This section contains a roughly alphabetical reference list of commands, with a description of the the actions invoked. (Strict alphabetic ordering has been sacrificed in one or two places in order to group together the texts for closely related commands.) Each command has its associated parameters listed with it, in the order in which they must be supplied. Optional parameters (for which defaults are provided) are given in [brackets].
Reads commands from a command file. Any prompts for unspecified mandatory parameters are given at the terminal. Command files can include blank lines and ‘comment cards’; the latter must have an exclamation mark [!] or asterisk [*] as the first character. (Comments may not be flagged with a first-column “C”, because this could easily be the first character of a command.)
On successful completion (or on failure to execute a command) control returns to the terminal.
IMPORTANT: Command files may not contain references to other command files (nor to themselves), for fairly obvious reasons.
DIPSO first searches for a file of the given name in the current directory (or whatever directory is
given in the file specification). If it fails to find it, it then looks for a file $OWNERDIR/<filename>.
Thus you can keep a set of frequently used command files in the directory assigned the
environment variable OWNERDIR (this assignment would normally be carried out in your login
scripts).
The default file type ([.typ]) is ‘.cmd’.
Adds the contents of the ‘current’ Y array to the values in STACK entry ‘n’, leaving the result
in the ‘current’ array. Both data sets must have monotonic X arrays for sensible results
to emerge. To perform the arithmetic, the data in the ‘current’ arrays are mapped onto
the STACK X grid. The addition is only performed at X values where there are valid Y
values in both data sets. In conjunction with GRID, the AADD command can be used to remap
data.
Divides the Y values in STACK entry ‘n’ by the Y values in the ‘current’ arrays. In order to do this, the ‘current’ data are mapped onto the X grid of the STACK data (both X grids must be monotonically changing). The output data are left stored in the ‘current’ arrays.
If both data sets are recognized by DIPSO as having X units in velocity space (WORV not equal 1; see
the TOV command for details), then the output data are corrected by the ratio of the WORVs (i.e. by the
ratio of the central wavelengths) to give true flux ratio as a function of velocity. If only one data
set is recognized as having X units of velocity, the division is carried out and an error is
reported.
At each X point, puts the larger of the STACK (entry ‘n’) and ‘current array’ Y values into the current
array. To do this, the data in the current arrays are mapped onto the STACK X grid. The function
returns Y values only at those X values where valid data occur in both data sets. AMAX may be
useful, when used with AMIN, for displaying the envelope to a set of spectra of a given
object.
At each X point, puts the smaller of the STACK (entry ‘n’) and ‘current array’ Y values into the current
array. To do this the data in the current arrays are mapped onto the STACK X grid. The function
returns Y values only at those X values where valid data occur in both data sets. (See also
AMAX).
Multiply the Y values in STACK entry ‘n’ by the values in the ‘current’ arrays. The data in the ‘current’ arrays are mapped onto the X grid of the STACK data in order to carry out this operation. (The X grids are both required to be monotonic). Results are left in the ‘current’ arrays.
Subtracts the ‘current’ Y array from STACK entry ‘n’. In order to carry out this operation, the ‘current’ data are mapped onto the X grid of the STACK data. Subtraction is only carried out at X values where there are valid Y data in both STACK and current arrays, and both X arrays must be monotonic. Results are left stored in the ‘current’ arrays.
Swaps the ‘top’ (i.e. numerically largest) STACK entry with the contents of the ‘current’ arrays.
Checks the current defaults for the lines and columns to be read in using ALASRD.
Tells ALASRD to read X and Y data from the specified columns of a file. (A “column” is a string of
alphanumeric characters separated from other columns by spaces; of course, the columns
which ALASRD actually acquires must contain exclusively numeric values.) When DIPSO
begins, these are set to 1 and 2 by default; they are not reset to these values after execution of
ALASRD.
Tells ALASRD to read only lines line1 to line2 (inclusive) of an input file. If line2 is specified as zero, this
is interpreted to mean end-of-file.
Reads data from a formatted file. The simplest structure which ALASRD (and DIPSO) can read is one
pair of X, Y values per record; this is the default file structure expected by ALASRD. More
elaborate files can be read in through prior use of the ALASLINS and ALASCOLS commands
(q.v.).
Gaps in the data are assumed to be flagged by Y values of zero, unless a different “brkval” is specified on the command line.
The default file type ([.typ]) is ‘.DAT’.
Writes the contents of the ‘current’ arrays into a formatted file. The X data are output in column 1, and the Y data in column 2; gaps in the data are flagged with a Y value of zero, unless a different value for “brkval” is specified on the command line.
The default file type ([.typ]) is ‘.DAT’.
Changes the angle at which PWRITE strings and MARK symbols are plotted. Theta is measured in
degrees, anticlockwise from the horizontal, and is initially set to zero.
Reads in Kurucz model atmosphere fluxes, from the data base kept in the directory
with environment variable SPECDAT. The data are stored in the database in the
form of astrophysical fluxes; however, ATLASRD multiplies them up by a factor
to
produce ‘actual’ fluxes. The x unit is Angstroms, and the y unit erg/cm2/s/A.
If MODE=0 (the default value) solar abundance models are acquired; otherwise the ‘low metal
abundance’ models (1/30 solar) are read in. A summary of the available solar abundance models can
be obtained using ATLIST, and the models normalised to cursor-selectable X,Y values using
ATNORM.
Other files in the SPECDAT database, including extended atmosphere and Non-LTE models (see
$SPECDAT/info.lis for details), can be read in using SP2RD $SPECDAT/filename.typ. The KHMEXT,
MLTE and MNLT models can all be ATNORMed if they are first subjected to TENY, but the
normalisation constant will be meaningless.
If you cannot access the model atmospheres, it may be because you are not currently using NDF data
format (see command USENDF). Most of the data in SPECDAT is only available in NDF format. Of
course, your node may not have the SPECDAT database installed, in which case complain to your
node manager.
Lists the T(eff) and Log(g) values used to specify the Kurucz solar abundance model atmosphere
fluxes that are accessible to ATLASRD.
Normalises model atmosphere fluxes stored in the ‘current’ arrays to the cursor position. The angular diameter implied by the normalising constant is printed at the terminal. The normalised fluxes are left in the ‘current’ arrays, and are plotted if mode=1 (the default). No plot is produced if mode=0.
If the cursor Y value is negative, the plot is assumed to be of X v Log10(FLUX). In this case, the Y
values plotted (and left in the ‘current’ arrays) are logs of the normalised model atmosphere fluxes;
however, the unnormalised atmosphere data in the ‘current’ arrays MUST be linear in Y to start off
with (i.e. in the form resulting from an ATLASRD). ATNORM can also be used to normalise black-body
fluxes generated with BBODY.
Calculates black-body fluxes [pi times B(nu)] at the x values of the grid in the ‘current’ arrays,
overwriting the y values therein. (GRID can be used to initialise appropriate x values.) The fluxes can
be normalised to observed data using the ATNORM command. The unit of ‘temp’ is Kelvin, the ‘x’ values
are assumed to be in Angstroms, and B(nu) is in erg/cm-2/s/A.
Turns on the BEEP following NOBEEP (also tests your terminal’s beeper).
Bins the contents of the ‘current’ arrays, which are expected to be in monotonically increasing X order. X1 is the start wavelength of the input data, and DX the X range over which binning is to take place; thus the first X value in the output data set will normally be approximately:
If you choose a value for DX which is less than the typical separation of the input X points you will probably get an unsatisfactory result. The binned data, which are the unweighted averages of the input X and Y values in each bin, are left in the ‘current’ arrays.
Automatic clearing of the plotting surface between plots. This is the default option on starting up. The
inverse function is NB.
Allows you to draw a ‘continuum’ using the cursor; the input X values must be in increasing order
(CDRAW is terminated with an ‘X(N+1).LE.X(N)’ type test). The result is stored in the ‘current’ arrays
(so you should first PUSH your spectrum so as not to lose it); subsequent rectification of data can be
carried out using ADIV. (See also CREGD, CREGS and PF).
If MODE=0 (the default value), the data stored are just those input with the cursor, giving a ‘join-the-dots’ spectrum. If MODE=1, a spline fit to the data points is carried out, using the subroutine INTEP described by Hill (Publ DAO). A smooth curve (which is supposed to be like one you might draw by hand through all the cursor points) is calculated on the grid of ‘x’ points of the arrays in the top (i.e. numerically largest) stack entry.
It is possible to input data to this routine from a file, rather than at the terminal. The data are expected
to be in the form (NDF or native DIPSO) produced by the WRITE command, and MODE is assumed to
be 1 (MODE=0 would have the same effect as a simple READ, and is therefore redundant). Thus,
using this option makes CDRAW act essentially as a straightforward spline interpolation
routine.
Clears the text surface. (Simply a PRINT *, ’ ’ loop; it will not, therefore, work on a 4010-type device,
for which you should use ERASE).
Clears all breaks (i.e. gaps in the data) from the current arrays. (The end-of-data is preserved, internally, as a break point, however.)
