.help revisions Jun88 noao.twodspec.longslit
.nf
longslit$transform/mkpkg
    Added missing dependency.  (10/11/99, Valdes)

longslit$transform/t_transform.x
    The REFSPEC keywords are now deleted if present.  (9/7/99, Valdes)

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

longslit$transform/
longslit$transform/fcgetcoords.x
    Added an error check for there only being one line or column measured.
    (7/21/99, Valdes)

longslit$transform/igsfit/igsfit.x
    Added an error check for an error in the fitting.  (7/21/99, Valdes)

transform/t_transform.x
    Updated for new interpolation types.  (1/4/99, Valdes)

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

transform/fcgetcoords.x
    Add an errchk on immap.  Without this the task would give a segmentation
    violation if for some reason it could not open the image section given
    in the identify database.  For example if the image was not present.
    (11/20/98, Valdes)

longslit.cl
    aidpars was incorrectly defined to be aidpars.cl instead of aidpars.par.
    (11/18/97, Valdes)

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

response.x
    The previous change had a typo in line 264 where the index should be
    j and not i.  (7/10/97, Valdes)

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

response.x
doc/response.hlp
    Change the behavior of the task with respect to the threshold parameter
    to agree with the help page.  Previously it replaced values below
    the threshold by the threshold value in both the normalization and
    the data prior to dividing.  The result would not be a unit response
    unless both the data and normalization were below the threshold.
    The new behavior gives a unit response if either the normalization
    or data are below the threshold.  The help page was slightly
    modified to make the behavior even clearer.  (5/15/97, Valdes)

doc/response.help
    Fixed formating typo.  (5/15/97, Valdes)

reidentify.par
    Change default threshold value to 0. (4/22/97, Valdes)

doc/fluxcalib.hlp
    Fixed missing task name in revisions section.  (4/22/97, Valdes)

demos$mktest.cl
demos$mktestt.cl
    Made the ARTDATA package parameters explicit.  (4/15/97, Valdes)

transform/fitcoords.x
transform/fcfitcoords.x
transform/fcgetcoords.x
transform/mkpkg
    Added error checking for missing database, missing database file,
    no coordinates, all INDEF coordinates.  (2/21/96, Valdes)

doc/illumination.hlp
    Fixed a formating error (font change).  (10/15/96, Valdes)

transform/fcgetcoords.x
    A rotated WCS is ignored in the same way as IDENTIFY.
    (1/4/96, Valdes)

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

doc/response.hlp
doc/illumination.hlp
doc/extinction.hlp
doc/fluxcalib.hlp
    Added note that DISPAXIS refers to the original dispersion axis in
    transposed images.  (7/31/95, Valdes)

longslit.cl
longslit.men
    Added the new SFLIP task to the package.  (7/18/94, Valdes)

transform/t_transform.x
    The last interval of the inversion surface could be distorted by the
    limitation of the inversion coordinats to be within the input image.
    This limit was removed (with the out of bounds checking taking place
    later).  (9/19/93, Valdes)

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

transform/fcgetcoords.x
transform/t_transform.x
    Modified to allow transposed axes.  (5/14/93, Valdes)

getdaxis.x	+
response.x
illumination.x
extinction.x
fluxcalib.x
transform/t_transform.x
    Access to the dispersion axis is now through the routine get_daxis.  This
    routine checks for transposed images.  (5/14/93, Valdes)

longslit.men
longslit.par
longslit.cl
standard.par +
sensfunc.par +
calibrate.par +
identify.par -
reidentify.par
demos$test.cl
demos$xgtest.dat +
demos$gtest.dat -
demos$xtest.dat -
    1.  Added commonly used tasks from the ONEDSPEC package.
    2.  Added additional package paraemters required by the ONEDSPEC tasks.
    3.  Modified the test playback for the new package and XGTERM.
    4.  Removed playbacks for XTERM and GTERM.
    (2/12/93, Valdes)

