ABS(x)
 Returns the absolute value(s) of x.
ACOS(x)
 Returns the arc cosine of x in radians. The result ranges from 0 to pi.
 Multiply by "#R.D" to get degrees.
ALOG(x)
 Returns the natural logarithm of x.
 See also: "ALOG10", "LOG1P"
ALOG10(x)
 Returns the decimal logarithm (base 10) of x.
 See also: "ALOG", "LOG1P"
AREA,bitmap [, seed, numbers] or AREA(bitmap [, seed, numbers])
 Identifies one or all disjoint areas in 2D integer array <bitmap> which 
 have value one.  Elements with values other than one are disregarded.  
 The identified areas in <bitmap> are filled with a number which uniquely 
 label them.  These numbers are either read from <numbers>, or started at 
 2 and incremented by 1 for each new area.  The function version of "AREA" 
 returns a 1D array containing (grouped by area) the indices of all 
 elements of <bitmap> which belong to an identified area.  The first 
 element of such an area has its index multiplied by -1.  <seed> contains 
 positions within the areas to be identified.  If it is not specified, 
 then all areas are identified.  Specify the seeds as indices to the 
 bitmap.  <numbers>, if defined, must have the same size as <seed>.  If no 
 area is found at the seed positions (or at all), then <bitmap> is not 
 modified, and the function returns the scalar -1.  
ARRAY(type,dimensions)
 Return an array of indicated <type> (0 through 4 for I*1, I*2, I*4, F*4, 
 and F*8, respectively) and <dimensions>.  The latter may be an array 
 containing the dimensions, or a list of up to 8 dimensions.  
 See also: "BYTARR, INTARR, LONARR, FLTARR, DBLARR"
ASIN(x)
 Returns the arc sine of x in radians. The result ranges from -pi/2 to pi/2.
 Multiply by "#R.D" to get degrees.
 See also: "ACOS, ATAN"
ASSOC(lun,array)
 Set up an associated variable for logical unit <lun>.
 Associates the lun with an array of type and structure of <array>.
 See also: " OPENR, DATA TYPES"
ATAN(x)
 Returns the arc tangent of x, in radians. The result ranges from -pi/2 to pi/2.
 Multiply by "#R.D" to get 
 degrees.  
 Reverse: " TAN"
 See also: "ATAN2, ACOS, ASIN"
ATAN2(x,y)
 Returns the angle between the point with coordinates (x,y) and the 
 positive x-axis, counting counterclockwise and in radians.  Multiply by 
 "#R.D" to get degrees.  
 See also: "ATAN"
ATOL(string, [base])
 Converts an ASCII string to a long integer with specified base. The base
 is optional and defaults to base 10. Only one integer is extracted from
 the string and the first unrecognized character ends the decoding. Uses
 include converting octal and hex strings to an integer.
ATOMIZE,symbol [,dump]
 For debugging.  Tries to resolve the ANA executable symbol with number 
 <symbol> and prints the result.  If <dump> is set and unequal to zero, 
 then encountered atomic entities (proper ANA variables) are also "DUMP"ed.  
 See also: " DUMP, SHOW"
BATCH [,value][,/QUIT]
 Toggle batch mode.  In batch mode, all input is read from the current
 input stream.  In direct (= non-batch) mode, commands are read from
 the current input stream, but requested data is read from the
 keyboard.  If <value> is zero, or not specified, then batch mode is
 entered.  Otherwise, direct mode is selected.
BESSEL_J0(x)
 Returns the value of the zero-order Bessel function of the first kind,
 at ordinate x.
 See also: "BESSEL_JN"
 Alias: "BJ0"
BESSEL_J1(x)
 Returns the value of the first-order Bessel function of the first kind,
 at ordinate x.
 See also: "BESSEL_JN"
 Alias: "BJ1"
BESSEL_JN(x,order)
 Returns the value of the Bessel function of the first kind, and of the
 given <order>, at ordinate x.
 See also: "BESSEL_J0, BESSEL_J1, BESSEL_YN"
 Alias: "BJN"
BESSEL_Y0(x)
 Returns the value of the zero-order Bessel function of the second kind,
 at ordinate x.
 See also: "BESSEL_JN, BESSEL_YN"
 Alias: "BY0"
BESSEL_Y1(x)
 Returns the value of the first-order Bessel function of the second kind,
 at ordinate x.
 See also: "BESSEL_JN, BESSEL_YN"
 Alias: "BY1"
BESSEL_YN(x,order)
 Returns the value of the Bessel function of the second kind, and of the
 given <order>, at ordinate x.
 See also: "BESSEL_Y0, BESSEL_Y1, BESSEL_JN"
 Alias: "BYN"
BJ0(x)
 Returns the value of the zero-order Bessel function of the first kind,
 at ordinate x.
 See also: "BJN"
 Alias: "BESSEL_J0"
BJ1(x)
 Returns the value of the first-order Bessel function of the first kind,
 at ordinate x.
 See also: "BJN"
 Alias: "BESSEL_J1"
BJN(x,order)
 Returns the value of the Bessel function of the first kind, and of the
 given <order>, at ordinate x.
 See also: "BJ0, BJ1, BYN"
 Alias: "BESSEL_JN"
BMAP(x)
 Maps x byte for byte into a byte (I*1) array.
 See also: "SMAP, WMAP, LMAP, FMAP, DMAP"
BY0(x)
 Returns the value of the zero-order Bessel function of the second kind,
 at ordinate x.
 See also: "BJN, BYN"
 Alias: "BESSEL_Y0"
BY1(x)
 Returns the value of the first-order Bessel function of the second kind,
 at ordinate x.
 See also: "BJN, BYN"
 Alias: "BESSEL_Y1"
BYN(x,order)
 Returns the value of the Bessel function of the second kind, and of the
 given <order>, at ordinate x.
 See also: "BY0, BY1, BJN"
 Alias: "BESSEL_YN"
BYTARR(dimensions)
 Return an I*1 array of the specified <dimensions>.  The latter may be an 
 array containing the dimensions, or a list of up to 8 dimensions.  
 See also: "ARRAY, INTARR, LONARR, FLTARR, DBLARR"
BYTE(x)
 Returns a byte (I*1) version of x.
 See also: "WORD, LONG, FLOAT, DOUBLE"
BYTGHOST(file, dimensions)
 Links an I*1 array structure with the specified <dimensions> to the 
 specified <file>.  <dimensions> may be an array, or a list of up to ten 
 dimensions.  No file checks are made.  
 See also: "GHOST ARRAY, ARRAY"
C,string,x,y [,size,theta,font,flag][,/DEV,/DVI,/IMG,/PLT,/RIM,/RPL]
 Calligraphy.  Synonym of CALLIG.
CALLIG,string,x,y [,size,theta,font,flag][,/DEV,/DVI,/IMG,/PLT,/RIM,/RPL]
 Draws characters from the Hershy set.  After the call, "!XC" and "!YC" 
 contain the coordinates of the point just after the last drawn character.  
 The characters are drawn in device-independent coordinates, which means 
 they will look squashed if the display is not square.  You may tweak this 
 using "!CRATIO", an extra scaling factor for the characters' y size.  
 <string> = characters to be drawn.  Control characters are preceded by
   a dollar sign $:
.ll-
.  0..9  -  font change to the specified font
.  A, a  -  move up, but keep same size.
.  E, e  -  move up, scale size by factor 0.7.  (exponent)
.  I, i  -  move down, scale size by factor 0.7. (index)
.  B, b  -  move down, keep same size
.  N, n  -  negate combined action of previous A, E, I, and B-s
.  $ > < -  draw $, >, or <
.  S, s  -  save current position
.  R, r  -  restore saved position
.  X, x  -  restore entry font
.  C, c  -  change size to number that follows
.  * .   -  draw a star or a bullet
..
.lll=
. <x, y>   = position in display.  Either pixels or device-independent units.
. <size>   = scale factor.  By default equal to 1.
. <theta>  = rotation angle in degrees.  By default zero.
. <font>   = number of the entry font.
. <flag>   = pen flag.  If equal to zero, the string isn't drawn.  "!XC" and
            "!YC" are set as if the string had been drawn, however.
..
 Coordinate system keywords are available.
 Alias: "C"
 See also:  "COORDINATE SYSTEMS"
CBRT(x)
 Returns the cube root of x.
CG2_OFF
 Redirects screen output to terminal screen.  Not everywhere.
 See also: "CG2_ON"
CG2_ON
 Redirects screen output to CG board.  Not everywhere.
 See also: "CG2_OFF"
CHECKMENU([menu])
 Returns 1 if an entry in menu <menu> has been selected but not yet
 read.  Returns 0 otherwise.
 See also: "MENU, WAIT_FOR_MENU"
CHECKWINDOW([window])
 Returns 1 if the mouse has been clicked in the specified ANA window,
 or in any ANA window; 0 otherwise.  "!IX, !IY, !XTIME, !WINDOW", and
 "!BUTTON" are modified appropriately if the mouse was clicked.
 See also: "XPLACE, CHECKMENU, XQUERY"
CLASSNAME(class)
 Returns the name of the class.
 See also: "TYPENAME, SYMCLASS"
CLOSE,lun
 Close the file on logical unit <lun>.