If no parameters are supplied, a listing of all available commands is given. If a minus sign is given as the only parameter, a one line description of each command is also displayed. If the minus sign is followed by a word, then only those commands which contain the given word in their descriptions (case insensitive) are displayed. As an example:
would display all commands which contain the string “continuum” in their descriptions.
Alternatively, commands for display can be selected by their function. All commands are grouped into one or more classes which are identified by single lower-case letters as follows:
Each parameter must be a word made up from a selection of these class identifiers. Commands must belong to all the classes in at least one of these words to be displayed. Classes can be negated (i.e. explicitly excluded) by specifying an upper case identifier (i.e. "A" means “not in class "a"”). As an example:
would list all commands which are to do with graphics but which are not control commands, and also all commands which are to do with title manipulation.
Note, if a file called $LDIPSODIR/comand.lis exists then the contents of the file are displayed without
processing (all parameters are ignored). It is recommended that such files be re-formatted to use the
new system (see section8.4).
Performs various illegal instructions with the object of trying to crash the program! It can be used to
test the behaviour of the signal/condition handler (see command HANDLER). The following values of
‘n’ causes the corresponding condition:
‘Continuum REGion Display’: plots the regions selected for ‘continuum’ fitting using CREGS.
The continuum windows are indicated by horizontal bars, which are drawn a fraction ‘h’ of the distance from the bottom to the top of the plot (0h1, default 0.8).
‘Continuum REGion List’: lists the current continuum windows selected with CREGS.
‘Continuum REGion Select’: select ‘continuum’ regions. These regions can be input as a parameter list; if no values are provided, the cursor is activated for interactive selection of continuum windows. When using the cursor, the input ‘X’ values must be in increasing order.
The selected regions can be checked using CREGD. This command will normally be used in conjunction
with PF. (See also CDRAW).
Implements automatic rotation of colours (when used with appropriate hardware). Each plot
begins with the colour specified by the last call to CSET (or colour 1, if CSET hasn’t been
called).
To cancel, use NCROT.
Set colour for Ikon plotting; n = 1-12 (1 = white). (See also CROT).
‘Cursor X range’: define the X range for plotting using the cursor.
‘Cursor X & Y range’: define the X and Y ranges for plotting using the cursor.
‘Cursor Y range’: define the Y range for plotting using the cursor.
Deletes STACK entries. At least 1, and up to 50, stack entries may be deleted; after deletion, the
remaining STACK entries are renumbered in sequence. Ranges of entries can be specified using the “-”
operator; e.g. DEL 2 4-6 8 will result in the deletion of entries 2, 4, 5, 6 and 8. DEL 1-50 will clear the
stack.
You are recommended to develop the habit of typing SL immediately after DEL, to check what has
happened.
Opens a GKS workstation (plotting device). If the argument isn’t provided then a table of legitimate workstations is listed, followed by a prompt. A full list of available workstations is given in SUN/83.
Dereddens data stored in the ‘current’ arrays. Negative points are multiplied by -1, dereddened, and the dereddened values again multiplied by -1.
The default value of R (=A(V)/E(B-V)) is 3.1.
The default value of MODE is 0, which results in a ‘standard’ Galactic-type law being used (Seaton 1979 for the UV, Howarth 1983 for optical-IR); any non-zero value results in an LMC-type law being used (Howarth 1983).
Draws a poly line between two points on the current graph.
Controls echoing, at the terminal, of commands (and comments) read in from command files. If mode=-1 (the default) there is no echoing, but you get a “Command sequence completed” message when the macro file is closed; if mode=0 this message is suppressed. Mode=1 results in commands being echoed; mode=2 echoes comment lines (i.e. those beginning with “!” or “*”); and mode=3 echoes commands and comments. All non-zero modes give a “Command sequence completedi” on completion.
This is not a DIPSO command, but a suite of programs accessed for Emission Line Fitting.
The available DIPSO/ELF commands are:
The most important ‘core’ commands are:
which are sufficient to give you a fit to some data.
“ELF DELete Coefficients”. Deletes entries from the stack of fit coefficients. Remaining
coefficient stack entries are renumbered, so use of ELFCSL immediately following ELFDELC is
recommended.
“ELF NEW Coefficients”. Clears the current array of fit coefficients (N.B. Not the DIPSO current
arrays!). The same result can be obtained using CLEAR inside the ELFINP editor.
“ELF INPut”. Allows you to Specify starting values and constraints prior to optimising the line fits with ELFOPT. The “batch” parameter controls what happens if the ELFINP command is invoked from within a command file. If “batch” is zero (the default), the user is prompted for the strings specifying the ELF constraints, starting values, etc (you are alerted to this by a change of prompt). If “batch” is non-zero, the strings are read from the script file. In either case, the last string read should be “QELF”.
The ELF command language allows starting values and constraints to be entered prior to optimising line fits. The language is similar to FORTRAN in the logical construction of commands; the following operators are recognised:
The last four have their conventional arithmetic meanings.
Five variables are recognised:
These refer to the line centre position, line width (FWHM), peak intensity, profile type, and degree of background polynomial. (If no value is specified for the profile type, it is assumed to be a Gaussian; i.e. P=1). To refer to a particular line, the variable must be followed IMMEDIATELY by the appropriate index (i.e. C1, W5, I2 etc). Numerical constants can be input as integer or decimal numbers, but exponents are not accepted. Blanks are permitted, but not within variable names or numerical constants.
Used to specify a starting value for a variable quantity, e.g.:
Starting values for peak fluxes are not required.
Used to set a fixed quantity, or relationship between two quantities; for example,
The command W2=W1 constrains the width of line 2 to be the same as that of line 1; P1=6 defines the profile type — in this case, to the first ‘numerical’ profile; D=0 fixes the background polynomial to have degree zero. (Note that the ‘=’ operator is the only one allowed with the D and P variables.) If D is undefined no background polynomial is incorporated. (Note that least-squares polynomials can be fitted to data by defining no lines and an appropriate value of D. For reasons of numerical stability, the polynomial coefficients are computed using an x scale centred on the mean x value of the data set; you are informed of the value of this offset zero-point.) Other examples should be self-explanatory.
These may ONLY be used with the variables C and W.
These may only be used with the variable I.
The other commands available are:
There is no command for deleting the specifications for a given line from the complete coefficient specification. If you need to ‘remove’ a line, it is necessary to define (using the = operator) the line intensity to zero, and the line width and position to suitable arbitrary values.
“ELF OPTimisation”. Initiates optimisation of fit coefficients (in the sense of minimising the sum of the squares of the deviations of the fit from the spectrum data in the DIPSO ‘current’ arrays). In the case of an analytical profile fit, the data may have velocity or wavelength as the unit of the X axis. Provided WORV is set correctly, the fitted flux will be in sensible units (ie those of the Y axis). In the case of a ‘numerical’ profile, spectrum data must be in velocity units. Note that line width is not permitted as a free parameter when fitting ‘numerical’ profiles.
ELFOPT may not be fast on a busy machine if there are many free parameters and/or datum points.
Also, for complex fits, ELFOPT may not converge on the correct solution. You may therefore want to
monitor the progress of a fit on an iteration by iteration basis. To do this, you should specify the
optional parameter “prmpt" to have a value of 1 or greater; you will then be prompted after
the first (and, optionally, subsequent) iteration to give you the opportunity to exit ELFOPT
cleanly.
The output from a completed fit consists of the optimised parameters and their errors (calculated in
the linear approximation, from the error matrix). The results may be stored for later inspection, or
output, using ELFPUSHC.
“ELFPUSH Coefficients”. Pushes the coefficients of the current ELF fit onto the stack of fit coefficients.
These results may be inspected, printed out, or recovered to the status of current fit coefficients using
ELFCSL, ELFVUC, ELFWRC, and ELFPOPC.
“ELF POP Coefficients”. Pops a set of fit coefficients from the ELF fit coefficient stack into the ‘current’
coefficient arrays (where they can be modified with ELFINP).
“ELF RESTore Coefficient stack”. Restores an ELF fit coefficient stack previously saved using
ELFSAVEC. The SAVEd stack of numerical profiles is restored only if there are no numerical profiles on
the profile stack at the time of the restore. This is to avoid ambiguity in definition of profile indices.
The conditions and features of the restore are the same as those for the DIPSO stack restore
(‘RESTORE’), except that the default filename and type are [ELFSAVE.ESTK].
“ELF SAVE Coefficient stack”. Saves the entire fit coefficient stack as an unformatted file that can be
subsequently recovered using ELFRESTC. The stack of stored numerical profiles (if any) is also saved.
The default filename and type are [ELFSAVE.ESTK]. The data file created is not an NDF and
may therefore not be readable on operating systems other than the one on which it was
created.
“ELF Coefficient Stack List”. Gives a summary of entries in the fit coefficient stack. Individual entries
can be inspected in more detail using ELFVUC.
“ELF View Coefficients”. Gives a listing, at the terminal, of entry ’n’ of the ELF fit coefficient stack. An
overview of the stack can be obtained using ELFCSL.