transform/fcgetcoords.x
    If the combine option is used and the images do not all have the same
    fit axis then a segmentation error would occur because of a mistake
    in where the MWCS and IMIO pointers are closed.  This was fixed
    and a warning message added.  (12/7/92, Valdes)

transform/fcgetcoords.x
    Features with INDEF user values are now excluded.
    (11/11/92, Valdes)

transform/t_transform.x
    Added DCLOG1 keyword.  This goes along with the changes in DISPCOR
    to allow multiple dispersion corrections.  (10/19/92, Valdes)

fluxcalib.x
    Loosened the wavelength limit checks so that an warning is only given
    if the image wavelengths extend outside the calibration wavelengths
    by more than a half pixel.  (9/10/92, Valdes)

demos/* +
longslit.cl
longslit.men
    Added a demos task with a test playback.  (7/24/92, Valdes)

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

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

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

transform/t_transform.x
    It was possible to end up with too few lines for MSIFIT.  A minimum
    buffer size is now enforced.  (6/18/92, Valdes)

transform/t_transform.x
    Modified to use MWCS.  (5/20/92, Valdes)

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

longslit$fluxcalib.x
longslit$doc/fluxcalib.hlp
    The output pixel type is now of type real.  If the input image is
    to be modified the calibration is done on a temporary image and
    renamed to the input image upon completion rather than being done
    in place.  Previously, flux calibrating a type short image would
    produce an image of all zeros.  (3/19/92, Valdes)

longslit$longslit.par
    Added observatory to package parameters.
    (2/6/92, Valdes)

longslit$transform/fcgetcoords.x
    In V2.10 IDENTIFY/REIDENTIFY measure feature positions in physical
    coordinates while FITCOORDS and TRANSFORM require logical coordinates.
    Therefore, the IDENTIFY database coordinates are transformed to
    logical coordinates when they are read.  (12/20/91, Valdes)

longslit$transform/igsfit/igsfit.x
    Removed the print statement about fitting because this caused the graphics
    to be overplotted on the previous graph for some unknown reason.
    (12/12/91, Valdes)

longslit$doc/extinction.hlp
longslit$doc/fluxcalib.hlp
longslit$doc/illumination.hlp
longslit$doc/response.hlp
    Added discussion and example about the DISPAXIS keyword.  (12/6/91, Valdes)

longslit$t_transform.x
    Fixed datatype declaration error for array tmp.  This was a harmless
    error.  (11/21/91, Valdes)

longslit$longslit.par
longslit$response.x
longslit$illumination.x
longslit$fluxcalib.x
longslit$extinction.x
longslit$transform/t_transform.x
    1.  Added dispaxis parameter to package parameters.
    2.  Modified all routines to use package dispaxis if not found in image
	all also write it to header.  (8/28/91, Valdes)

longslit$transform/t_transform.x
    Removed W0 and WPC from output image.  (8/28/91, Valdes)

longslit$transform/igsfit/igssolve.x
    The case of a single trace along x handled by igs_solve3 was using the
    yorder instead of the xorder in one place.  (7/11/91, Valdes)

longslit$transform/t_transform.x
    The interative inversion was made more stable by using a fudge factor.
    This was needed to make the LONGSLIT test procedure work on HPUX.
    (9/17/90, Valdes)

longslit$identify.par
longslit$reidentify.par
    Updated parameter files for the new version.  (8/23/90, Valdes)

longslit$transform/t_transform.x
    Changed the computation of the output grid from a cumulative addition of
    the pixel increment to a direct calculation to avoid cumulative
    round off errors in high resolution data. (7/19/90, Valdes)

longslit$doc/lslit.ms +
    Added copy of the SPIE paper on the LONGSLIT package.  It is in MS TROFF
    format.  Postscript copies may be obtained from the FTP archive.
    (7/4/90, Valdes)