 See also: "OPENR, OPENW, OPENU"
COLORPIXEL(color_index)
 Maps <color_index> (possibly obtained with "SCALE") to pixel values that
 can be displayed on screen with "TVRAW".
 See also: "TVRAW, TVMAP, TV, SCALE"
COMPRESS(image,xscale [,yscale])
 Returns <image> compressed by integer factors <xscale> in the first,
 and <yscale> (defaults to <xscale>) in the second dimensions.
 See also: "EXPAND, REGRID, REGRID3"
CONCAT,x,y [,...]
 Concatenation  
.rl-
. Calls: "CONCAT,x"       - x is returned with one more dimensions
                        (equal to 1).  Scalars get upgraded to
                        numerical arrays, strings to string arrays.
.        "CONCAT,x,y,..." - <y> and up to 30 more scalars or arrays or strings
                        are appended to x.
..
 Undefined arguments are ignored (as long as at least one argument is
 defined), so starting and appending to an array may be done without
 initialization.  Either all arguments must be numerical, or all strings.
 Alias: "["
CONTOUR,image [,levels,xmin,xmax,ymin,ymax,style,dashsize,autocontour]

 Draw a contour plot of the part of <image> with coordinates between
 <xmin> and <xmax>, and <ymin> and <ymax>, with contour levels
 determined by <levels>: if <levels> is a scalar, then that number of
 levels is used.  If a 1D array, then those contour levels are used.
 If not specified, then "!CONTOUR_NLEV" is used.  <style> specifies
 the line style used in drawing the contours and supersedes the value
 of "!CONTOUR_STYLE", which is the default.  <dashsize> specifies the
 relative size of the dashes (if any) in the chosen line style, and
 defaults to 1.0.  <autocontour> indicates whether the contour levels
 must be calculated from the spread of the data ("/AUTOCONTOUR") or if
 the values in <levels> or "$CONTOURS" must be used
 ("/NOAUTOCONTOUR").  This key supersedes the value of "!AUTOCON",
 which is the default.
 Various options may be set through system variables:
.ll-
. "!contour_nlev"      -   number of levels (defaults to 5)
. "!autocon"           -   flags automatic calculation of contour levels
. "$contours"          -   specifies contour levels when "!autocon" is zero;
                           used contour levels are returned in it
. "!contour_mode"      -   selects type of automatic countouring:
                           0: linear;  1: square root;  2: logarithm
. "!contour_box"       -   flags drawing of one-bin box at origin
. "!contour_border"    -   thickness of border around contour plot
. "!contour_ticks"     -   size of contour tick marks
. "!contour_tick_pen"  -   thickness of contour tick marks
. "!contour_tick_fac"  -   ?
. "!contour_dash_lev"  -   contour level below which contours are dashed.
..
COORDTRF,xold,yold[,xnew,ynew][,/DVI,/DEV,/IMG,/PLT,/RIM,/RPL]
 "[,/TODVI,/TODEV,/TOIMG,/TOPLT,/TORIM,/TORPL]"
 Transforms coordinates <xold,yold> to a different coordinate system.
 Returns the results in <xnew,ynew> if specified, or in <xold,yold>
 otherwise.  The source coordinate system is indicated by one of the
 keywords "DVI, DEV, IMG, PLT, RIM," or "RPL," and the target
 coordinate system by one of those keywords but with "TO" prefixed.
 By default, both the source and target coordinate systems are "DVI."
 See also: "COORDINATE SYSTEMS"
COS(x)
 Returns the cosine of x (which is in radians).  Multiply by "#D.R" to
 get radians from degrees.
 Reverse: " ACOS"
 See also: "SIN, TAN"
COSH(x)
 Returns the hyperbolic cosine of x.
 See also: "TANH, SINH"
CRUNCH,array,byte_slice,out
 Tries to compress <array> and writes the result in byte array <out>, 
 which must be predefined and large enough to contain it.  <byte_slice> 
 specifies how many bits per number are passed to <out> without 
 compression.  Only "BYTE" (I*1) and "WORD" (I*2) arrays are handled.  If 
 <byte_slice> is too small or too large, then less efficient compression 
 or even expansion may occur.  Try three less than the array's entropy for 
 <byte_slice>.  
 Associated system variables:
.ll-
. "!CRUNCH_SLICE"   -      the specified <byte_slice>
. "!CRUNCH_BITS"    -      the number of bits the last compression yielded
. "!CRUNCH_BPP"     -      the number of bits per compressed pixel.
..
 Reverse: "DECRUNCH"
D[,symbol,...][,/FIXED,/SYSTEM,/ZERO,/LOCAL,/CONTEXT,/FOLLOW]
 Shows info about the given <symbol>.  Up to 32 symbols may be specified.
 Alias: "DUMP"
 See also: "ATOMIZE, SHOW"
DBLARR(dimensions)
 Return an F*8 array of the specified <dimensions>.  The latter may be an 
 array containing the dimensions, or a list of up to 10 dimensions.  
 See also: "ARRAY, BYTARR, INTARR, LONARR, FLTARR"
DBLGHOST(file, dimensions)
 Links an F*8 array structure with the specified <dimensions> to the 
 specified <file>.  <dimensions> may be an array, or a list of up to ten 
 dimensions.  No file checks are made.  
 See also: "GHOST ARRAY, ARRAY"
DEBUG
 Experimental routine.
DECOMP,m
 Replaces square matrix <m> with LU decomposition.
 See also: "DSOLVE"
DECRUNCH,code,var
 Decompresses <code> into <var>.
 Reverse: "COMPRESS"
DEFAULT,var1,value1 [,var2,value2,...]
 Assigns values <value1>, <value2>,...  to variables <var1>, <var2>,...  
 if these are undefined.  
 See also: "DEFINED, DELETE"
DEFINED(var)
 Returns 1 if <var> is defined, 0 otherwise.
 See also: "DELETE, DEFAULT"
DELETE,symbol
 Deletes <symbol> by making it undefined.  The "/POINTER" switch (mode
 1) indicates that pointers should themselves be deleted, rather than
 their targets.
 See also: "SWITCH, DEFINED"
DETREND(x [,n])
 Returns x with <n>th-degree polynomial fit removed.
 See also: "TREND"
DIAGNOSTIC [,file]
 Opens file <file> to accept diagnostic output.  <file> defaults to
 "diagnostic.ana".
 See also: "DIAGNOSTIC OUTPUT"
DIFFER(x [,axis,order][,/CENTRAL])
 Returns the running difference of x:  each element of the result 
 contains the difference between the corresponding element of x and its 
 |<order>|-th predecessor.  The predecessor of the first elements are 
 taken to be zero.  If <axis> is not specified (or negative), then x is 
 treated as a 1D array.  Otherwise, differencing is done along the 
 indicated axis (dimension), and is started over for each of the other 
 dimensions.  If <order> is positive, then the unavailable value of each 
 result vector are set to zero;  if negative, then the value from the 
 original is copied (i.e. reverse of "RUNSUM").  The default <order>
 is 1.  The "/CENTRAL" keyword indicates that central differencing is
 performed.
 Reverse: "RUNSUM"
DIMEN(array [,coordinate])
 Returns the size of <array> in the indicated coordinate, which defaults
 to the first (number 0).
 See also "NUM_DIM, NUM_ELEM, SYMCLASS"
DISTR,target,bins,values   and   DISTR(bins,values)
 Distributes the <values> over the corresponding <bins>.  The function 
 form returns an array containing the gathered totals in bins 0 through 
 "MAX(values)".  If <values> contains only 1s, then "DISTR(bins,values)"
 equals "HIST(bins)".  The subroutine form adds the values to the proper 
 bins of <target>.  
 See also: "HIST"
DMAP(x)
 Maps x byte for byte into a double (F*8) array.
 See also: "SMAP, BMAP, WMAP, LMAP, FMAP"
DOUB(x)
 Returns a double (F*8) version of x.
 See also: "BYTE, WORD, LONG, FLOAT"
 Alias: "DOUBLE"
DOUBLE(x)
 Returns a double (F*8) version of x.
 See also: "BYTE, WORD, LONG, FLOAT"
 Alias: "DOUB"
DSOLVE(m,v)
 Solution of linear system of equations
.lll=
. <m> = LU decomposition of equations matrix
. <v> = vector to be matched
..
 See also: "DECOMP"
DSUM(x [,axis])
 Returns the F*8 total of all values in x, or the sum along the
 <mode>th dimension (if <mode> is a nonnegative scalar), or the sum
 for all integer classes in <mode>, where each element of <mode>
 indicates the class that the corresponding member of x belongs to.
 See also: "FSUM"
DUMP[,symbol,...][,/FIXED,/SYSTEM,/ZERO,/LOCAL,/CONTEXT,/FOLLOW]
 Shows info about the given <symbol>.  Up to 32 symbols may be specified.
 By default, info on all named symbols except constants ("#NAME"),
 system variables (predefined "!NAME"), and variables local to
 subroutines and functions is displayed.