“ELF WRite Coefficients”. Writes the contents of the fit coefficient stack to a file (default type is .DAT). The information given includes the starting specifications for the fit, the results (if any), and the line fluxes in units corresponding to the data stored in the DIPSO stack.
“ELF Line FIX”. Allows use of the cursor to define a ‘fixed’ (i.e. non-optimisable) line in the data. Its principal use is to (in effect) take out features in the far wings of lines that are being optimised; these features might otherwise adversely affect the final fit.
Typing ELFLFIX brings up the cursor; two hits are then required. The first locates the centre and peak
flux of the feature. (These data are added to the ‘current’ fit coefficients in the same way as using ‘=’ in
ELFINP.) The second locates the half intensity point, on either side of the feature, to fix the half-width
at half maximum (from which the FWHM follows). (Hitting the same point twice leaves the width
undefined.)
“ELF Profile INput”. Transfers a numerical profile from DIPSO stack entry ‘n’ to storage in the profile stack. The data must have units of velocity along the X axis; spectra to which numerical profiles are fitted must also be in velocity space.
“ELF Profile List”. Lists the contents of the profile stack (by giving the title of the original DIPSO stack entry).
“ELFPUSH fit”. Pushes the result of an ELF fit onto the DIPSO stack, as a continuous (i.e. no breaks) spectrum. If no arguments are specified, the complete fit is pushed, into the next available DIPSO stack position. If ‘n1’ is specified, only the fit for line n1 is pushed. If ‘n2’ is also specified, the fits for lines n1 to n2 inclusive are pushed. If n1 is specified, the background polynomial (if any exists) is also pushed, before the line(s).
The string value being used for the specified environment variable is displayed. An error is reported if the no value is available.
DIPSO uses environment variables to specify various directories (eg DIPSODIR), and also the prompt string (DIPSOPROMPT). Default values for these environment variables can be supplied on the DIPSO command line in a comma separated list of “name=value” pairs. These default values will be used if the corresponding environment variables are not defined.
Erases plotting zone “n” (c.f TZONE). This is done rather slowly unless n=0 (the default).
Measures equivalent widths. The cursor is used to define two pairs of (X,Y)points,
between which the equivalent width is measured with respect to a linear ‘continuum’; if
you need a more complex continuum, the data can be preprocessed using CREGS together
with PF, or using CDRAW. To terminate the equivalent width measuring session, define
X2X1.
Note that EW expects to measure spectra with monotonically increasing X values.
Errors on the measured equivalent widths are calculated using the prescriptions given by Howarth
and Phillips (MNRAS 222, 809, 1986); these errors are likely to make most sense for interstellar line
measurements. The assumed nature of the errors is controlled by the value of MODE; the default for
MODE is 0. In this case, if the data have been processed with PF immediately before using EW, an
estimate of the continuum signal-to-noise is available. This results in automatic generation of
‘statistical’ errors on the equivalent width measurements, under the (conservative) assumption that
noise (rather than signal-to-noise) is constant. Such an assumption is likely to be reasonable for IUE
data. Note that on termination of the EW command the ‘statistical’ error is LOST. This is a
feature, not a bug, and is intended to stop you making mistakes through oversight. If, after
terminating EW, you find that you want to do further measurements on the same data, then the
sequence:
(where ‘m’ is the STACK entry into which the data are PUSHed) will normally recover the error
estimates. (The CREGS continuum regions are remembered, and PF 0 will fit a horizontal line to these
regions in the already rectified data, recomputing the error estimate. The Y coordinate of this line
should be 1.0).
If MODE is given a non-zero value, ‘statistical’ errors are calculated under the assumption of Poisson
statistics. Such an assumption may be considered by some (though not necessarily the author of this
document) to be appropriate to IPCS data. Since the uncertainties on individual points are not stored,
but are calculated (on a square-root basis) during the EW measurement, it is essential that unrectified
data are interrogated in this mode. It may, of course, prove convenient to overplot a ‘continuum’ to
assist interpretation.
In addition to the ‘statistical’ (i.e. random) errors calculated by EW, allowance for systematic errors in
setting the zero and continuum levels can be made, using the EWERR command.
WARNING: EW operates on the data stored in the ‘current’ arrays. Be careful not to make the mistake
of plotting STACK data, and then trying to measure equivalent widths off of the data on the screen.
The calculated value of EW is multiplied by the factor WORV.
Provides the program with estimates of the systematic errors in continuum and zero-level placement
(expressed as percentages of the continuum level). These errors are then incorporated into subsequent
error analysis when determining equivalent widths with EW.
These systematic errors are assumed to propagate quadratically (see MNRAS 222, 809, 1986).
Exits DIPSO, saving the stack in a file called EXIT.STK, or EXIT_STK.sdf.
Expands components of a plot by a factor “Factor" with respect to the default size on a given device.
(Factor is absolute, not relative; e.g. the sequence EXPAND 2, EXPAND 3 gives 3x enlargement, not
6x.)
If no value is supplied for clist, then the supplied expansion factor applies to all the components of
the plot listed below. If a value is supplied for clist, it must be a string containing some sub-set of the
characters A, I, N, T, M and P. Each character refers to a different component of the plot as
follows:
The expansion factor is only applied to those components of the plot for which the corresponding
letters are included in the clist parameter. The expansion factor for all other components is left
unchanged.
Results in filled-in MARK symbols (c.f NOFILL).
Measures fluxes with respect to (=above) a linear ‘continuum’ defined using pairs of cursor hits. More complex
continua must be rectified out, using CDRAW, or CREGS together with PF. To end a FLUX measuring session,
input X2X1.
Note that FLUX expects to measure spectra with monotonically increasing X values.
Fluxes are obtained by trapezoidal integration between X,Y points. Linear interpolation is used across ‘breaks’ in the data (and a warning given).
If a file name is supplied (the default file type is .DAT) then the x1, x2 and flux values are output to the
file for future reference. The file remains open until (i) you exit from the program; (ii) a new file name
is provided; or (iii) FLUX 0 is typed in. (The last option closes any file that is open [but does not open a
file called 0.DAT].) This means that, for example, FLUX can be exited, the X and Y ranges changed and
a new plot generated, then FLUX re-entered, and data will continue to be output to the previously
opened file.
WARNING: FLUX operates on the data stored in the ‘current’ arrays. Be careful not to make the
mistake of plotting STACK data, and then trying to measure fluxes off the data on the screen. The
integrated flux is multiplied by the factor WORV.
Selects font quality. Permitted values of ‘n’ are 0 (hardware characters), 1 (SGS characters), and 2
(NCAR characters). These are in increasing order of elegance, and execution time! The resulting fonts
are used for all text on the plot (i.e. apply to XLAB, YLAB, PWRITE, and TITLE when the title is
plotted).
FONT 0 (the default) is not always elegant, and because of problems with text rotation some devices
will write the Y axis label “upside down”. If this really worries you you can get round it with YLAB
(e.g. use YLAB xulF).
The elaborate FONT 2 style gives access to a wide range of special characters (Greek, italics,
mathematical symbols, etc.) These cannot be reproduced in this document, but are given in full in
SUN/90 (which documents the routines which DIPSO utilises). The following formats serve to
illustrate the possibilities:
Thus to give the formula for the area of a circle as a title:
Note that the single quotes surrounding the style definition strings are mandatory, as are the uppercase specifications. Note, too, that the correspondence between Greek and Roman characters is not usually as obvious as P=“pi”; it is important to check a printed (not line-printer!) copy of SUN/90 for details.
An angstrom symbol (Å) is provided as a special character: ‘.A’ (in ‘PRU’ style). Other complex
characters can be constructed (see SUN/90); for example, a mass-loss rate symbol (M surmounted by
a dot) is obtained using "’PRU’M’H:-85V105PRU’.’H:85V-105’".
Because the codes for special characters are flagged by apostrophes, life gets complicated if you want to include a normal apostrophe in a tring. But not too complicated: you just give it twice. For example, if you want a title that says:
you must specify:
Results in a formatted write of the contents of the ‘current’ arrays. The resulting data are in a form suitable for output on a line printer. The default file type ([.typ]) is ’.DAT’.
Defines a plotting area in absolute (device-independent) terms. The xsize and ysize parameters are in
cm; the locator index follows the numeric keypad, i.e. 1=bottom left, 9=top right etc. The locator index
defaults to 5 (centre of plotting zone). To return to (device-dependent) plotting zones use
TZONE.
Subzones within the specified FRAME can be set using FRZONE.
Defines subzones within a plotting area specified by FRAME. The parameters x1 etc. are in the range 0-1,
and define the fractional position of the subzone within the frame. An example of the use of FRAME and
FRZONE can be found in DIPSODIR:DEMO2.CMD ($DIPSODIR/demo2.cmd.
Filters high-frequency components (which you may identify with noise) from the real and
imaginary parts of a Fourier Transform (see FTRANS). FTFILTER expects the real part of the FT
to be in stack entry “n1", and the imaginary part to be in entry “n2" (which defaults to
n1+1).