====
V2.9
====

longslit$transform/igsfit
longslit$transform/t_transform.x
longslit$fluxcalib.x
longslit$extinction.x
    Added use of CD keywords in addition to CDELT.  (3/8/90, Valdes)

longslit$transform/igsfit/igsfit.x
    1.  Changed incorrect usage of abscissa/ordinate.
    2.  Cleared prompts after input.
    (3/6/90, Valdes)

longslit$transform/fcgetcoords.x
    Fixed problem in which database files where opened within a loop but
    only closed once outside a loop.  (5/6/89, Valdes - reported by Schaller)

longslit$illumination.x
    1.  Added error checking to handle missing DISPAXIS keyword.
    2.  Changed to dynamically allocated strings.
    (2/28/89, Valdes)
 
longslit$ilsetbins.x
    1.  The "bins" string is now checked for null after stripping any
	leading whitespace with xt_stripwhite.
    2.  The ":bins" command with no argument will not clear the bins now.
    3.  An error message is printed if two many sky bins are defined
	using the cursor.
    (1/26/89, Valdes)
 
longslit$fluxcalib.x
    1.  Changed CRPIXn keyword and variable to type real.
    2.  Added the ONEDSPEC flag for flux calibration.
    (1/26/89, Valdes)
 
longslit$response.x
longslit$illumination.x
    Added header keywords CCDMEAN and MKILLUM for compatibility with CCDRED.
    (12/14/88 Valdes)

longslit$transform/t_transform.x
    Changed the computation of x1, x2 and y1, y2 to natural units if logx and
    logy were set to yes. These numbers were being erroneously computed in
    log units leading to an erroneous transformation if the user specified the
    coordinate limits with x1,nx,dx and y1,ny,dy. (10/26/88 Davis)

longslit$t_longslit.x
    Changed the units of w0 to be log (w0) if log=yes. (9/21/88 Davis)

longslit$ilsetbins.x
longslit$transform/igsfit/igsfit.x
noao$lib/scr/ilsetbins.key
noao$lib/scr/igsfit.key
    Added 'I' interrupt key.  (4/20/88 Valdes)

longslit$mkpkg
longslit$longslit.cl
longslit$x_longslit.x
longslit$transform/mkpkg
longslit$transform/igsfit/mkpkg
longslit$transform/x_transform.x -
longslit$transform/libpkg.a -
longslit$transform/fitcoords.par -> longslit$fitcoords.par
longslit$transform/transform.par -> longslit$transform.par
    Merged tranform executable with the longslit executable.  (4/7/88 Valdes)

longslit$transform/extinction.x
    Was incorrectly doing in place correction.  (3/24/88 Valdes)

longslit$ilsetbins.x
    Increased bin string from SZ_LINE to 2048 chars.  Some users have attempted
    to define a large number of bins which fails when the string limit is
    reached. (1/4/88 Valdes)

longslit$transform/fluxcalib.x
    Was incorrectly doing in place correction.  (11/5/87 Valdes)

longslit$transform/transform.x -
longslit$transform/trtransform.x -
longslit$transform/trgetsurface.x -
longslit$transform/trsftomsi.x -
longslit$transform/trsetoutput.x -
longslit$transform/t_transform.x +
longslit$doc/transform.hlp
    The task TRANSFORM in  the  LONGSLIT  package  is  used  to
    interpolate images  onto a user defined coordinate system given as
    surface functions U(X,Y)  and  V(X,Y)  where  (X,Y)  are  the
    untransformed  image  pixel coordinates  and  (U,V) are the user
    coordinates.  The surface functions are derived from a set of measured
    points  using  the  task  FITCOORDS.  With  Version  2.6  of  IRAF
    the  algorithm  used  to  invert  the user coordinate surfaces, U(X,Y)
    and V(X,Y) --> X(U,V) and Y(U,V),  has  been changed.   Previously,
    surfaces  function  of  comparable  order to the original surfaces were
    fit to a grid of points,  i.e.  (U(X,Y),  V(X,Y), X)  and (U(X,Y),
    V(X,Y), Y), with the same surface fitting routines used in FITCOORDS to
    obtain the input user coordinate surfaces.  This  method of  inversion
    worked  well in all cases in which reasonable distortions and
    dispersions were used.  It was selected because  it  was  relatively
    fast.   However,  it  cannot  be  proved  to  work  in all cases; in
    one instance in  which  an  invalid  surface  was  used  the
    inversion  was actually  much poorer than expected.  Therefore, a more
    direct iterative inversion algorithm is  now  used.   This  is
    guaranteed  to  give  the correct  inversion  to  within a set error
    (0.05 of a pixel in X and Y).  It is slightly slower than the previous
    algorithm but it  is  still  not as major a factor as the image
    interpolation itself.