 Mode Keywords: (default: 4)
.lrl-
. "FIXED"   -  1 - display all constants ("#NAME")
. "SYSTEM"  -  2 - display all system variables (predefined "!NAME")
. "ZERO"    -  4 - display all new zero-context variables
. "LOCAL"   -  8 - display all new local variables
. "CONTEXT" - 24 - the following argument is the context whose
                   variables are displayed.
..
 Alias: "D"
 See also: "SHOW, DUMP_FUNC, DUMP_SUBR, DUMP_STACK, ATOMIZE"
DUMP_FUNC
 Shows all compiled user functions and the variables created therein.
 See also: "DUMP_SUBR, DUMP, SHOW"
DUMP_STACK
 Shows all variables currently on the variables stack in the same format
 as is used in "DUMP".  The displayed variable names are the names of the
 variables that were pushed on the stack and are memory aids only.
 See also: "PUSH, POP, DUMP"
DUMP_SUBR
 Shows all compiled user subroutines and the variables created therein.
 See also: "DUMP_FUNC, DUMP, SHOW"
ECHO
 From now on, code that's being compiled will be echoed to the screen.
 Reverse: "NOECHO"
ENDIAN,x
 Swap bytes in each element of numerical array x according to its
 data type in order to switch between littleendian and bigendian
 representation of numbers.  In "FLOAT" and "LONG" numbers, bytes
 0 and 3 are swapped, and bytes 1 and 2.  In a "DOUBLE" number, bytes
 0 and 7 are swapped, and also bytes 1 and 6, bytes 2 and 5, and bytes
 3 and 4.  In a "WORD" number, bytes 0 and 1 are swapped.  A "BYTE"
 number is left unchanged.
ERASE [window]
 Erases the indicated or current window.
ERF(x)
 Returns the value of the error function at ordinate x.
 Ranges between -1 and +1.  The chance of exceeding value <v> when
 drawing from a standard normal distribution is "0.5 -
 0.5*ERF(v/SQRT(2))". 
 See also: "ERFC"
ERFC(x)
 Returns one minus the value of the error function at ordinate x.
 Ranges between 0 and 2.
 See also: "ERF"
ERROR,text[,symbol]
 Generates an error with error message <text>, pointing at symbol
 <symbol>. 
ESMOOTH(x [,order,axis])
 Exponential asymmetric smoothing.  Only elements before and including the 
 current element enter into the smooth result corresponding to the current 
 element.  The weight of each element drops exponentially with its 
 distance to the current element.  The decay length is <order> and 
 defaults to 3.  The dimension along which smoothing occurs is <axis>.  If 
 <axis> is not specified, then x is treated as a one-dimensional array.  
 See also: "SMOOTH, GSMOOTH, RUNSUM, RUNCUM"
EVAL(x)
 Compiles and evaluates a string, interpreting it as ANA code.  Non-strings
 are just evaluated.
 See also: "EXECUTE"
EXECUTE(x)
 Compiles and executes a string, interpreting it as ANA code.
 See also: "EVAL"
EXIT[,status]
 Exits ANA with <status>.
EXP(x)
 Returns the natural antilogarithm of x.
 Reverse: "ALOG"
EXPAND(array,xscale [,yscale,smooth])
 Returns an expanded version of 1D or 2D <array>.
.lll=
. <xscale> = magnification in 1st dimension (may be non-integer; may be < 1)
. <yscale> = magnification in 2nd dimension (defaults to <xscale>)
. <smooth> = smoothing type: 0 -> bilinear interpolation; else -> nearest
             neighbor (defaults to "!TVSMT")
..
 See also: "COMPRESS, REGRID, REGRID3"
EXPM1(x)
 Returns the natural antilogarithm of x, minus 1.
 Reverse: "LOG1P"
F0H,file [,header]
 Reads header string of FZ <file>, and displays it or stores it in <header>.
 See also: "F0READ, F0WRITE, FCWRITE"
 Alias: "F0HEAD, FZH, FZHEAD"
F0HEAD,file [,header]
 Reads header string of FZ <file>, and displays it or stores it in <header>.
 See also: "F0READ, F0WRITE, FCWRITE"
 Alias: "F0H, FZH, FZHEAD"
F0R,var,file [,header]
 Retrieves <var> and <header> from (compressed or uncompressed) FZ <file>.  
 If <header> is not specified, then the header is displayed on the screen.  
 The function form returns 1 if successful, otherwise 0.  
 See also: "FZHEAD, FZWRITE, FCWRITE"
 Alias: "FZREAD, FZR, F0READ"
F0READ,var,file [,header]
 Retrieves <var> and <header> from (compressed or uncompressed) FZ <file>.
 If <header> is not specified, then the header is displayed on the screen.
 The function form returns 1 if successful, otherwise 0.
 See also: "FZHEAD, FZWRITE, FCWRITE"
 Alias: "FZREAD, F0R, FZR"
F0W,var,file [,header]
 Writes <var> and <header> string into uncompressed FZ <file>.
 See also: "FCWRITE, FZREAD, FZHEAD"
 Alias: "FZWRITE, FZW, F0WRITE"
F0WRITE,var,file [,header]
 Writes <var> and <header> string into uncompressed FZ <file>.
 See also: "FCWRITE, FZREAD, FZHEAD"
 Alias: "FZWRITE, FZW, F0WRITE"
FCW,var,file [,header]  or  FCW(var,file [,header])
 Writes <var> and <header> string into compressed FZ <file>.
 Function version returns 1 on success, 0 on failure, and no error
 messages.
 See also: "FZWRITE, FZREAD, FZHEAD"
 Alias: "FCWRITE"
FCWRITE,var,file [,header]  or  FCWRITE(var,file [,header])
 Writes <var> and <header> string into compressed FZ <file>.
 Function version returns 1 on success, 0 on failure, and no error
 messages.
 See also: "FZWRITE, FZREAD, FZHEAD"
 Alias: "FCW"
FIND(array,key[,offset,mode][,/EXACT,/INDEX_GE,/VALUE_GE,/FIRST)
 Finds the first occurrence of each element of <key>, starting at
 (the corresponding element of) <offset> in <array>.  <mode> indicates
 the search mode, which may also be specified by the mode keywords
 mentioned in parentheses.  The <mode>s are:
.rll-
.  0 - "/EXACT"    - Seek exact match, return offset or -1 if none found.
.  1 - "/INDEX_GE" - Seek first element greater than or equal to
            the key.  Return offset or -1, if none found.
.  2 - "/VALUE_GE" - Seek first element greater than or equal to
  	    the key.  Return found element, or -1 if none found.
.  4 - "/FIRST"    - Return a result for each combination of the
	    second and higher dimensions.
..
 E.g., if "X" is a 3*2*4 array, then "FIND(X,1)" returns the index of
 the first occurrence of "1" in "X", irrespective of the dimensional
 structure of "X", and "FIND(X,1,/FIRST)" returns a 2*4 array with
 the results of "FIND(X(*,0,0),1), FIND(X(*,1,0),1)," etcetera.
 <key> may be a scalar or an array.
 <offset> may be a scalar or an array of the same size as <key>.
FIND_MAXLOC(image)
 Returns the indices of all local maximums in <image>.  Each value that is 
 larger than the eight values immediately surrounding it is a local 
 maximum.  The edges are not considered.  Of a multi-pixel local maximum, 
 only one position is returned.  Of a plateau that is larger than three by 
 three pixels, the algorithm cannot determine whether it is a local 
 maximum or not, and one position in the plateau is returned.  
FIX(x)
 Return a long (I*4) version of x.
 See also: "BYTE, WORD, FLOAT, DOUBLE"
 Alias: "LONG"
FJREAD,array,file [,header]
 Retrieves <array> and <header> from JPEG-compressed <file>.
 If <header> is not specified, then the header is displayed on the screen.
 See also: "FJWRITE, FCWRITE, FZWRITE"
FJWRITE,array,file [,header,quality]
 Writes 2D "BYTE" <array> and <header> to <file> in JPEG-compressed format 
 with compression quality parameter <quality>, which ranges between 0 and 
 100 and defaults to 75.  JPEG compression discards part of the 
 resolution.  The higher <quality> is, the larger the file size and the 
 less information is lost in the compression.  You may omit <header> or 
 <quality> or both.  
 See also: "FJREAD, FCWRITE, FZWRITE"
FLOAT(x)
 Return a float (F*4) version of x.
 See also: "BYTE, WORD, LONG, DOUBLE"
FLTARR(dimensions)
 Return a F*4 array of the specified <dimensions>.  The latter may be an 
 array containing the dimensions, or a list of up to 10 dimensions.  
 See also: "ARRAY, BYTARR, INTARR, LONARR, DBLARR"
FLTGHOST(file, dimensions)
 Links an F*4 array structure with the specified <dimensions> to the 
 specified <file>.  <dimensions> may be an array, or a list of up to ten 
 dimensions.  No file checks are made.  
 See also: "GHOST ARRAY, ARRAY"
FMAP(x)
 Maps x byte for byte into a float (F*4) array.
 See also:  "SMAP, BMAP, WMAP, LMAP, DMAP"
FORMAT_SET[,string]
 Sets or displays the format strings !FORMAT_I, !FORMAT_F and !FORMAT_S 
 used as defaults by PRINT and PRINTF.  
 See also PRINTING FORMATS, FPRINT
FPRINT,format,arg1 [,arg2,...]