FTFILTER operates by constructing an “exponent-squared edge function", and multiplying the FT by
that filter (see Bracewell, “The Fourier Transform and its Applications", for details). The filter has a
value of unity up to frequency f1, then drops like exp(D[nu]**2), where D[nu] is the change in
frequency. The e-folding scale is determined by specifying f2, the frequency at which the filter drops to
a value of 0.01.
The filtered real and imaginary parts are pushed onto the stack, together with the filter used.
The frequencies f1 and f2 will depend critically on your application; defaults are provided merely to provide an illustration for first-time users. These defaults are (Nu1+Nu2)/2 for f1, and MIN(Nu2,1.1*f1) for f2, where Nu1 and Nu2 are the first and last frequencies in the data set to be filtered.
Computes the inverse fourier transform from the real and imaginary parts of the FT, where the real part is in stack entry n1, and the imaginary part is in stack entry n2 (which defaults to n1+1). The result is pushed onto the stack (and WORV=1 assumed).
Computes the Fourier Transform of stack entry “n”, pushing the resulting real and imaginary parts onto the stack, together with the power spectrum. A fraction “p” of the data set is endmasked (at each end) with a cosine bell; “p” defaults to 0.05. For best results you should subtract a continuum (or at the very least the mean value) from the data, in order to maximise the effectiveness of the endmasking.
FTRANS does not use the Fast Fourier Transform algorithm. It does not, therefore, require
the number of datum points to be an integer power of two (at the price of being slow for
large data sets). However, it is necessary for the data to be uniformly sampled; if FTRANS
determines that your data set does not contain regularly sampled X values it will automatically
perform a 4-point Laguerre interpolation in order to simulate such data. For applications
where the data sampling rate is very irregular, or where there are substantial gaps (e.g. in
light-curves), it is recommended that PDGRAM (and PDGWINDOW) be used for Fourier analysis
applications.
The X values in any data set must be monotonically increasing.
It is recommended that you familiarise yourself with (e.g.) Brault & White, A&A 13, 169 (1971) before
using FTRANS.
Creates a grid of uniform wavelength points, starting at Lambda(1)=MIN(X1,X2). The Y values are set
to zero. This command can be used with AADD to remap data, or to set up a wavelength grid for (e.g.)
NEBCONT. Data are left in the ‘current’ arrays.
Controls the design of plots.
The DIPSO error handler makes the program uncrashable (in principle!). Setting level=0 results in the system error handler being invoked; fatal errors will kill the program. Any level greater than zero (such as the default value of 1) causes the program to issue a warning message when an error occurs (including control-C interupts). The current command is aborted and the user is returned to the main DIPSO command prompt.
To familiarise yourself with the condition handler you can type ‘CRASH’.
Calculates a theoretical (fully damped) profile for the Hydrogen Lyman-alpha line (1216 A), with a column density ‘N’. The profile is calculated at the X grid in the current arrays, with a continuum level of unity. If the parameter N is less than 30 it is assumed to represent log(N).
The most convenient way of using this function is to calculate a theoretical profile, use ADIV to divide
observed data (in the STACK) by it, then plot the resulting data. The ‘best’ value of the
interstellar H I column is that which gives the ‘flattest’ continuum around 1216. Of course,
interpretation is your responsibility, and the intrinsic stellar Lyman-alpha profile should
always be considered, together with the possibility that the interstellar line is not fully
damped.
If no value is supplied for ‘string’ then general help information for the DIPSO package is
displayed. More verbose assistance on a particular command can be obtained by typing ‘HELP
command
name’.
The information is displayed in one of two formats: plain text, or hypertext. The hypertext
information is taken from SUN/50 and is displayed using a World-Wide-Web browser. By default the
Mosaic browser is used, but this can be changed by assigning the required browser command to the
environment variable HTX_BROWSER before entering DIPSO. For example, to use Netscape issue the
following command before starting DIPSO:
If a browser of the requested variety has already been fired-up prior to starting DIPSO, it is “hi-jacked” to show the requested help information. Otherwise, a new browser is fired-up.
If this command fails, saying that the “showme” command cannot be found, try defining the showme
command explicitly using USEHTX.
In plain text mode, the information is taken from $DIPSODIR/dipso.hlp and $DIPSODIR/help.lis, and is
displayed in the DIPSO command window.
By default, the plain text format is used, but this can be changed using the USEHTX command.
Typing an interrogative (“?” or “?command
name”)
has the same effect as typing HELP.
Plots to be done histogram-style (as opposed to POLY or MARK).
Rotates between Hist and Poly plots on a given diagram, starting with whichever style is currently in
force (selected with HIST or POLY command); can be useful when comparing (e.g.) observations and
models. If mode is 1 (default is 0) then the first data set is plotted in the current style (i.e.
histogram if HIST is in force), and all subsequent data sets plotted on the given diagram will
appear in the alternative style. The cycle can be re-initialised at any time with HIST or
POLY.
Negated with NHPROT.
Estimates the area under the current arrays using simple trapezoidal integration (and will therefore only work successfully for data sets with monotonically increasing X values).
Applies Laguerre interpolation to data in the current arrays, to give a regularly sampled data set with the same number of points as the original. The parameter “n" controls the order of the interpolation.
INTERP does (n-1)th order interpolation, where ‘n’ must be an even number; thus only odd orders
greater than zero are handled. For example, n=2 gives 1st order (i.e. linear) interpolation; n=4 gives 3rd
order (i.e. quartic) interpolation. (INTERP cannot supply quadratic interpolation, since this would
require n=3 and odd values of ‘n’ are not allowed.)
The data must have monotonically increasing X values.
Note that spline interpolation can also be executed, using the CDRAW command.
Supplies atomic data for use with ISCALC and ISCOG. If a wavelength (in Angstroms) is provided, then
DIPSO searches for a file called ATOMIC.DAT in the current directory; if it fails, it looks for one
in a directory with the environment variable OWNERDIR; and if it again fails, it opens
$DIPSODIR/ATOMIC.DAT. Once a file has successfully been opened, DIPSO searches for a line whose
wavelength falls within 0.1A of the argument wavelength to ISATM. If it finds it, it uploads the
associated atomic data.
The file $DIPSODIR/ATOMIC.DAT uses data based on MNRAS 222, 809 (1986) and references therein, and
provides a model for the required format if you want to set up your own file. “Comment cards"
prefixed by “*” or “!” are permitted.
If a wavelength is not provided, or if ISATM fails to find a data set corresponding to a specified
wavelength, then a “data edit” mode is entered. This allows you to change or input atomic data, or to
carry out further searches through an ATOMIC.DAT file; typing H in this mode gives more
information.
(Incidentally, in case you’re wondering, ATOMIC.DAT requires statistical weights in order to permit
calculation of damping constants.)
Calculates a theoretical absorption profile for an interstellar cloud (or other plane-parallel slab of absorbing material with negligible forward scattering and a Gaussian line-of-sight velocity distribution for the absorbers), and pushes the result onto the stack.
Before this command can successfully be invoked, the atomic data must already have been loaded
(with ISATM), a cloud model defined (with ISINP), and a set of cloud ‘options’ defined (with
ISOPT).
Calculates a Curve-Of-Growth for a cloud model defined with ISINP, using atomic data
input via ISATM. The COG is pushed onto the stack, with X values (N.f.lambda) in units of
(cm-2.dimensionless.Angstroms) and Y values (W[lambda]/lambda) assuming equivalent width and
wavelength to be measured in the same unit.
Invokes a cloud “editor”, which permits an interstellar cloud model to be defined prior to calculating
theoretical profiles with ISCALC. Up to 18 clouds are permitted; each is specified by a velocity
dispersion parameter, “b” (see, e.g., MNRAS 222, 809 [1986] for a definition), a central velocity, V, and a
column density, N.
Typing HELP while in the cloud editor provides more information.
Loads a variety of options required before ISCALC can be successfully executed. These include
specifications for (optional) convolution with an instrumental resolution function, blending with other
lines, etc. ISOPT invokes an option “editor”; typing H in this edit mode gives more information.
Alternatively, if a filename is specified on the command line, an attempt is made to read options from
that file. Such a file should contain information matching the ISOPT editor input; for example, if the
file contains:
line profiles will be calculated over (at least) the range -100 km/s to +100 km/s; default values for all
other options will be assumed. (The default file name is ISOPT, and the default extension is
.DAT.)
Applies the ‘aging’ corrections for the IUE cameras described by Bohlin and Grillmair (Ap. J. Sup., 66,
209, 1988; SWP) and by Clavel et al. (ESA IUE Newsletter No. 26, p. 65, 1986; LWR). The
‘camera’ parameter is 2 for LWR and 3 for SWP; the ‘day’ parameter is day number in the
year. The ‘aperture’ parameter is relevant only to SWP data, and is 1 for trailed spectra,
2 for small aperture spectra, and 3 (the default) for large aperture point source spectra.
IUECOR works on data in the current arrays, replacing fluxes therein by their corrected
values.