    The  event  which  triggered  this  change was when a user
    misidentified some arc lines.  The dispersion function which was
    forced  to  fit  the misidentified   lines   required  curvatures  of
    a  couple  of  hundred angstroms over 100 pixels at a dispersion of
    10  angstroms  per  pixel.  It  was  possible  to do this to the user's
    satisifaction with a surface function of xorder=6 and yorder=7.
    TRANSFORM inverts  this  surface  by fitting  a  function  with the
    same orders (it uses a minimum of order 6 and the order of the  input
    surface  function).   The  transformed  arc image  was  then examined
    and found to have residual wavelength errors 5 times larger expected
    from the residuals  in  the  dispersion  solution.  With  such  a
    large  curvature  in  the  dispersion surface function it turned out
    that to  maintain  errors  at  the  same  level  the  fitting function
    required  orders of 12.  (To determine this required a special version
    of TRANSFORM  and  the  new  double  precision  surface  fitting
    routines).   When  the  lines  were  correctly identified the
    dispersion function had much lower curvatures and required lower orders
    in the  fit and  gave a good transformation of the arc image.  The
    conclusions drawn from this event are:

    1. An incorrect dispersion solution can appear  to  be  correct  if
    the misidentified lines are at the end and a high enough order is
    used.

    2.   This  requires  high  order  surface  functions  in  FITCOORDS
    and TRANSFORM.

    3.  The algorithm used in TRANSFORM  in  V2.5  and  earlier,  while
    not failing,  does  give  unexpectly  large  residuals in the
    linearized arc spectrum in this case.  A cautious user should transform
    arc images  and examine them.

    4.   In  the  future  a more direct inversion algorithm is guaranteed
    to give residuals in the transform consistent with  the  residuals  in
    the dispersion solution even when the dispersion function is not
    realistic.
    (9/14/87 Valdes)

longslit$transform/trgetsurface.x
longslit$transform/fcfitcoords.x
longslit$transform/fcdbio.x
longslit$transform/trsftomsi.x
longslit$transform/trsetoutput.x
longslit$transform/igsfit/igsfit.x
longslit$transform/igsfit/igscolon.x
longslit$transform/igsfit/igssolve.x
longslit$transform/igsfit/igsget.x
longslit$transform/igsfit/xgs.x +
    Modified routines using the GSURFIT routines to call an interface routine
    which allows calling the double precision versions of these procedures
    without changing the single precision data arrays (a double precision
    copy is made within the interface).  Thus, FITCOORDS and TRANSFORM now
    use double precision arithmetic when doing surface fitting and evaluating.
    This removes the problems experienced with high order surfaces.
    (8/14/87 Valdes)

longslit$transform/igsfit/igsfit.x
longslit$transform/igsfit/igsget.x
longslit$transform/igsfit/igscolon.x
longslit$doc/fitcoords.hlp
noao$lib/scr/igsfit.key
    Added a listing of the fitted surface values at the corners of the
    image.  This allows evaluating the fit. (8/8/87 Valdes)

longslit$transform/fitcoords.x
    Added check against using blanks in fitname prefix instead of null
    file. (7/3/87 Valdes)

====
V2.5
====

longslit$extinction.x
longslit$extinction.par
longslit$doc/extinction.hlp
    Valdes, May 26, 1987
    1.  EXTINCTION now uses the same extinction files used by the ONEDSPEC
	package.
    2.  The parameter name for the extinction file has been changed from
	"table" to "extinction" to be consistent with the ONEDSPEC parameter.
    3.  The help page was updated.