 Print arguments under guidance of format string <format>.  The last 
 format entry is repeated as necessary to service all arguments.  
 Arguments are cast to the type appropriate for the next format entry.  
 One format entry is used per argument, unless the first character of 
 <format> is a minus sign, in which case on format entry per output number 
 is used.  C format strings are legal.  
 See also: " FORMAT STRINGS"
FPRINTF,lun,format,arg1 [,arg2,...]
 Print arguments to a file on logical unit <lun> under guidance of format 
 string <format>.  The last format entry is repeated as necessary to 
 service all arguments.  Arguments are cast to the type appropriate for 
 the next format entry.  One format entry is used per argument, unless the 
 first character of <format> is a minus sign, in which case on format 
 entry per output number is used.  C format strings are legal.  
 See also: " FORMAT STRINGS"
FREADF,lun,format,arg1 [,arg2,...]
 Read arguments from a file on logical unit <lun> under guidance of
 format string <format>.  <format> consists of (ignored) whitespace,
 characters which are expected on input, and format items.  All C
 scanf format types are legal, expect %n and %p.  Additionally, the
 number of elements to read into an array can be indicated with the #
 notation: %d6# reads 6 LONGs into an array. 
 See also: "FORMAT STRINGS" (which is mostly valid for reading, too).
FSTRING(format, argument1 [,argument2 ...])
 Transforms the arguments to an ASCII string under guidance of C-format 
 string <format>.  The argument is modified to suit the data type expected 
 by the format.  
 See also: "STRING, FPRINT, PRINTING FORMATS"
FSUM(x [,mode])
 Returns the F*4 total of all values in x, or the sum along the
 <mode>th dimension (if <mode> is a nonnegative scalar), or the sum
 for all integer classes in <mode>, where each element of <mode>
 indicates the class that the corresponding member of x belongs to.
 See also: "DSUM"
FZH,file [,header]    or FZH(file [,header])
 Reads header string of FZ <file>, and displays it or stores it in <header>.
 The function form returns 1 when successful, 0 otherwise.
 See also: "F0READ, F0WRITE, FCWRITE"
 Alias: "F0H, F0HEAD, FZHEAD"
FZHEAD,file [,header]    or FZHEAD(file [,header])
 Reads header string of FZ <file>, and displays it or stores it in <header>.
 The function form returns 1 when successful, 0 otherwise.
 See also: "F0READ, F0WRITE, FCWRITE"
 Alias: "F0H, FZH, F0HEAD"
FZR,var,file [,header]   or FZR(var,file [,header])
 Retrieves <var> and <header> from (compressed or uncompressed) FZ <file>.
 If <header> is not specified, then the header is displayed on the screen.
 The function form returns 1 if successful, otherwise 0.
 See also: "FZHEAD, FZWRITE, FCWRITE"
 Alias: "FZREAD, F0R, F0READ"
FZREAD,var,file [,header]    or FZREAD(var,file [,header])
 Retrieves <var> and <header> from (compressed or uncompressed) FZ <file>.
 If <header> is not specified, then the header is displayed on the screen.
 The function form returns 1 if successful, otherwise 0.
 See also: "FZHEAD, FZWRITE, FCWRITE"
 Alias: "FZR, F0R, F0READ"
FZW,var,file [,header]  or  FZW(var,file [,header])
 Writes <var> and <header> string into uncompressed FZ <file>.
 Function version returns 1 on success, 0 on failure, and no error
 messages.
 See also: "FCWRITE, FZREAD, FZHEAD"
 Alias: "FZWRITE, F0W, F0WRITE"
FZWRITE,var,file [,header]  or  FZWRITE(var,file [,header])
 Writes <var> and <header> string into uncompressed FZ <file>.
 Function version returns 1 on success, 0 on failure, and no error
 messages.
 See also: "FCWRITE, FZREAD, FZHEAD"
 Alias: "FZW, F0W, F0WRITE"
GAMMA(x)
 Returns values of the gamma function.  If the "/LOG" keyword is
 specified, then the natural logarithm of the absolute value of the
 gamma function is returned.
GETMIN9,array,x,y
 Returns the position of the interpolated minimum of 3 by 3 element <array>
 in <x, y>.
GRIDMATCH(image, compare_image, grid_x, grid_y, window_x, window_y,
          gausswidth [,mode])
 Determines and returns displacement of one image compared to another
.lll=
. <image>         = image under investigation
. <compare_image> = reference image
. <grid_x,grid_y> = coordinates of grid of (centers of) subimages
. <window_x,..y>  = dimensions of subimages
. <gausswidth>    = width of gaussian mask
. <mode>          = flag. If set, quality parameter is also returned,
                   as third "coordinate" of result.
..
GSMOOTH(image [[,axis],kernel)
 Returns a smoothed version of <image>.  If <kernel> is a scalar, then
 smoothing is done with a Gaussian kernel of FWHM <kernel>.  If
 <kernel> is an array, then smoothing is done with weights given by
 <kernel>.  In that case, <kernel> must have an odd size and the
 central element of <kernel> is the center of smoothing.  <axis>
 indicates the dimension along which the smoothing is performed.  Keyword
 "NORMALIZE" (1) determines if a user-defined kernel is
 normalized (internally) to a total of 1 before use.  By default, the
 kernel is normalized.  Keyword "PARTIAL" (2) determines the weight of
 assumed zeros beyond the edges of <image>:  if set (default), then
 weighted sums near edges are normalized with partial sums of the
 weights that fall between the edges, otherwise normalization is
 always with the full sum of weights (i.e. the zeros beyond the edges
 weight in according to the kernel).
 See also: " SMOOTH, ESMOOTH"
HAIRS
 Alias: "XPLACE"
HELP[,topic][,/EXACT,/ROUTINE,/TREE,/SUBROUTINE,/FUNCTION,/LIST]
 Yields information on the specified <topic> (a string).  To see all the
 options, try "HELP".  To start at the top of the help tree, try "HELP,'ANA'".
 <topic> is case-insensitive.
 Mode Keywords:  (default: 30)
.lrl-
. "EXACT"      -  1 -  asks for an exact match (not left-partial)
. "ROUTINE"    -  2 -  searches through the list of internal routines
. "TREE"       -  4 -  searches the help tree
. "SUBROUTINE" -  8 -  searches for a user-defined subroutine
. "FUNCTION"   - 16 -  searches for a user-defined function
. "LIST"       - 32 -  lists all matches
..
HIST(x[,/FIRST])
 Returns the histogram (binsize 1) of x.  The histogram runs between
 values "!HISTMIN" (zero or negative) and "!HISTMAX", and cannot exceed
 "!MAXHISTSIZE" bins.
HISTR(x[,/FIRST])
 Returns the normalized (to between 0 and 1) running sum of the histogram
 of x.
IGNORE
 When this command is issued, all following commands, except for the
 "RESUME" command, are ignored.  This can be used to "comment out" large
 chunck of code.  Nested "IGNORE"s are supported, i.e. code is only then
 again interpreted, when as many "RESUME"s have been issued as have
 "IGNORE"s.
 See also: " RESUME"
INDEX(array[,/AXIS])
 Returns an index array of a sorted version of <array>, i.e.
 "array(INDEX(array))" is sorted in ascending order.  If keyword "/AXIS"
 (1) is specified, then indexing occurs along the 0th dimension,
 otherwise the array is seen as a large 1-dimensional one.
 See also: "SORT"
INDGEN(var [,coordinate])
 Replaces each element of <var> with its index (or the specified <coordinate>).
 See also: "ZERO"
INSERT,target,source [,origin,axes]
 Inserts <source> in <target> starting at position <origin>, each subsequent
 <source>-axis along the <target>-axes in <axes>.  <origin> defaults to
 all zeroes, and <axes> to "[0,1,2,...]".  All arguments must be (general)
 arrays.
INT(x)
 Returns a word (I*2) version of x.
 See also "BYTE, LONG, FLOAT, DOUBLE"
 Alias: "WORD"
INTARR(dimensions)
 Return an I*2 array of the specified <dimensions>.  The latter may be
 an array containing the dimensions, or a list of up to 10 dimensions.
 See also: "ARRAY, BYTARR, LONARR, FLTARR, DBLARR"
INTGHOST(file, dimensions)
 Links an I*2 array structure with the specified <dimensions> to the
 specified <file>.  <dimensions> may be an array, or a list of up to ten
 dimensions.  No file checks are made.
 See also: "GHOST ARRAY, ARRAY"
ISARRAY(var)
 Returns 1 if <var> is an array, zero otherwise.
 See also:  "ISSCALAR, ISSTRING, SYMDTYPE, SYMCLASS"
ISBRIGHTBEAD(im)
 Returns 1s where a pixel in image <im> shows negative curvature
 in the four directions N-S, E-W, NE-SW, and NW-SE, and zeros
 elsewhere.  The one-pixel wide edge is not considered (ie. always
 zero).
ISSCALAR(var)
 Returns 1 if <var> is a scalar, zero otherwise.
 See also:  "ISARRAY, ISSTRING, SYMDTYPE, SYMCLASS"
ISSTRING(var)
 Returns 1 if <var> is a string, zero otherwise.