Turns axis labelling back on (after use of NLAB).
If set ‘True’, the X axis will be plotted on a log10 scale.
If set ‘True’, the Y axis will be plotted on a log10 scale.
Replaces the X values in the current arrays by their base 10 logarithms.
Replaces Y values in the current arrays by their base 10 logarithms.
Alters the weight (i.e. ‘heaviness’) of all plotted lines, on devices which support this feature. The
weight must be 1-5 (initial setting is 1). See also TWEIGHT.
Plots to be done using symbols, as opposed to HIST or POLY. The symbols may be ‘designed’ using
MSET, and expanded and rotated using EXPAND and ANGLE.
Calculates the mean and standard deviation of the Y data stored in the ‘current’ arrays.
Merges STACK entries S1 and S2, with weights WT1 and WT2 respectively. This function is an
alternative to the AADD and YMULT commands, differing in the respect that if there is a gap in one
(but not both) data sets being manipulated, then the MERGEd data set will not have a
gap. The data sets in the STACK are both required to have monotonically increasing X
values.
If there is any region of overlap between the two data sets, then the Y values of the data set associated with the larger value of X(1) are mapped onto the X grid of the other data set. This remapping is done using a triangular filter, so that some slight smoothing can result if the sampling rates in the two data sets are very different. Remapping is carried out ONLY in any overlap region.
This function is intended for merging data sets with (normally) overlapping, but not necessarily
identical, X ranges. In particular, it can be used for merging IUE SWP and LWR flux calibrated spectra,
and UV and optical data (e.g. UBV fluxes; see UBVRD). Because remapping is done linearly, it is also
suitable for averaging similar spectra.
If MODE=zero (the default value) and the data look like IUE SWP and LWR spectra (on the grounds
of the X ranges), a wavelength-dependent weighting is used which reflects the inverse sensitivity
function of IUE. If you ARE dealing with IUE data, then inputting the exposure times as the WTs and
using MODE=0 will result in (essentially) a MERGE at the ’FN’ level, which is probably more correct
than one at the ‘absolute flux’ level. If MODE is positive, the additional IUE weighting is switched off.
(Note that if you are NOT working with IUE data MODE=0 and MODE=1 will give the same
result).
If MODE=-1, a straight addition of data in the overlap region is carried out. This option may be useful for (eg) combining emission line models with a zero background level. If MODE=-2 (or less), the two data sets are multiplied in the overlap region. This option may be useful for (e.g.) combining absorption line models with a background level of unity. If MODE is negative, ‘dummy’ values of Wt1 and Wt2 MUST be supplied.
You should only merge data which you expect to have similar Y values over any range of overlap. Merging dissimilar data sets (e.g. SWP and LWR IUE spectra which have not been flux calibrated) will give unsatisfactory results.
Writes a file which can subsequently be read into MONGO. It is possible to make MONGO leave gaps
in plots, corresponding to datum points with a notifiable ‘bad’ y value; the MONGOWR command inserts
a ‘bad’ point with y value ‘badval’ (default 0) into breaks in the data set to facilitate use of this
option.
Most of the more important MONGO functions are reproduced in DIPSO - you shouldn’t have to use this command!
Implements automatic rotation of plotting symbols. Each plot begins with the symbol type specified
by the last use of MARK (or symbol 1, if MARK hasn’t been called).
NOT IMPLEMENTED AT PRESENT.
Turned off with NMROT.
Selects the MARK plotting symbol. ‘NVERT’ is the number of vertices the symbol has; ‘STYLE’ takes on
values 1-4. Style=1 is a polygon, Style=2 a ‘star’ design, style=3 is an asterix, Style=4 gives an arrow
symbol, for plotting lower or (with ANGLE) upper limits.
‘No Box’: switches off automatic clearing off the plotting frame between plots. (Inverse function is
BOX). On a device or zone change, the ‘box’ is switched ‘on’ for the first plot, regardless of any NB
call.
Stops automatic rotation of the colour table when Ikon plotting and resets the colour index to 1
(white). (See CROT and CSET).
Calculates a theoretical nebular recombination continuum. The requisite data are comparatively numerous, and so are accessed from a file. The contents of the file must be as follows:
where:
A specimen file is given in $DIPSODIR/NEBCONT.DAT. If NEBCONT fails to find the file in the current
directory, it tries to look in a directory assigned the environment variable OWNERDIR (which you can
define in your .login script or your LOGIN.COM procedure).
All data in the file are reproduced at the terminal. By default, (mode1=mode2=mode3=0), new values are prompted for (simply hit ‘return’ to get the file values). By setting mode1=1, abundances are listed but not prompted for (fluxes, C, temperatures and densities still prompted). Mode2=1 does the same for the H(beta) flux and ‘C’; and mode3=1, the same for temperatures and density.
The results are calculated at the X co-ordinates of data in the ‘current’ arrays, and pushed onto the
stack. If you are not modelling real data, GRID can be used to set up an X array.
The default file type is .DAT.
Has the same effect as ECHO -1.
Selectively turns off axis labelling (modes not yet implemented; NLAB currently turns off X and Y axis
labels, and title).
Stops rotating plot style between Hist and Poly (after HPROT).
Stops automatic rotation of plotting symbols and sets the MARK index to 1. (See MROT and
MARK).
Stops the terminal beeping every time you make a boo-boo.
Results in MARK symbols being plotted as open (unfilled) figures; c.f. FILL.
Stops automatic rotation of line attributes and resets the TLINE index to 1 (continuous lines). (See TROT
and TLINE).
Stops rotating everything, resetting the colour, MARK and TLINE indexes to 1 (i.e. has the same effect as
NCROT,NMROT,NTROT).
Return to autoscaling of the X axis (after using XR, XMIN, XMAX, CXR and/or CXYR).
Return to autoscaling of the X & Y axes (after using XR, YR, XMIN, XMAX, YMIN, YMAX, CXR, CYR, and/or
CXYR).
Return to autoscaling of the Y axis (after using YR, CYR and/or CXYR).
Gives a BEEP (unless you’ve specified NOBEEP), and causes nothing to happen until you hit the return
key. (You might find a use for this command in command files.)
‘Poly fit’: fits a polynomial of degree ‘n’ through data points in the HIGHEST (i.e. numerically largest)
STACK entry, using continuum windows defined using CREGS. PF expects these data to be in order of
increasing X value. Y values, obtained from the parameters of the ‘best fit’, are calculated at the grid of
X points of the data in the highest STACK entry, between the first and last ‘continuum window’
X values. The resulting curve is left in the ‘current’ arrays (overwriting anything there
previously).
Because the ‘continuum window’ data are stored internally, it is possible to carry out a sequence of trial fits in an attempt to obtain a satisfactory polynomial representation of the continuum. Thus a typical sequence of commands to rectify data might be as follows:
(Then ADIV, TITLE, and so on).
Note that a slightly different approach to polynomial fitting is possible using ELFINP and ELFOPT
(which can give you the coefficients of a least-squares fit).
Locates the peak of the data set in the current arrays (notionally, but not necessarily, produced with
PDGRAM). PDGPEAK does this by locating the largest Y values, then fitting a parabola to the point and the
adjacent ones on either side. The peak and central values listed are those calculated from this
parabola.
Replaces the data in the current arrays (which should have monotonically increasing X values) with their “unevenly spaced data periodogram”, as defined by Scargle (ApJ 263, 835, 1982).
The parameter fl is the lowest frequency at which the periodogram is to be evaluated, and defaults to zero; fh, the highest frequency, defaults to 0.5*(n-1)/(x[n]-x[1]), where there are “n” points in the current arrays; df, the frequency interval, defaults to 1.0/(x[n]-x[1]); and p, the proportion of the data set endmasked at each end (using a cosine bell), defaults to 0.05.
N.B. For large data sets the default frequency parameters may lead to very large numbers of points in the periodogram, which will accordingly take a substantial time to evaluate.
The periodogram has X units which are the inverse of the original X units (ie will normally be in units of spatial or temporal frequency).
Replaces the data in the current arrays (which should have monotonically increasing X values) with
their window function, as defined by Scargle (ApJ 263, 835, 1982). The parameters fl, fh, and df have
the same meanings and default values as for the PDGRAM function, and the caution given for that
function regarding lengthy evaluation times applies also to PDGWINDOW.
One would normally calculate a data set’s window function in addition to its periodogram (PDGRAM) to ensure that any peaks in the latter are not an artefact of the sampling frequency.
Invert the Y axis on plotting (i.e. the ‘bottom’ Y value becomes the ‘top’ value, and vice versa).
Reverse the X axis (ie the X value at the left-hand edge of the plot becomes that at the right-hand edge, and vice versa).
Plot data (‘PM’ comes about ‘for historical reasons’). If you haven’t used the PPROMPT command, then
PM without any arguments will plot the data stored in the ‘current’ arrays; otherwise up to 50 STACK
entries can be plotted. If you have set PPROMPT to TRUE, then if you didn’t provide any arguments for
PM on the command line DIPSO will prompt you for some when it comes to do the plot. In order
to be able to plot both ‘current’ and ‘STACK’ data, the ‘current’ arrays are awarded the
‘honorary’ STACK entry number 0 for this command only. Ranges of stack entries may
be specified using the “-” operator; e.g PM 1 3-5 will result in entries 1, 3, 4 and 5 being
plotted.