longslit$longslit.cl
longslit$identify.par +
longslit$reidentify.par +
    Valdes, April 16, 1986
    1.  Parameters for IDENTIFY and REIDENTIFY are now separate for the
	LONGSLIT package.

longslit$fluxcalib.x
    Valdes, March 16, 1987
    1.  A reference off the end of the sensitivity image due to an error
	in a do loop index was fixed.

longslit$transform/trtransform.x
    Valdes, February 26, 1987
    1.  Add a warning if the header parameter DISPAXIS is not found.  This
	affects whether coordinate information for ONEDSPEC is produced.

longslit$*.x
    Valdes, February 17, 1987
    1.  Required GIO changes.

longslit$transform/igsfit/igsdelete.x
longslit$transform/igsfit/igsundelete.x
    Valdes, October 16, 1986
    1.  Real line type specified in gseti call changed to integer.
	This caused a crash on AOS/IRAF.

longslit$doc/fluxcalib.hlp
    Valdes, October 8, 1986
    1.  Added a short paragraph discussing calibration of logarithmicly
	binned spectra.

longslit$response.x
longslit$response.par
longslit$doc/response.hlp
    Valdes, August 18, 1986
    1.  RESPONSE was modified to allow separately specifying the image
	section to be used to determine the response (the numerator)
	and the image section used to derive the normalization spectrum
	(the denominator).  The help page was also modified.

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

longslit$doc:  Valdes, July 9, 1986
    1.  Help page and menu file (noao$lib/scr/ilsetbins.key) for ILLUMINATION
	were updated since they mention colon commands which do not exist.
    2.  Help page for EXTINCTION updated to reflect new name for extinction
	file.
    3.  Date of help page for FITCOORDS updated to because of new window
	command.

longslit$fitcoords.x:  Valdes, July 7, 1986
    1.  Keys 'a' and 'e' replaced with the general 'w' window package.
    2.  Help page updated.

longslit$response.x, illumination.x:  Valdes, July 3, 1986
    1.  RESPONSE and ILLUMINATION modified to use new ICFIT package.

transform/fitcoords.x,fcgetcoords.x,fcgetim.x: Valdes, July 1, 1986
    1.  Added routine to remove image extensions.  This was necessary
	to prevent having two legal image names and to avoid creating
	database files with the image extensions.

=====================================
STScI Pre-release and SUN 2.3 Release
=====================================

longslit$illumination.x:  Valdes, June 17, 1986:
    1.  It was possible to request a higher order image interpolator
	than the number of bins being interpolated causing an error.
	A check was added to use a lower order interpolator if the
	number of bins is too small.

longslit$*.ext; Valdes June 2, 1986
    1.  Moved the extinction data files to "noao$lib/onedstds/".
	Modified the parameter file for EXTINCTION appropriately.

longslit$fluxcalib.x:  Valdes, May 13, 1986
    1.  Modified FLUXCALIB to allow any combination of log or linear wavelength
	coordinates for the input image and the sensitivity image.

longslit$fluxcalib.x:  Valdes, May 1, 1986
    1.  Modified FLUXCALIB to use image templates instead of file templates.

longslit$tranform/transform.par: Valdes, May 1, 1986
    1.  Changed default value of parameter database to "database" from
	"identify.db"
    2.  Changed help page to reflect change in default parameter.

longslit$tranform/transform.x: Valdes, April 21, 1986
    1.  Task TRANSFORM crashed when flux conservation was turned off.  This
	was caused at the end by attempting to free memory allocated for
	flux conservation.  The transformed image is still ok.  This
	bug has been fixed.
    2.  Help page for TRANSFORM updated to include timing information.

longslit$ilsetbins.x: Valdes, April 7, 1986
    1.  Fixed use of STRIDX with a character constant to STRIDXS.

longslit: Valdes, Mar 24, 1986
    1.  RESPONSE, ILLUMINATION, EXTINCTION, and FLUXCALIB modified to
	fix history writing bug.