 See also:  "ISARRAY, ISSCALAR, SYMDTYPE, SYMCLASS"
IST(x [,n,mode])
 String equivalent of number x.
.lll=
. <n>    = length of the string (modes 0 and 1)
. <mode> = 0: pad with blanks, 1: pad with zeros, 2: fits snugly
..
ISTRING(x [,n,mode])
 String equivalent of number x.
.lll=
. <n>    = length of the string (modes 0 and 1)
. <mode> = 0: pad with blanks, 1: pad with zeros, 2: fits snugly
..
LIMITS [,xmin,xmax,ymin,ymax]
 Sets or shows limits on the coordinate ranges shown in the next plot.
LMAP(x)
 Maps x byte for byte into a long (I*4) array.
 See also: "SMAP, BMAP, WMAP, FMAP, DMAP"
LOCAL_MAX(array, position [,/SUBGRID,/BOUND])
 Finds position and value of local maximums in <array>, starting at
 the positions (indices) in <position> and walking along the steepest
 gradient.  <position> must be a scalar or array of indices to
 <array>.  By default, a local maximum on the grid is returned.
 Keyword "/SUBGRID" requests the estimated value at the subgrid position,
 with the subgrid position forced to be within one px in any
 coordinate from the integer position.  Allow far subgrid results by
 also specifying "/NOBOUND".
 The last found maximum is put in "!LASTMAX" and "!LASTMAXLOC" (the
 latter always integer), and all found values are returned.
 See also: "LOCAL_MAXLOC, LOCAL_MIN, LOCAL_MINLOC, MAX, MAXLOC"
LOCAL_MAXLOC(array, position [,/SUBGRID])
 Finds position and value of local maximums in <array>, starting at
 the positions (indices) in <position> and walking along the steepest
 gradient.  <position> must be a scalar or array of indices to
 <array>.  By default, a local maximum on the grid is returned.
 Keyword "/SUBGRID" requests the subgrid position,
 with the subgrid position forced to be within one px in any
 coordinate from the integer position.  Allow far subgrid results by
 also specifying "/NOBOUND".
 The last found maximum is put in "!LASTMAX" and "!LASTMAXLOC" (the
 latter always integer), and all found positions are returned as
 indices to <array>.
 See also: "LOCAL_MAX, LOCAL_MIN, LOCAL_MINLOC, MAX, MAXLOC"
LOCAL_MIN(array, position [,/SUBGRID])
 Finds position and value of local minimums in <array>, starting at
 the positions (indices) in <position> and walking along the steepest
 gradient.  <position> must be a scalar or array of indices to <array>.
 By default, a local minimum on the grid is returned.  Keyword
 "/SUBGRID" requests the estimated value at the subgrid position,
 with the subgrid position forced to be within one px in any
 coordinate from the integer position.  Allow far subgrid results by
 also specifying "/NOBOUND".  The last found minimum is put in
 "!LASTMIN" and "!LASTMINLOC" (the latter always integer), and all
 found values are returned. 
 See also: "LOCAL_MINLOC, LOCAL_MAX, LOCAL_MAXLOC, MIN, MINLOC"
LOCAL_MINLOC(array, position [,/SUBGRID])
 Finds position and value of local minimums in <array>, starting at
 the positions (indices) in <position> and walking along the steepest
 gradient.  <position> must be a scalar or array of indices to <array>.
 By default, a local minimum on the grid is returned.  Keyword
 "/SUBGRID" requests the estimated value at the subgrid position,
 with the subgrid position forced to be within one px in any
 coordinate from the integer position.  Allow far subgrid results by
 also specifying "/NOBOUND".  The last found minimum is put in
 "!LASTMIN" and "!LASTMINLOC" (the latter always integer), and all
 found positions are returned as indices to <array>. 
 See also: "LOCAL_MIN, LOCAL_MAX, LOCAL_MAXLOC, MIN, MINLOC"
LOG(x)
 Returns the natural logarithm of x.
 Reverse: "EXP"
LOG10(x)
 Returns the decimal logarithm (base 10) of x.
LOG1P(x)
 Returns the natural logarithm of one more than x.
 Reverse: "EXPM1"
LONARR(dimensions)
 Return an I*4 array of the specified <dimensions>.  The latter may be
 an array containing the dimensions, or a list of up to 10 dimensions.
 See also: "ARRAY, BYTARR, INTARR, FLTARR, DBLARR, STRARR"
LONG(x)
 Return a long (I*4) version of x.
 See also: "BYTE, WORD, FLOAT, DOUBLE"
 Alias: "FIX"
LONGHOST(file, dimensions)
 Links an I*4 array structure with the specified <dimensions> to the
 specified <file>.  <dimensions> may be an array, or a list of up to ten
 dimensions.  No file checks are made.
 See also: "GHOST ARRAY, ARRAY"
LOWCASE(s)
 All letters in <s> converted to lower case.
 Alias: "LOWER"
 See also: "UPCASE, UPPERCASE"
LOWER(s)
 All letters in <s> converted to lower case.
 Alias: "LOWCASE"
 See also: "UPCASE, UPPERCASE"
MAKEMANUAL[,/A4,/STYLE]
 Turns the ANA help file "$ANADIR/anahelp" into a LaTeX file called
 "$ANADIR/ana_manual.tex".  If a file "$ANADIR/ana_manual.sty" exists,
 it is used for preamble to the LaTeX file, otherwise some default
 preamble is used.  Keyword "/A4" includes a4.sty.
MATCH(set, target)
 Returns a 1 for each element of <target> which is in <set>, a 0 otherwise.
MATMUL(m1,m2)
 Product of matrices <m1> and <m2>
MAX(array [,mode])
 Returns the maximum value of <array>.  If <mode> is a scalar, then
 returns the maximum values obtained by scanning along the <mode>th
 dimension for all values of the other coordinates.  If <mode> is an
 array, then its elements indicate the class each corresponding
 element of <array> belongs to and the maximum value in each class is
 returned, with the maximum of class 0 or the most negative class in
 the first element of the result.
 E.g. if  "X = [[[1,7]],[[3,5]]]"  then
   "MAX(X) = 7   MAX(X,0) = [7,5]  MAX(X,1) = [3,7]"
 See also: " MAXLOC, MIN, MINLOC, !LASTMAX, !LASTMAXLOC, !LASTMIN,
 !LASTMINLOC, LOCAL_MAX, LOCAL_MAXLOC, LOCAL_MIN, LOCAL_MINLOC, FIND_MAXLOC"
MAXLOC(array [,mode])
 Returns the coordinates of the maximum value of <array>.  If <mode> is a 
 scalar, then returns the coordinates of the maximum values obtained by 
 scanning along the <mode>th dimension for all values of the other 
 coordinates.  If <mode> is an array, then its elements indicate the class 
 each corresponding element of <array> belongs to and the position of
 the maximum value in each class is returned, with the maximum of
 class 0 or the most negative class in the first element of the result.
 E.g. if  "X = [[[1,7]],[[3,5]]]" then
   "MAXLOC(X) = [1,0]  MAXLOC(X,0) = [1,1]  MAXLOC(X,1) = [1,0]"
 See also: " MAX"
MEAN(x [,mode])
 Returns the F*4 arithmetic mean of the values in x.  If <mode> is a
 scalar, then the mean is determined along the indicated dimension.
 If <mode> is an array, then its elements denote the class of each
 corresponding element of x and the means of all classes are
 returned, with as first element of the result the mean of class 0 or
 the most negative class.
 See also: "TOTAL, SDEV"
MEMSTAT
 Yields information about memory use.  Doesn't work in Utrecht.
MENU,menu [,x,y] [,title,item1 [,...] ]
 Define and/or (re)display a menu.  The menu number is <menu>.
 The menu is displayed at position <x,y>, or at its old position.
 The menu entries are the old ones, or <title>, <item1>,
 etc., which must be strings.  The strings may also be specified
 in one string array (i.e. MENU,menu,x,y,string_array).
 If the position is specified, then the menu is created and/or
 resized.  If only the text is specified, then the menu is not
 resized.  If only the menu number is specified, then the menu
 is redisplayed.
 See also: "MENUHIDE, MENUITEM, MENUZAP, CHECKMENU, WAIT_FOR_MENU,
           MENUPOP"
MENUHIDE,menu
 Hides menu number <menu>.  The menu contents are preserved, and
 the menu may be redisplayed at any time with MENU.
 See also: "MENU, MENUZAP"
MENUITEM,menu,item,text
 Replaces item number <item> in menu number <menu> with <text>.  (The
 first item is number 1.)  The size of the menu is not changed.
 See also: " MENU"
MENUPOP,menu [,title,item1 [,...] ]
 Define and/or (re)display a menu.  Same as "MENU", but gets position
 to display menu at from current mouse position.
 See also: " MENU, MENUHIDE, MENUZAP, CHECKMENU, WAIT_FOR_MENU"
MENUREAD,menu,menu_item,query_string,answer_string
 Displays <query_string> in item <menu_item> of <menu>, accepts
 a user-typed string and returns that string in <answer_string>.  The
 first menu item has <menu_item> equal to 1.
 See also: " MENU, MENUITEM"
MENUZAP,menu
 Discards menu number <menu>.  The menu contents are destroyed.