Even if the BOX is switched ‘on’, all the data associated with a single PM command are plotted in a
single frame; i.e. a command sequence like:
will have a different result from:
WARNINGS:
may produce a different plot to:
Plotting to be done ‘join-the-dots’ style (as opposed to MARK or HIST).
Pop STACK entry ‘n’ into the ‘current’ arrays.
PPROMPT governs the default action of PM. The value of the ‘switch’ argument is T(rue) or F(alse), and
on startup is set to F. If you set it to T, then the PM command will prompt for an argument list if none is
provided on the command line. This could be useful if, for example, you want to include PM in
command files where you may not know in advance which stack entries you want to plot. If the
switch is set to F, then typing PM without any argument results in the data in the current arrays being
plotted.
‘Plot Spectrum’: this command, which acts on the contents of the ‘current’ arrays, is intended specifically for plotting long stretches of a single spectrum (e.g. high resolution IUE data) in sequential frames. The parameters are:
Plots are alternated in zones 5 and 6. After each frame is plotted, you are prompted as to whether or not you want to continue plotting.
WARNING: On completion of this command the X range and plotting zone in force will be those used for the last frame; you will probably want to change them.
‘Pushes’ the contents of the ‘current’ arrays onto the top of the STACK.
Writes a character string ‘string’ at co-ordinates x,y. If FONT 2 is active, the “locator index”,
‘i’, determines the location of the string with respect to the x,y coords: if i=1, the string
is written to the lower left (i.e. the top right corner of an imaginary box containing the
string ends at x,y); if i=2 the ’box’ is horizontally centred on, and vertically below, x,y; i=3
has the top left corner of the ’box’ at x,y; i=4,5,6 correspond to the ‘box’ vertically centred
and to the left, centre, right of x,y, respectively; i=7,8,9 are similar, but above x,y. These
locator indexes are chosen so that the numeric keypad available on most keyboards acts as a
mnemonic.
If the string contains commas, it must be enclosed in double quotes ("). It is recommended that you
develop the habit of using such quotes in any case.
The ANGLE, EXPAND, and FONT commands can be used to modify the appearance of the string on
output.
Reports the (relative) sizes of the grid and graph windows for zone “Zn”, together with the absolute
area of the plotting surface (if available). See the TPORT command for details.
Applies a ‘quick’ Gaussian smoothing (FWHM of filter = 2.354*Sigma) to Y values in the ‘current’
arrays. (Defines gaussian weights at a grid of delta(x) values, then linearly interpolates when
smoothing). Much faster, and scarcely less accurate, than SM.
Quit DIPSO, without saving the stack. Can be abbreviated to “Q”.
Reads the values from two specified catalogue columns in the the current arrays. The file containing the catalogue is given by “file”. This must be a full path name (no shell meta-characters such as ~, $, etc can be used here). The file can be a FITS table (binary or formatted), or an STL table (see SUN/190 ). The names of the columns holding the X and Y values are given by “xcol” and “ycol”. A full list of available column names is displayed if an unknown column name is supplied.
Reads a disk file which has been written with WRITE. This and RESTORE are the only input modes that
preserve ALL the information associated with a DIPSO data set, and are the recommended
options. The disk data can be in either NDF or “native DIPSO” format, depending on the flag
established by the USENDF command. When specifying an NDF, do not include any file
type.
If the specified NDF was not created by DIPSO, then an attempt will be made to create appropriate X values based on the WCS component in the NDF (or the FITS header cards in the FITS extension if there is no WCS component). The second and third command parameters, “sys” and “unit”, may be used to indicate the spectral system and units in use within the NDF (these are only used if the system and units are not clearly specified in the NDF). The system value supplied must be a standard FITS-WCS value such as “FREQ”, “WAVE”, “VRAD”, “VOPT”, etc. Likewise, the unit should be a standard FITS-WCS such as “m”, “nm”, “Angstrom”, “Hz”, “GHz”, “m/s”, “km/s”, etc. The spectral axis values within the NDF will be converted automatically from their original system to either wavelength in Angstroms or optical velocity in km/s for storage in the DIPSO X array. The original system and unit adopted for the NDF during this conversion are reported. If the user does not specify a system and/or unit, then a default of “wavelength in metres” will be adopted unless the contents of the NDF suggest some other default.
Recalls a previous command line or prompt response, putting it into the current keyboard input buffer
for acceptance (by pressing return) or further editing. The RECA statement can be issued
when-ever dipso prompts the user for input. If no parameters or qualifiers are given, the string
which the user last typed in is recalled. If a non-numeric parameter value is given, the most
recent string to be given by the user which starts with the supplied string parameter is
recalled. If a numeric parameter is given, the corresponding string is recalled where “1” is
the last (i.e. previous) string, “2” is the last-but-one, etc. If the /all qualifier is given (i.e.
“RECA/ALL” ), a list of the last 20 strings given by the user is displayed, and the user is then
re-prompted.
Lists a RECORDed string at the terminal (as a reminder!)
Records a string of commands for subsequent REPLAY. The string must be enclose in double quotes; it
can itself contain double quotes, but must then be delimited by pairs of double quotes. The string may
not incorporate RECORD or REPLAY commands, but the syntax, and general validity, of the string is
otherwise not checked until it is REPLAYed.
To cancel a current RECORDing, use a null string.
If REPLAY is included in a command line, then it is replaced in that command line by the contents of
any RECORDed command string.
This command can be used to reduce the number of error messages displayed if one of the NDF accessing commands fail. The command takes a single integer parameter. Level = 2 is the initial value and results in all error messages being delivered to the screen. Level = 1 replaces all errors deriving from a single failed command with a single (less detailed) error message. Level = 0 suppresses all error reports.
Restores STACK data previously dumped using SAVE. RESTORE does not require the current stack to be
empty; however, if it is not then retrieval may not be complete, depending on the available stack
space. (Notification is given of incomplete retrieval, of course). The data can be in either NDF or
“native DIPSO” format, depending on the flag established by the USENDF command. When specifying
an NDF, do not include any file type. The default NDF name is ‘SAVE_STK’. If not using NDFs, the
default file name is ‘SAVE.STK’.
Changes the title of stack entry “n”. The string (which may be null) must be enclosed in double quotes
if it contains commas. (The use of double quotes is in any case recommended.) Special characters may
be incorporated (see the FONT command).
Restrict the X range of data in the ‘current’ arrays (i.e. throw away data outside the range X1 to X2). Only works properly on data sets with monotonically increasing or decreasing X values.
‘Restrict Y range’; throws away datum points in the current arrays which have Y values outside the limits Y1 and Y2. These limits apply to the entire arrays unless X1 and X2 are supplied, in which case the operation is carried out only within that X range.
Dumps the contents of the STACK in a form suitable for subsequent reacquisition using
RESTORE. The data can be in either NDF or “native DIPSO” format, depending on the flag
established by the USENDF command. When specifying an NDF, do not include any file
type. The default NDF name is ‘SAVE_STK’. If not using NDFs, the default file name is
‘SAVE.STK’.
It is possible to specify subsets of the stack for saving, but then a file or NDF name is mandatory. The entire contents of the stack are saved by default.
Input data from the terminal to the current arrays; terminate with “\”. Each input line should contain a pair of X and Y values but nothing else. If a value of "brkval" is specified on the command line, then Y values matching it are assumed to flag gaps in the data.
When a VT emulant is being used, this command result in text being scrolled between lines n1 and n2
only. Thus you can scroll text to, say, the bottom half of the screen (SCROLLVT 14 25) while
plotting on the top half (TZONE 5). If specified, both n1 and n2 must be in the range 1 to 25,
an at least two lines must be allowed for. Full screen scrolling results if no line range is
given.
The supplied command is executed within a Bourne shell running in a child process which terminates
when the command is completed. Note, the command is assumed to have failed if the status value
returned by the command is non-zero. An alternative to using the SHELL command is to suspend the
process running DIPSO by pressing control-Z, issue the required command, and then re-awaken the
dipso process by typing “fg”. The SHELL command is provided for use in DIPSO command
files.
Stack list: gives the entry number, number of points, first and last X values, and the first characters of the title associated with each STACK entry. By default, the entire stack is listed (N1 = 1, N2 = no of stack entries), otherwise the first (N1) and last (N2) entries to be displayed can be specified. If N1=0, the contents of the ‘current’ arrays are also summarised. The “-" operator can be used to separate N1 and N2.
Stack List WRite. Operates identically to SL, but outputs results to a file STACK.LIS in the current
directory, instead of to the terminal.
Smooths Y data in the ‘current’ arrays with a Gaussian filter (FWHM=2.354*Sigma). QSM gives a much
faster, and scarcely less accurate, result.