longslit: Valdes, Mar 21, 1986
    1.  APDEFINE, APEXTRACT, and SETIMHDR removed from this package.
    2.  APDEFINE, APEXTRACT, and SETIMHDR help pages removed.
    3.  LONGSLIT menu revised.

longslit$response.x: Valdes, Mar 20, 1986
    1.  There was a bug in RESPONSE which turned the interactive fitting
	off if the answer was only "no" instead of "NO".  This has been
	fixed.
    
longslit$illumination.x:  Valdes, Mar 11, 1986
    1.  ILLUMINATION has a new parameter for the interpolation type.
    2.  The help page for ILLUMINATION has been updated

===========
Release 2.2
===========
From Valdes Feb 11, 1986:

1.  APEXTRACT sets the BEAM_NUM beam number to zero for all extractions.
The aperture numbers are used to generate the record extensions.
------
From Valdes Feb 7, 1986:

1.  Images package loaded with longslit.
------
From Valdes Feb 3, 1986:

1.  Fixed bug in setting the aperture number in APDEFINE.  It was interpreting
the input value as a real number and storing it in an integer variable.
------
From Valdes Jan 23, 1986:

1.  Buffering limits removed in TRANSFORM.

2.  Bug fixed in coordinate setting in TRANSFORM.

3.  Bug fixed in undeleting points in FITCOORDS.
------
From Valdes Jan 3, 1986:

1.  FITCOORDS has been modified.  The 'z' zoom option now queries for
the type of zoom.  The types are feature, constant x, constant y, and
constant z.  This allows examining dispersion solutions at different
columns or lines.
------
From Valdes Nov 20, 1985:

1.  TRANSFORM now exits with an error if a database record is not found
rather than giving a warning and continuing on.
------
From Valdes Nov 15, 1985:

1.  FITCOORDS and TRANSFORM modified to use directory/text databases
rather than single text databases.  This new database structure is what
is now created by IDENTIFY and REIDENTIFY.
------
From Valdes Nov 7, 1985:

1.  The task MKSCRIPT has been made a basic system task.  It is no longer
loaded in the LONGSLIT package but is always available.
------
From Valdes Nov 1, 1985:

1.  New task MKSCRIPT has been added.  It is loaded out of the IMRED.GENERIC
package.  See the help page for the task and the revisions for GENERIC.

2.  Task FITCOORDS has been modified in several ways:
	a.  The images in a list of images can be fit separately or
	    combined into a single fit based on the value of the parameter
	    COMBINE.
	b.  Points delete interactively are recorded in a deletion list
	    and may be used in subsequent fits.
	c.  The last interactive plot or a default non-interactive plot
	    is recorded in a plotfile (if specified).  The plots in the
	    plot file can be spool or examined after the fact.

See the new help for this task.
------
From Valdes Oct 22, 1985:

1.  New parameter "exposure" in FLUXCALIB.  This parameter specifies the
image header keyword corresponding to the exposure time to be used in
calibrating the images.

2.  FLUXCALIB and EXTINCTION have been changed to take a list of input
images and a list of output images.  The output images may be the same
as the input images.
------
From Valdes Oct 4, 1985:

1.  Response and illumination modified to include the parameters for
low and high rejection and rejection iteration.
------
From Valdes Oct 1, 1985:

1.  The package has been reorganized.  Task extract has been moved to
a new package twodspec.echelle.  The source code for identify and reidentify,
which are actually one dimensional tools, have been moved to the onedspec
package though they are still loaded with the twodspec package.

2.  New task fluxcalib flux calibrates long slit images using the flux
calibration file produced by onedspec.sensfunc.

3.  Illumination can now handle using a single illumination bin.

4.  Task revisions renamed to revs.  Note that this is a temporary task.
------
From Valdes September 25, 1985:

1.  New task setimages added.  This task sets parameters in the image headers
defining the dispersion axis and, optionally, strings for the coordinate
types and coordinate units.  This strings, if defined, are used in other
tasks for identifying and labeling graphs.

2.  Because the dispersion axis is now defined in the header the axis
parameter in tasks response and illumination have been removed.