 See also: "MENU, MENUHIDE"
MIN(array [,mode])
 Returns the minimum value of <array>.  If <mode> is a scalar, then
 returns the minimum values obtained by scanning along the <mode>th
 dimension for all values of the other coordinates.  If <mode> is an
 array, then its elements indicate the class each corresponding
 element of <array> belongs to and the minimum value in each class is
 returned, with the minimum of class 0 or the most negative class in
 the first element of the result.
 E.g. if  "X = [[[1,7]],[[3,5]]]"  then
   "MIN(X) = 1   MIN(X,0) = [1,3]  MIN(X,1) = [1,5]"
 See also: " MAX, MAXLOC, MINLOC, !LASTMAX, !LASTMAXLOC, !LASTMIN,
 !LASTMINLOC, LOCAL_MAX, LOCAL_MIN, LOCAL_MAXLOC, LOCAL_MINLOC, FIND_MAXLOC"
MINLOC(array [,axis])
 Returns the coordinates of the minimum value of <array>.  If <mode> is a 
 scalar, then returns the coordinates of the minimum values obtained by 
 scanning along the <mode>th dimension for all values of the other 
 coordinates.  If <mode> is an array, then its elements indicate the class 
 each corresponding element of <array> belongs to and the position of
 the minimum value in each class is returned, with the minimum of
 class 0 or the most negative class in the first element of the result.
 E.g. if  "X = [[[1,7]],[[3,5]]]" then
   "MINLOC(X) = [0,0]  MINLOC(X,0) = [0,0]  MINLOC(X,1) = [0,1]"
 See also: " MIN"
NOECHO
 Disables echoing of compiling command lines.
 Reverse: "ECHO"
NOT(x)
 Returns a 1 for each zero element of x, and a 0 otherwise.
NUM_DIM(x)
 Returns the number of dimensions in x.
 See also: "NUM_ELEM, SYMCLASS, SYMDTYPE"
NUM_ELEM(x)
 Returns the number of elements (or characters) in x.
 See also: "NUM_DIM, DIMEN, SYMCLASS, SYMDTYPE"
OP1CMD,string    or OP1CMD(string)
 Sends <string> to optical disk player/recorder #1.  The subroutine
 displays the (printable part of the) response string on screen; the
 function form returns that string.
 See also: "CG2_ON, CG2_OF"
OPENR,lun,file [,/GET_LUN]   or OPENR(lun,file [,/GET_LUN])
 Opens <file> (a string) on logical unit number <lun> for reading.
 The "/GET_LUN" switch instructs ANA to find and use the next free
 logical unit and to return the found unit number in <lun>.
 The function form returns 1 when successful, 0 otherwise.
 See also: "OPENW, OPENU, CLOSE, PRINTF, ASSOC, REWIND"
OPENU,lun,file [,/GET_LUN]  or OPENU(lun,file [,/GET_LUN])
 Opens <file> (a string) on logical unit number <lun> for updating
 (i.e. appending at the end).
 The "/GET_LUN" switch instructs ANA to find and use the next free
 logical unit and to return the found unit number in <lun>.
 The function form returns 1 when successful, 0 otherwise.
 See also: "OPENW, OPENR, CLOSE, PRINTF, ASSOC, REWIND"
OPENW,lun,file [,/GET_LUN]  or OPENW(lun,file [,/GET_LUN])
 Opens <file> (a string) on logical unit number <lun> for writing.
 The "/GET_LUN" switch instructs ANA to find and use the next free
 logical unit and to return the found unit number in <lun>.
 The function form returns 1 when successful, 0 otherwise.
 See also: "OPENR, OPENU, CLOSE, PRINTF, ASSOC, REWIND"
OPLOT[,XDATA],YDATA[,SYMBOL,LINE,DASHSIZE,BREAKS]
 "[,/DEV,/DVI,/IMG,/PLT,/RIM,/RPL]"
 Plots <ydata> vs. <xdata>, indicating the points with symbols of type
 <symbol>, connected by lines of type <line>, in the previous plot,
 breaking the plot (i.e. moving but not drawing) at the indices
 in array <breaks>.  <dashsize> scales the size of the dashes in
 non-solid lines.
 See also: "PLOT, COORDINATE SYSTEMS, Plot Modes"
PDEV[,device]
 Sets the plot device, or displays all valid choices.
PEN[,width,gray]
 Shows or sets the width and gray level of the plot pen.
PIT(x [,y,n])
 Coefficients of <n>th-degree polynomial fit of <y> to x.
 <n> defaults to 0, x to indgen of <y>
PLOT[,XDATA],YDATA[,SYMBOL,LINE,DASHSIZE,BREAKS
 ",XTITLE,YTITLE,TITLE,BOUNDINGBOX]"
 "[,/DEV,/DVI,/IMG,/PLT,/RIM,/RPL,/NOERASE]"
 Plots <ydata> vs. <xdata> with the points indicated by symbols of
 style <symbol>, connected with lines of style <line>, and with labels
 <xtitle> and <ytitle> along the axes and <title> above the plot.
 Breaks the plot, i.e. moves but does not draw, at the indices in array
 <breaks>.  <boundingbox> determines which limits are stored in "!BXB,
 !BXT, !BYB, !BYT":  0 -> determined by "!BB_UPDATE";  1 -> plot
 window + small xy labels; 2 -> plot window + big xy labels +
 title;  3 -> plot window + small xy labels + title;  4 -> plot window
 + big xy labels + title;  5 -> plot window only;  6 -> everything.
 Only the relative order of arguments of the same data class (scalar,
 array, string) matters; i.e. "PLOT,xtitle,ydata,breaks,ytitle" is a
 valid call.  The first 6 keywords are coordinate system keywords.
 "/ERASE," which is selected by default, clears the device before
 drawing the plot.  <dashsize> scales the size of the dashes in
 non-solid lines.
 See also: "OPLOT, COORDINATE SYSTEMS, Plot Modes"
POLATE(x,y,xnew)
 Linearly interpolates in table <y> vs. x at <xnew>.
 x must be in ascending order.
 2D -> series of 1D interpolations.  1D x and 2D <y> allowed.
 "TABLE" is more general and "POLATE" seems to have a bug
 somewhere.
 See also: "TABLE"
POLY(x,c)
 Evaluate polynomials with coefficients <c> at x
POP,var [,var...]
 Pops variables from the variables stack (last in - first out) and
 assigns them to the specified <var>s.  Up to 32 variables may be
 specified.
 See also: "PUSH, DUMP_STACK"
POSTIMAGE,image [,x1,x2,y1,y2]
 Write <image> (or the part with coordinates between <x1> and <x2>, <y1> and
 <y2>) to file junk.eps in the current directory.
POSTRAW,1,1, ana_postraw,
POSTREL
 Properly finish the current PostScript file.
 Alias:  "POSTRELEASE"
POSTRELEASE
 Properly finish the current PostScript file.
 Alias: "POSTREL"
POWER(x)
 Returns the power spectrum of x along the first dimension.
PRINT,x [,...]
 Displays the contents of symbol x on the screen.  Up to 32 arguments
 may be specified.
 See also: "PRINTF"
 Alias: "T, TY, TYPE"
PRINTF,lun,x [,...]   or  PRINTF(lun,x [,...])
 Writes the contents of symbol x to the file opened on logical unit <lun>.
 Up to 32 arguments may be specified.  The function form returns 1 upon
 exit when successful, 0 otherwise.
 See also: "PRINT"
PROJECTION[,matrix,TRANSLATE,ROTATE,SCALE,PERSPECTIVE,OBLIQUE,]
 "[/RESET,/ORIGINAL]"
 Sets, resets, or modifies the current 3D-transformation/projection.  The 
 following keywords and arguments are always treated in the order in which 
 they are mentioned here:  Keyword "/RESET" fills transformation matrix 
 "#P3D" with the identity transformation.  <matrix> is a 16-element (4 
 by 4) matrix to be entered into "#P3D".  <translate> is a 3-element 
 vector, <perspective> for perspective projections either a 3-element 
 vector containing the relative screen coordinates of the center of the 
 projection and the distance that the eye is behind the projection plane, 
 or just the distance, <rotate> a 3-element vector containing the rotation 
 angles around the x, y, and z axis, <scale> a 3-element vector containing 
 the scale factors in the x, y, and z directions, <oblique> a 2-element 
 vector for isometric projections (first element:  size, second element:  
 angle).  "/ORIGINAL" selects the original (data-cube related) coordinate
 system instead of the current (screen-related) coordinate system.  All
 arguments but <matrix> may be specified by keyword.  
PUSH,var [,var...]
 Pushes the specified variables (32 or less of them) onto the variables
 stack.
 See also: "POP, DUMP_STACK"
QUIT[,status]
 Exits ANA with <status>
RANDOMN(dimensions[,seed])
 Returns an F*4 array with the specified <dimensions> (at most 8 of them),
 filled with pseudo-random numbers with a normal distribution
 around zero.
 See also:  "RANDOMU"
RANDOMU(dimensions[,seed])
 Returns an F*4 array with the specified <dimensions> (at most 8 of them),
 filled with pseudo-random numbers with a uniform distribution
 around zero.