Cuts out data from the ‘current’ arrays. Pairs of X values can be provided as optional parameters; if no
parameters are given, the cursor is used to define regions to be SNIPped. Each pair of X values, whether
input at the terminal or defined by cursor, must be in order of increasing X (though this constraint
does not apply to successive pairs of values). To get out of SNIP mode when using the cursor, define
X2X1.
WARNING: SNIPonly works successfully on data sets that have monotonically increasing X values.
Remember, ‘snipping’ is done on data held in the ‘current’ arrays. Be careful not to plot STACK data
and mistakenly think that it is the plotted data that are being edited. Furthermore, it is not possible
to recall snipped-out data points, so it is wise to maintain an untouched data set on the
STACK.
Reads a SPECTRUM format 0 file into the ‘current’ arrays. This is the format that data output from
IUEDR are normally in (see SUN/37 for details); for other purposes, READ (and WRITE) are
recommended. Maximum number of points allowed is 20,000. WORV is assumed to be
1.0.
The data can be in either NDF or “native DIPSO” format, depending on the flag established by the
USENDF command. When specifying an NDF, do not include any file type.
If YES (the default) is supplied for the second (optional) logical parameter any zeros in the data file are treated as gaps in the data. If NO is supplied, any zeros are treated as good data values. This replaces the undocumented PANICRD command.
Outputs a SPECTRUM format 0 file (see the IUEDR documentation SUN/37 for a description of the
SPECTRUM formats). But you should be using WRITE! The data output are those stored in the ‘current’
arrays.
The data can be in either NDF or “native DIPSO” format, depending on the flag established by the
USENDF command. When specifying an NDF, do not include any file type.
Reads a SPECTRUM format 1 file (see SUN/37 for details) into the ‘current’ arrays. Maximum number of points allowed is 20,000. WORV is assumed to be 1.0. The default file type ([.typ]) is ‘.DAT’. This command cannot read NDFs.
Reads a SPECTRUM format 2 file (see SUN/37 for details) into the ‘current’ arrays. Maximum number of points allowed is 20,000. WORV is assumed to be 1.0. The default file type ([.typ]) is ‘.DAT’. This command cannot read NDFs.
Replaces the X values in the current arrays by their square root (throwing away any negative values).
Replaces the Y values in the current arrays by their square root (throwing away any negative values).
Returns information on the current status (device number, X and Y ranges, etc.).
Adds a string to the end of the current title.
Replaces the X values in the current arrays by 10.0**X.
Replaces Y values in the current arrays by 10**Y
Controls the appearance of tickmarks on plots. The arguments dx and dy control the spacing between major (labelled) tickmarks on the x and y axes, respectively; nx and ny control the number of spaces between major ticks (i.e. no. of minor ticks = nx/y minus 1).
A zero for any argument results in the relevant aspect of the plot design returning to automatic
control; just typing TICKS (no arguments) returns to fully automatic tickmark design. Negative
arguments result in no tickmarking.
If you supply grossly inappropriate arguments (e.g. trying to squeeze too many numbered ticks onto an axis) then you will certainly find the resulting plot not to your satisfaction. Beware, in particular, of forgetting to scale by appropriate powers of ten (which may be subsumed into the axis label). Unless you are very familiar with the range of the data being plotted, it is wise to do initial plots with autodesign in force.
Change the title associated with data in the ‘current’ arrays. Null strings are accepted as such unless
TPROMPT has previously been set TRUE (q.v. TPROMPT). The string must NOT contain any control
characters (which would probably cause a crash in the graphics library routines). Since a comma is
normally interpreted as the end of a string, it is necessary to enclose strings containing commas in
double quotes; e.g.
gives:
and
respectively. It’s a good idea to develop the habit of using double quotes regularly, even if you don’t use commas often in strings.
The price paid for being able to include commas in strings is that other usage of quotes in TITLE is
forbidden. (Actually, it’s not, but the rules are so complex as to effectively forbid use of
quotes.)
If “logical” is T or Y, then the command TITLE will prompt for an input string if none is provided; if F
or N (the default), it won’t.
Change line attributes (continuous, dot-dash etc). The index ‘line’ must be in the range 1-5 (1 =
continuous lines). (See also TROT).
If the X and Y data in the current arrays are in units of Hz and
erg/cm2/s/Hz, TOFLAMBDA will convert to Angstroms and erg/cm2/s/A.
If the X and Y data in the current arrays are in units of Angstroms and erg/cm2/s/A, TOFNU will
convert to Hz and erg/cm2/s/Hz.
Convert X values from wavelength to velocity (the inverse of TOW), where Wav0 is the appropriate rest
wavelength in Angstroms.
No changes are made to the Y array. This would normally mean that subsequent measurements made
with EW or FLUX would be in rather strange units (e.g. km/s, or erg/cm2/s/[km/s]). To avoid this
anomaly, an internal variable WORV (for Wavelength OR Velocity) is associated with each data set.
This has the value 1.0 by default, but is reset to Wav0/c (where c is the speed of light in km/s) on
using TOV.
(If you are reading in data sets with velocity as the X co-ordinate it is usually safest to convert to wavelength [TOW] then back to velocity [TOV] in order to obtain an appropriate WORV value. This is unnecessary if you use READ/WRITE or SAVE/RESTORE to move data in and out of DIPSO.)
WARNING: If you have set the X range, you’ll probably need to change it to get anything on the
plotting surface after using TOV or TOW!
Convert X values from velocity to wavelength (see TOV), and resets WORV to 1.0.
Implements automatic rotation of line attributes. Each plot begins with the line style defined by the
last use of TLINE (or style 1, if TLINE hasn’t been called).
Switched off with NTROT.
Defines a plotting subzone (no. “Zn”, where
100Zn8).
The dimensions of the subzone (Xmin...Ymax) are normalised such that the total dimensions available
(regardless of device) are 0 to 1 in both axes. (The corresponding physical dimensions can be
discovered using QAREA.)
In general, a plot consists of axes, axis labels, and a header, as well as the data. Technically, the plotting subzone is a “graph window”; within this graph window a “grid window" is present. The grid window is exactly filled by the axes, and so will normally be smaller than the graph window (to leave room for the labelling). DIPSO will normally work out default grid window dimensions, but you can define your own (WXmin...WYmax). This might be useful if you want contiguous axes in different plots, for example.
Just as the graph window is defined in terms of fractions of the available plotting area, so the grid window is defined in terms of fractions of the available graph window. Thus parameters WXmin etc. must also be in the range 0 to 1 (regardless of the graph window dimensions).
To access a given subzone, use TZONE.
Removes leading blanks from the title associated with the ‘current’ arrays.
Copies the title from STACK entry ‘n’ to the ‘current’ arrays.
Alters the weight (i.e. ‘heaviness’) of data curves plotted with PM, on devices which support this
feature. No other lines (axes, etc) are effected. The weight must be 1-5 (initial setting is 1). See also
LWEIGHT.
Selects zone “zn” for plotting. The zone numbers are:
Additional zones can be user-defined if required (see TPORT). When using such additional zones
special care is needed to avoid overplotting previous zones (use ERASE to get a “page throw” on hard
copy devices); this is taken care of automatically with zones 0-8.
Converts UBV magnitudes to fluxes, and stores the results in the ‘current’ arrays. If the U, B and/or V
magnitude is unknown, 0 should be entered. (Should you want to actually input a value of 0, I’m
afraid that you’ll have to use something like 0.0001, or put in a value of (e.g.) 5 and then carry out
some arithmetic using TENY, YMULT and LOGY).
The data are plotted at assumed wavelengths of 3600, 4400 and 5500 Angstroms using lines 2*dx wide (default value of dx: 50 Angstroms). Conversion from magnitudes to fluxes is carried out using
where C is 20.94, 20.51 and 21.12 for U, B and V respectively. (These values are from the absolute flux measurements of Vega made by Tug et al, Oke & Schild, and Hayes & Latham, for V; and a normalised Kurucz Vega model for U and B).
If ‘logical’ is Y, YES, T or TRUE, the HELP command will display help information in hypertext format,
using a World-Wide-Web browser. If N, NO, F or FALSE is supplied for ‘logical’, help information will
be displayed in plain text format in the DIPSO command window. If no value is supplied for ‘logical’,
YES is assumed.
If supplied, ‘showme’ should contain a string giving the unix command used to run the Starlink
showme utility. The supplied command string will be used to initiate hypertext help sessions until a
new value is supplied. The command used initially is “<dir>/showme”, where <dir> is the path to the
directory containing Starlink executable files which was used when DIPSO was installed (typically
/star/bin).
If ‘logical’ is Y, YES, T or TRUE, data accessing commands such as READ, WRITE, SAVE, RESTORE, etc, will
use NDF structures (see SUN/33) to store data in. If N, NO, F or FALSE is supplied, they will use
“DIPSO native” files as used by DIPSO prior to version 3.0 (but note, that these files will NOT be
portable from one operating system to another). If no value is supplied for ‘logical’, YES is
assumed.
