	on the SUN workstation.  Modified GET_AIRM to expect a double
	precision airmass variable.

onedspec:  Valdes, March 13, 1986
    1.  Modified IDENTIFY and REIDENTIFY to be double precision.  It uses
	the double precision ICFIT and CURFIT procedures.
    2.  The help pages for IDENTIFY and REIDENTIFY were updated for the
	changes since Release 2.2
    3.  Fixed bug in SLIST which printed W0, WPC, and AIRMASS incorrectly
	on the SUN workstation.  Pargr was used instead of pargd.  Also fixed
	possible problem with assigning INDEFR to a double variable.

onedspec:  Valdes, March 11, 1986
    1.  SENSFUNC was not putting the dispersion correction flag, DC-FLAG,
	in the header for the sensitivity image.  This causes LONGSLIT.FLUXCAL
	to fail.  This has been fixed.

onedspec:  Valdes, March 6, 1986
    1.  Added parameter to SPLOT to allow selection of the graphics output
	device.
    2.  Help page for SPLOT modified.
    3.  New parameter file for SPLOT.  Also installed in IMRED packages.

onedspec:  Valdes, Feb 27, 1986
    1.  IDENTIFY and REIDENTIFY have been modified to do shifts in user
	coordinates instead of pixel coordinates.  This applies to the 's'
	and 'x' keys in IDENTIFY and to REIDENTIFY.  The shift specified in
	REIDENTIFY is now in user coordinates.  Unless otherwise specified
	the shifts printed by these tasks are in user coordinates instead
	of pixels.
    2.  A new key has been added to IDENTIFY.  The key 'r' resets the
	current feature to the position of the cursor.  This replaces the
	need to mark the new position and then delete the old position.
    3.  The output of 's' and 'x' in IDENTIFY is slightly different.
===========
Release 2.2
===========
From Valdes Feb 28, 1986:

1.  Fixed bug in FLATDIV which printed the image title as garbage.  Also
the output record number is increment for each input spectrum regardless
of whether the input spectrum is found, has already been flatted, or
is flattened.
------
From Valdes Feb 24, 1986:

1.  Removed junk file identify/isdir.x.
------
From Valdes Feb 14, 1986:

1.  t_sensfunc.x, userstd.x, and fudge.x have been modified to allow
the grey scale correction to be determined interactively even when
points are deleted.

2.  Fixed bug in DISPCOR to allow interpolation between solutions.  This
did not work before.
------
From Valdes Feb 10, 1986:

1.  FLATDIV has been modified to do in-place flattening when the input
and output spectra are the same.
------
From Valdes Jan 24, 1986:

1.  In IDENTIFY the 'l' always does a fit first before identifying
additional lines.
------
From Valdes Jan 21, 1986:

1.  HELP pages updated.

2.  The log information written by REIDENTIFY has been made more compact
and a option to futher reduce this log information "verbose" has been added.
------
From Valdes Jan 17, 1986:

1.  Bugs fixed affecting SPLOT and DISPCOR.
------
From Valdes Jan 6, 1986:

1.  Problem with cursor key 'o' in SENSFUNC fixed.

2.  The 's' shift option in IDENTIFY has been modified.  It now prints
the initial shift, the mean pixel shift, and the mean fractional shift
in user units.  This can be conveniently used for determining velocity
shifts from a standard.
------
From Valdes Jan 2, 1986:

1.  If the HA field was missing from a field it was being initialized to
0. which is a valid HA value.  This has been changed to initialize to -100.
This value will force recomputation of the HA when determining the air mass.

2.  A bug in computing the air mass when the HA is not defined was found
and fixed.
------
From Valdes Dec 30, 1985:

1.  A bug in DISPCOR when using a reference image and the directory
structured database has been fixed.
------
From Valdes Dec 9, 1985:

1.  NORMCONTINUUM has been renamed to CONTINUUM and modified to have the
output type as a hidden parameter.

2.  The standard line lists have been put in the directory stdlines$.
------
From Valdes Nov 26, 1985:

1.  SPLOT modified to use gtools graphics options.  These options are
accessed with :/ commands; i.e. ":/xwindow x1 x2" sets the x display
window.

2.  SPLOT parameter "auto" replaced by parameter "options" which allows
several plotting options to be given.  The options are given as a list of
possibly abbreviated strings.  The two options currently defined are
"auto" and "zero".  Auto is the same as before; it replots the graph
after any command that changes the graph.  Zero makes the initial
default for the graph have zero as the minimum Y.
------
From Valdes Nov 15, 1985:

1.  Modified IDENTIFY, REIDENTIFY, and DISPCOR to use directory type database
structure.  Instead of a single massive database textfile separate
database text files are created for each image in the database directory.
------
From Valdes Oct 28, 1985:

1.  Increased the efficiency of widstape from 7 seconds per spectrum to
about 2 seconds per spectrum by using low level formating.
------
From Valdes Oct 23, 1985:

1.  Bug fix to allow zero entries in the calibration files.
------
From Valdes Oct 9, 1985:

1.  Cursor parameter added to the tasks flatfit, splot, and standard.

2.  Defined widstape from ONEDSPEC package in the DATAIO package.  The
source and executable, however, still reside in ONEDSPEC (x_onedutil.e).
Widstape and widsout should be combined and the source put in DATAIO
at some point.
------
From Valdes Oct 7, 1985:

1.  Parameter indirections removed.

2.  Tasks IRS and IIDS moved to the IMRED package.  ONEDSPEC need not
be loaded directly.  The usually method should be to load IMRED and then
the appropriate instrument package.
------
From Valdes Oct 4, 1985:

1.  Add script task normcontinuum to fit the continuum of spectra and
output a continuum normalized spectrum.  This script is based on
images.fit1d.
------
From Valdes October 1, 1985:

1.  The source code for identify and reidentify has been moved from the
longslit package to the onedspec package since these tasks are essentially
one dimensional.

------
From Valdes August 19, 1985:

1.  Makelib file created to maintain archive for the onedspec package.
The archive is libods.a.  Makefile modified to use the library.
This removes all the .o files making directory easier to list.

2.  An attempt to write to an existing image in splot requires the
user to confirm.  Overwriting an existing image now maintains the pixel
files correctly.

3.  New script task revisions pages the package revision file.
.endhelp