 See also:  "RANDOMN"
READ,var [,...]
 Reads values for up to 32 variables from the keyboard or batch file.
 Arrays and strings must be predefined.
 See also: "READARR, READF, PRINT, OPENR"
READARR,array
 Reads array elements from the keyboard until they are exhausted and
 returns the read elements in an array.  I.e.: the number of array elements
 need not be specified in advance.
 See also: "READ, PRINT"
READF,lun,var [,...] or READF(lun,var [,...]) 
 Reads values for up to 31 variables from the formatted file which is open 
 on logical unit <lun>.  Arrays and strings must be predefined.  The 
 function form returns 1 when successful, 0 otherwise.
 See also: "PRINTF, READ, READU" 
READU,lun,var [,...] or READU(lun,var [,...])
 Reads values for up to 31 variables from the unformatted file which is 
 open on logical unit <lun>.  Arrays and strings must be predefined.
 The function form returns 1 when successful, 0 otherwise.
 See also: "READ, READU"
RECORD,file
 Starts appending typed command lines to file <file>.
 The default file name is "record.ana".  "RECORD,''" stops recording.
REDIM,array,dimensions [,...]   or REDIM(array,dimensions [,...])
 Changes the dimensions of <array> to the indicated <dimension>.  Up to 8
 dimensions may be specified in a list or in an array.  The function form
 returns the modified <array>.
 See also: "DIMEN, NUM_DIM"
REGRID(image,xgrid,ygrid,dx,dy)
 Maps <image> onto a specified grid and returns mapped image.
 Uses nearest neighbor approximation.
.lll=
. <xgrid, ygrid> = coordinates of grid points of mapped image in coordinate
            system of old image
. <dx, dy>       = grid point spacing in mapped image coordinates (integer)
..
 See also: "COMPRESS, EXPAND, REGRID3"
REGRID3(image,xgrid,ygrid,dx,dy)
 Maps <image> onto a specified grid and returns mapped image.
 Uses bicubic interpolation.
.lll=
. <xgrid, ygrid> = coordinates of grid points of mapped image in coordinate
            system of old image
. <dx, dy>       = grid point spacing in mapped image coordinates (integer)
..
 See also: "COMPRESS, REGRID, EXPAND"
RESTORE,x,file [,header]
 Restores data from a file.
 Alias: " FZREAD"
RESUME
 This command negates the effect of exactly one "IGNORE".  If as many
 "RESUME"s have been issued as "IGNORES", then subsequent code is again
 interpreted.  If code is not currently ignored, then "RESUME" has no
 effect.
 See also: "IGNORE"
RETURN [,value]
 May only be used inside a user-defined subroutine or function.
 Forces exit from the routine.  In case of a user function, <value> is
 the value of the function call in the expression in which it is used.
REVERSE(array [,dimensions...])
 Reverses <array> in up to 8 <dimensions>, or completely.
REWIND,lun     or REWIND(lun)
 Rewinds the file which is open on logical unit <lun>.
 The function form returns 1 when successful, 0 otherwise.
 See also: "OPENR, OPENW, OPENU, CLOSE"
RFIX(x)
 Returns a long (I*4) version of x, entering the nearest integer neighbor
 to each element of x.
RUNCUM(x, [[,axis],width])
 Running sum (cumulative) of x with fixed window size.
.
. <width> - number of bins to be totalled over (defaults to 3)
           both even and odd widths are allowed.
. <axis>  - the axis along which summing is performed.  If not specified,
           or negative, then x is regarded as a 1D array.
. "!SMOOTH_EDGE" - determines handling at the edges:
           0: fill with nearest full-width value
           1: fill with nearest smaller-width value
..
 See also: "RUNSUM, SMOOTH"
RUNSUM(x [,axis,order])
 Running sum of x.  By default without regard to array
 structure.  If <axis> is defined, then summing is done along the indicated
 dimension, starting over for each of the other dimensions.
 <order> is the order of the summation, akin to n-th order differencing,
 and defaults to 1.
 The result has the same type as x, except when <axis> isn't specified or
 is negative, in which case summing is done in floating point.
 Reverse: " DIFFER"
 See also: " SMOOTH, RUNCUM"
S[,symbol]
 Show info on given symbol.
 <symbol> = either a partial symbol name (string) or a symbol number.
            All symbols that contain the partial name are displayed.
 By default, info on all symbols is displayed.
 See also: "DUMP"
SC,x,s,c
 Sine-cosine transform of x.
 <s, c> = sine and cosine coefficients; element I corresponds to wavelength
          "NUM_ELEM(X)/I".
 x is truncated to even length.
 For quickest calculation, the length of x should only contain small
 prime factors, like 2, 3, and 5.
 Reverse:  "SCB"
SCALE(image [,low,high])
 Returns an linearly intensity-scaled byte version of <image>.  <low> and
 smaller numbers are mapped to "!SCALEMIN," <high> and greater numbers
 to "!SCALEMAX."  <low> and <high> default to the lowest and highest
 values in <image>.
SCB,x,s,c
 Sine-cosine back-transform to x.
 <s, c> = sine and cosine coefficients: element I corresponds to wavelength
          NUM_ELEM(X)/I.
 Reverse: "SC"
SDEV(x [,mode])
 Sample standard deviation of (F or D version of) x;  if <mode> is
 a nonnegative scalar, then returns standard deviations of all samples
 along the specified axis.  If <mode> is an array, then its members
 indicate the class that each corresponding element of x belongs to
 and a standard deviation is returned for each class.  The first
 element of the result contains the standard deviation of class 0, or
 of the most negative class.
 See also:  "MEAN, TOTAL"
SETBACKGROUND,port,color
 Sets the background color of window <port> to <color> (string).
SETENV,string
 Sets or displays an environment variable.
 If <string> does not contain an equal sign "=", then it is taken to
 be a file or directory name and the current expansion of that name is
 displayed.  If <string> has the form "'VARIABLE=VALUE'", then "VALUE"
 is filename expanded and environment variable "VARIABLE" (which must
 not have a preceding "$") is set to the expanded "VALUE".
SETFOREGROUND,port,color
 Sets the foreground color of window <port> to <color> (string).
SHOW[,symbol]
 Show info on given symbol.
 <symbol> = either a partial symbol name (string) or a symbol number.
            All symbols that contain the partial name are displayed.
 By default, info on all symbols is displayed.
 See also: "DUMP, ATOMIZE"
SHOWSTATS
 Displays some statistics about the current ANA session; mostly
 helpful in debugging the ANA program.  Depending on the way in which
 your version of ANA was compiled, the following columns may feature
 in the answer:  NALL - number of items for which memory is allocated,
 TALLOC - total number of allocated bytes, IGN - ignore level
 (IGNORE-RESUME), STD - statement depth, EXL - execution level, CPL -
 compilation level, LST - list stack index, NVAR - number of defined
 named variables, TVAR - index of temporary variables, EXE - index
 of permanent executables, TEXE - index of temporary executables,
 SSTK - symbol stack index, CTXT - current execution context.

 At the main execution level, the values of IGN, STD, EXL, CPL, LST,
 SSTK, and CTXT should all be zero.  TVAR ought to be equal to its
 default value (5000 in Utrecht) and TEXE to its default value plus
 one (20001 in Utrecht).
 See also: "DEBUGGING"
SHOW_F [,match_string]
 Lists all internal ANA functions whose names contain <match_string>.
 Alias: " SHOW_FUNC, SHOW_FUN"
 See also: "DUMP_FUNC, SHOW_SUBR"
SHOW_FUN [,match_string]
 Lists all internal ANA functions whose names contain <match_string>.
 Alias: " SHOW_F, SHOW_FUNC"
 See also: "DUMP_FUNC, SHOW_SUBR"
SHOW_FUNC [,match_string]
 Lists all internal ANA functions whose names contain <match_string>.
 Alias: " SHOW_F, SHOW_FUN"
 See also: "DUMP_FUNC, SHOW_SUBR"
SHOW_S [,match_string]
 Lists all internal ANA subroutines whose names contain <match_string>.
 Alias: " SHOW_SUBR, SHOW_SUB"
 See also: "DUMP_SUBR, SHOW_FUNC"
SHOW_SUB [,match_string]
 Lists all internal ANA subroutines whose names contain <match_string>.
 Alias: " SHOW_S, SHOW_SUBR"
 See also: "DUMP_SUBR, SHOW_FUNC"
SHOW_SUBR [,match_string]
 Lists all internal ANA subroutines whose names contain <match_string>.
 Alias: " SHOW_S, SHOW_SUB"
 See also: "DUMP_SUBR, SHOW_FUNC"
SIEVE([array,] condition)
 Returns a 1D array containing those <array> elements for which the
 corresponding element of <condition> is non-zero.  If <condition>
 isn't met anywhere, then scalar -1 is returned.  If <array> is not
 specified, then the indices of non-zero elements of <condition> are
 returned (i.e. equivalent to "SIEVE(INDGEN(LONG(condition)),condition)" )
SIN(x)
 Returns the sine of x (which is in radians).  Multiply by "#D.R" to
 get radians from degrees.
 Reverse: "ASIN"
SINH(x)
 Returns the hyperbolic sine of x.