Reads data from the IUE ‘Uniform’ Low Dispersion Atlas, ULDA, which have been output using the USSP (see SUN/20). The filename extension defaults to ‘.ULD’. This command cannot read NDFs. Each point in the USSP spectrum has an associated error index, called epsilon, with the following meanings:
The epsilons are not necessarily reliable indicators of data quality. DIPSO rejects points on input if they are flagged with an epsilon less than or equal to ‘epsmin’ (which defaults to -251), leaving the spectrum in the current arrays. It is forbidden to set epsmin less than or equal -1600, since this would result in totally unflagged, certainly bad, data being acquired.
If epsmin is given a value greater than zero, then all datapoints are read into the next available stack
entry, and the epsilon array into the subsequent stack entry. More subtle doctoring of the data is then
possible, using USSPCLIP (q.v.). However, it is recommended that the default epsmin be accepted
unless you really know what you’re doing, and have good reasons to choose a different
value.
Clips points out of IUE USSP spectra which have ‘epsilons’ less than or equal to epsmin. The data are
expected to have been previously read into the stack using the USSPRD command (with its epsmin
parameter given a positive value); ‘n1’ is the stack entry of the flux data, and ‘n2’ (which
defaults to n1+1) that of the epsilon array. The clipping is done over the wavelength range
w1ww2
(default: full wavelength range).
If mode=1 (the default), VCORR ‘unshifts’ X values back to a zero-velocity reference frame by replacing
the values with
where C is the velocity of light in km/s.
If mode=2, VCORR applies a velocity shift to the data by changing the X values:
If the mode=1 or 2 the Y values are not changed. If mode=-1 or -2 the Y values are also adjusted such that is constant – e.g. if mode=-1 then
and similarly, mutatis mutandis, for mode=-2.
Writes the contents of the ‘current’ arrays into a disk file suitable for subsequent re-reading using
READ. The data can be in either NDF or “native DIPSO” format, depending on the flag
established by the USENDF command. When specifying an NDF, do not include any file
type.
If the name of an existing NDF is given for the second (optional) parameter then the specified NDF is
used as a “model” for the output NDF. The output NDF is initialised to hold a copy of the model NDF
(including all extensions). The data values in the current array are then inserted (if possible) into the
output DATA array at their correct wavelength positions. This provides a mecahanism for creating
NDFs which look like they have been created by other packages. For instance, if you want to use
DIPSO to process the NDF “my_dat” originally created by the JCMTDR package, then you would use
“read my_dat” to read it in as normal, and then you could use “write new_dat my_dat” to write the
processed data to NDF “new_dat”, copying all the JCMTDR extension information (etc)
from the original “my_dat” NDF. The resulting NDF could then be put straight back into
JCMTDR.
Replaces the X values in the current arrays with ABS(X).
Adds a constant, “c", to the X values in the current arrays.
Replaces the X values in the current arrays by [XINT(X)].
Divides the X values in the current arrays by a constant, “c".
Replaces values in the current X array with INT(X).
Replaces the X values in the current arrays by their inverse.
Multiplies the X values in the current arrays by a constant, “c".
Replaces X values in the current arrays by NINT(X).
Subtracts a constant, “c" from the X values in the current arrays.
Cross-correlates the data set in stack entry n1 with the data set in stack entry n2. (Autocorrelation functions can be calculated by defining n1=n2.) Entry n1 contains the “stationary” data.
The range over which the cross-correlation function is evaluated is controlled by the parameters lolag and hilag, which must be in the same units as the stack entries. The default lag is given by:
where there are ‘n” datum points in stack entry n1, with X values from x[1] to x[n]. Then lolag defaults to -lag, and hilag to +lag. The parameter p is the fraction of each data set endmasked (at each end, with a cosine bell), and defaults to 0.05.
The cross-correlation is carried out in the units of the X arrays (which have to be monotonically
increasing, in the same units for both data sets, and at least partially overlapping if you want sensible
results). This is for generality. However, a typical application would be to find velocity shifts between
data sets in wavelength space — but a velocity shift is a function of wavelength in wavelength space.
The way to get round this is to take logs of both data sets (LOGX), then evaluate the correlation
function. Then:
where C is the speed of light (making sure that your units all match up). The arithmetic can all be
done in DIPSO (TENX, XSUB, XMULT).
IMPORTANT: XCORR just does the cross-correlation; it is not a ‘black box’. I can’t think of a situation
in which you shouldn’t first rectify your data (with CREGS and PF, or CDRAW, followed by ADIV)
then subtract the continuum (YSUB 1), for both stack entries, in order to minimise edge
effects.
“Justifies” the X range — i.e. sets X limits to exactly match the range of the data being plotted
(c.p. XT).
The label for the X axis; the default is ‘Wavelength’. If the string contains commas it must be enclosed
in double quotes (c.f TITLE).
Set the maximum X value for plotting to ‘x’. Negated with NX.
Set the minimum X value for plotting to ‘x’. Negated with NX.
Set the X range; x1 is the left-hand value for the plot, x2 the right-hand value. Negated using
NX.
Reverses the ordering of the data in the current X arrays, maintaining X-Y pairing
Sorts the X values in the current arrays into increasing values, maintaining X-Y pairing. Breaks in the data are lost.
“Trims” the X range of a plot — i.e. sets X limits which are some integer multiple of the distance
between tickmarks (c.p. XJ).
Obtain X values using the cursor. Do a cursor hit at the same place twice to get out of XV
mode.
Swaps the contents of the X and Y arrays, maintaining the break arrays unchanged. (Use CLRBRK if this
leads to unwanted results.)
Obtain X and Y values using the cursor. Hitting the same place twice exits XYV mode.
Replaces the Y values in the current arrays with ABS(Y).
Adds a constant, ‘c’, to the Y values stored in the ‘current’ arrays.
Replaces the Y values in the current arrays by [YINT(Y)].
Divides the Y values in the ‘current’ arrays by a constant, ‘c’.
Replaces Y values in the current arrays with INT(Y).
Replaces the Y values in the current arrays by their inverse.
Multiplies the Y values in the ‘current’ arrays by a constant, ‘c’.
Replaces Y values in the ‘current’ arrays with NINT(Y).
Subtract a constant, ‘c’, from the Y values stored in the ‘current’ arrays.
“Justifies” the Y range — i.e. sets the Y limits to exactly match the range of the data being plotted
(c.p. YT).
Replaces the label for the Y axis; the default is ‘Flux’. If the string contains commas it must be enclosed
in double quotes (c.f TITLE).
Sets the maximum Y value for plotting; negated with NY.
Sets the minimum Y value for plotting; negated with NY.
Sets the Y range for plotting; y1 is the lower value for the plotting frame, y2 the upper. Return to
autoscaling with NY or NXY.
“Trims” the Y range of a plot — i.e. sets Y limits which are some integer multiple of the distance
between tickmarks (c.p. YJ).
Obtain Y value. If an X value is supplied, then the corresponding Y value is obtained (by linear interpolation) from the data in the current arrays. Otherwise, the graphics cursor is brought up to permit Y values to be measured from the terminal; hit the same place twice to get out of this mode.
Replaces values in the ‘current’ Y arrays by Y*(X**power). (Some people like to plot data as F*[Lambda**4], for example.)
Calculates a (black-body) Zanstra temperature, using the observed flux, F(obs)
(in erg/cm2/s), of a recombination line. The lines for which this calculation can be performed are (H I)
4861; (He I) 4471, 5876; and (He II) 1640, 4686. The ‘line’ parameter is the wavelength, in Angstroms, of
the selected line.
The calculation requires the location of a ‘continuum’ point longwards of the ionization edge of the
ion concerned. This is obtained from the cursor; thus a plot, in erg/cm2/s/A vs Angstroms (or log10
thereof) is mandatory. Moreover, for a valid result the cursor hit has to correspond to the
dereddened stellar flux; judicious prior use of NEBCONT, ASUB and DRED may therefore be
useful.
T(neb) is the electron temperature of the ionized nebula, and may be input in units of K or 10,000K. This parameter is required because of the (fairly weak) temperature dependence of the ratio, R(line), of the effective recombination coefficients to the ion and line concerned (for a discussion of the physics in the Zanstra method, see e.g. Osterbrock, ‘Astrophysics of Gaseous Nebulae’). The adopted temperature dependences of R(line) are:
where the parameter t = T(neb)/(10,000K).
E(B-V) is used to deredden the line flux (only — i.e. not the continuum flux); a galactic reddening law with R=3.1 is adopted. If this is inappropriate to your data, deredden F(obs) to your own prescription and use E(B-V)=0 (which is the default value).
The output from this command consists of the Zanstra temperature and a normalising constant,
C(norm.). The latter quantity is the number by which a black-body spectrum calculated using BBODY
must be multiplied (YMULT) to make it pass through the (X,Y) co-ordinates selected using the
cursor.
Congratulations on reading this far (unless you’ve cheated, and skipped straight to the end...).
Features, bugs, complaints and comments should be addressed to the Starlink support mailing list
(starlink@jiscmail.ac.uk); but please check the documentation first!