3.  Task transform now adds coordinate information to the image headers.

4.  New task extinction corrects images for extinction.

------
From Valdes September 23, 1985:

1.  Reidentify has been significantly speeded up when tracing a 2D image
by eliminating most database accesses.
------
From Valdes August 6, 1985:

1.  A bug in the absorption feature centering was fixed.
2.  Numerous cosmetic changes in the graphics are being made.  These will
be documented later.
------
From Valdes August 1, 1985:

1.  The icfit package has been modified to allow resetting the x and
y fitting points with keys 'x' and 'y'.  This is useful in identify
to reset the user coordinates directly in the fitting package.

2.  The :features command in identify now takes an (optional) file name
directing the feature information to the specified file.  Without a
file the terminal is cleared and the information written to the terminal
with a pause at the end.  With a file name the information is appended to
the specified file.

3.  A couple of small bugs in the handling of INDEF user coordinates in
identify have been fixed.

4.  The default pixel range in the icfit package when called from identify
is now the full image range rather than the range of points to be fit.

5.  The image section in identify is now used with :image just as it is
used for images given as arguments to the task.  Explicit image sections
must be given, however, in database :read and :write because the optional
names to these commands need not be image names.
------
From Valdes July 30, 1985:

1.  The tasks lsmap, lstrans, and reidentify have been changed so that
the user may specify a list of log files instead of just one logfile.
Now it is possible to have log output be written to the terminal
as well as a disk file.  This is now the default.
------
From Valdes July 27, 1985:

1.  The default user coordinate when marking a feature in identify
is the pixel coordinate if there is no coordinate function.

2.  When entering a user coordinate in identify after a (m)ark or
(u)ser key the coordinate typed by the user is matched against the
line list and the line list value substituted if a match is found.
Thus, for wavelengths the user only needs to enter the wavelength to
the nearest Angstrom and the decimal part will be found from the 
coordinate list.

3.  Response and illumination have been modified to work along either
image axis.  A new parameter "axis" has been added to select the
axis.  For response the axis should be along the dispersion (default
is along the columns) and in illumination the axis is that slit position
axis (the default is along the lines).  These changes in conjunction
with the new flat1d, fit1d, and background make the orientation of the
longslit images arbitrary!

4.  The values in the default parameter files for response, illumination,
identify, reidentify, lsmap, and lstrans have been changed.  This will
cause user parameter files to be out of date.  Sorry about that.
------
From Valdes July 26, 1985:

1.  Background has been modified to use new fit1d task.  It now does
column backgrounds without transposes and allows image sections.
------
From Valdes July 23, 1985:

1.  Task lsrevisions has been renamed to revisions.  The intent is that
each package will have a revisions task.  Note that this means there may
be multiple tasks named revisions loaded at one time.  Typing revisions
alone will give the revisions for the current package.  To get the system
revisions type system.revisions.

2.  Background now does both line and column backgrounds.
______
July 18, 1985:

1.  Help page for extract is available.
2.  Help page for lsrevisions is available.
______
July 17, 1985:

1.  Extract has been modified to allow interactively setting the
extraction limits for each trace.  If this is not needed then answer
NO to the query.  Any changes made in lower and upper remain
in effect to subsequent traces.  The lower and upper limits are written
to the database.  Older database tracings are still useable as before.
______
July 16, 1985:

1. A new task, lsrevisions, has been added to record revisions to the
beta test version of the package.

2. A help page for identify is now available!

3. A default one dimensional image section is available in the tasks
identify, reidentify, and extract.  This allows use of two dimensional
images (without an image section) to be used without bothering with
the image section.  It is also a little more general than regular image
sections in that a special format in terms of lines or columns can be given.
The default section is the "middle line".

4. Extract has been changed to allow:

	a. Recording the traced curves.
	b. Using the traced curves from one image to extract from another image.
    
This is done by having three query parameters giving the name of the
image to be traced or which was previously traced, a list of input
images from which to extract, and a list of output rootnames
one for each input image.
.endhelp