 See also: "COSH, TANH"
SKIPC(s1,s2)
 Returns part of string <s1> starting with the first character which is not
 in string <s2>, or an empty string
SKIPF,lun,nfiles
 Skip <nfiles> files on logical unit (tape drive) <lun>.  The tape
 is positioned just after the <nfiles>th read end-of-file.  Negative
 <nfiles> cause the tape to be rewound |<nfiles>| files.
 See also: " SKIPR, REWIND, TAPRD, TAPWRT"
SKIPR,lun,nrecs
 Skip <nrecs> records on logical unit (tape drive) <lun>.  Negative
 values of <nrecs> cause the tape to be rewound |<nrecs>| records.
 See also: " SKIPF, REWIND, TAPRD, TAPWRT"
SMAP(x [,/TRUNCATE])
 Maps x byte for byte into a string.  If keyword "/TRUNCATE" is set,
 then truncates the string at the first zero byte.
 See also: "BMAP, WMAP, LMAP, FMAP, DMAP"
SMOOTH(x, [[,axis],width])
 Boxcar smoothing of x.
.ll-
. <width> - number of bins to be smoothed/totalled over (defaults to 3)
           both even and odd widths are allowed.
. <axis>  - the axis along which summing is performed.  If not specified,
           or negative, then x is regarded as a 1D array.
. "!SMOOTH_EDGE" - determines handling at the edges:
           0: fill with nearest full-width value
           1: fill with nearest smaller-width value
..
 See also: "RUNCUM, RUNSUM, GSMOOTH, ESMOOTH"
SORT(array[,/HEAP,/SHELL,/AXIS])
 Sorts the elements of <array> in ascending order, either along the 0th
 dimension (if keyword "/AXIS" (4) is specified), or as one big
 1-dimensional array.  Uses heap sort if keyword
 "/HEAP" (1) is used, or shell sort if keyword "/SHELL" (2) is used,
 or otherwise the default method, which is heap sort if "!SORT_FLAG"
 equals 0, or shell sort if it equals 1.
 See also: "INDEX"
SPAWN,string [,/SILENT]  or  SPAWN(string)
 Spawns <string> as a shell command.  If called as a function, then
 the return value from the shell is returned.  If called as a routine,
 then if keword "/SILENT" (1) is specified, then "SPAWN" never
 generates an error.  If "/SILENT" is not set (the default), then an
 error is generated if the shell command returns a non-zero value. 
SQRT(x)
 Returns the square root of x.
 See also: "CBRT"
STEP [,mode]
 Enables stepping through ANA statements.  Just before excution of
 each selected statement, it and a prompt ">executeLevel?" are displayed.
 Strike a key to continue.  The following keys have some side effect:
.ll-
.  s -  step to the next command at any level
.  n -  step to the next command at the current level
.  e -  step to the next command at a lower level
.  z -  enter and execute a command at the "dbg>" prompt
.  x -  discontinue stepping
.  q -  abort the current calculation
.  ? -  describe all keys with a side effect
..
 See also: "TRACE"
STORE,x,file [,header]
 Stores data x and header string <header> in the file with name <file>.
 Alias: " FZWRITE"
STR,1,1,ana_string,
STRARR(dimensions)
 Return a string array of the specified <dimensions>.  The latter may be an
 array containing the dimensions, or a list of up to 10 dimensions.
 See also: "ARRAY, BYTARR, INTARR, FLTARR, DBLARR, LONARR"
STRCOUNT,2,2,ana_strcount,
STRETCH(image,grid)
 Stretches 2D <image> according to the displacement grid <grid> as
 derived with "GRIDMATCH".
STRING([format,] argument1 [,argument2 ...])
 Transforms the arguments to an ASCII string under guidance of format
 string <format>.  If the argument is not of the data type expected by
 the format, then the appropriate standard format ("!FORMAT_I, !FORMAT_F"
 or "!FORMAT_S") is used.
 See also: "FSTRING, PRINT, PRINTING FORMATS"
STRLEN(s)
 Length of string <s>
STRLOC(s1,s2)
 Returns part of string <s1> which starts with substring <s2>,
 or an empty string
STRPOS(s1,s2)
 Returns position of substring <s2> in <s1>, if found.  -1 otherwise
STRR(s1,s2,s3 [,n])
 Replace up to <n> occurrences of substring <s2> in string <s1>
 with substring <s3>.  <n> defaults to infinity.
STRREPLACE(s1,s2,s3 [,n])
 Replace up to <n> occurrences of substring <s2> in string <s1>
 with substring <s3>.  <n> defaults to infinity.
STRSKP(s1,s2)
 Returns part of string <s1> which starts after substring <s2>, or
 an empty string
STRSUB(s1,i1,i2)
 Returns part of string <s1> consisting of elements <i1> through <i2>
STRTRIM(s)
 Returns string <s> with trailing blanks, tabs, and control characters removed
SUN_B,2,2,ana_sun_b,
SUN_D,2,2,ana_sun_d,
SUN_P,2,2,ana_sun_p,
SUN_R,2,2,ana_sun_r,
SWAB,x
 Swap bytes in x
SWAPB,x
 Swap bytes in x
SWITCH(var1,var2)
 Switch the identity of <var1> and <var2>.
SYMCLASS,symbol
 Returns the symbol class of <symbol>.
  Possible values:
.ll-
.    1 - scalar
.    2 - string
.    4 - array
.  255 - undefined
..
 See also: "SYMDTYPE, DIMEN, NUM_ELEM, NUM_DIM, ISARRAY, ISSCALAR,
 ISSTRING, CLASSNAME"
SYMDTYPE,symbol
 Returns the data type of <symbol>.
  Possible values:
.rll-
.    0 - byte    - I*1
.    1 - word    - I*2
.    2 - long    - I*4
.    3 - float   - F*4
.    4 - double  - F*8
.   15 - string
.  255 - undefined
..
 See also: "SYMCLASS, DIMEN, NUM_ELEM, NUM_DIM, TYPENAME"
T,x [,...]
 Displays the contents of symbol x on the screen.  Up to 32 arguments
 may be specified.
 See also: "PRINTF"
TAB[,column]
 Returns a blank string just big enough to span from the current cursor
 column to the indicated <column>, or to the next tab position (8-column).
TABLE(x,y[,index],xnew[,/MATCH]) or TABLE(x,y[,mode],xnew[,/TABLE])
 Interpolates in table <y> vs. x at positions <xnew>.
 The first dimension of x and <y> specifies the entries in one
 interpolation table.  Second and higher dimensions of x and <y>
 specify individual interpolation tables.
 x must be in ascending order in its first dimension.  Dimensions
 shared between x and <y> must be equal.  Variables are repeated
 to accomodate omitted dimensions.
   If array <index> is specified, then each element of <index>
 determines which interpolation table to use for the corresponding
 element of <xnew>.  In that case, dimensions shared between <index>
 and <xnew> must be equal.  Non-integer elements of <index> are
 treated with linear interpolation.  The result gets all dimensions
 of the largest of <xnew> and <index>.
   If scalar <mode> is specified, then this indicates whether to
 interpolate all elements of <xnew> in each interpolation table <y>
 vs. x (<mode> = 0, keyword "/NOMATCH"), or to select each next
 interpolation table for each next 1D vector from <xnew> (<mode> = 1,
 keyword "/MATCH").  In the former case, the result has all the
 dimensions of <xnew> and all dimensions of the largest of x and
 <y>.  In the latter case, all dimensions shared between x, <y>, and
 <xnew>, except the first one, must be equal.  The result gets the
 first dimension of <xnew>, and all higher dimensions of the largest
 of x, <y>, and <xnew>.  By default, <mode> = 0.
   "TABLE" is more general than "POLATE", and "POLATE" seems to have
 a small bug.
TAN(x)
 Returns the tangent of x (which is in radians).  Multiply by "#D.R" to
 get radians from degrees.
TANH(x)
 Returns the hyperbolic tangent of x.
 See also: "COSH, SINH"
TAPEBUFIN,2,5, ana_tapebufin,
TAPEBUFOUT,2,5, ana_tapebufout,
TAPE_STATUS,0,0, ana_tape_status,
TAPRD,2,2, ana_taprd,
TAPWRT,2,2, ana_tapwrt,
TENSE(x,y,xf [,tension][,dlsh,drsh])
 Interpolate functions using splines under tension.
.lll=
. <x, y>  = input points. x must be in ascending order.  (1D arrays)
. <xf> = x coordinates at which to interpolate (1D array)
. <tension> = amount of tension (default: 1), must not be less than 1.
. <dlsh, drhs> = first derivatives at left hand and right hand side.
..
 Returns y-values for the specified <xf>
TENSE_CURVE(x,y,xf [,tension][,dlhs,drhs])
 Interpolate arbitrary open curves using splines under tension.
.lll=
. <x, y>  = input points.  (1D arrays)
. <xf> = independent variable at which to interpolate, in units of curve
        length (ranging from 0 to 1) (1D array)
. <tension> = amount of tension (default: 1), must not be less than 1.
. <dlsh, drhs> = first derivatives at left hand and right hand side.
..
 Returns a 2D array containing x and y coordinates of interpolated points.
TENSE_LOOP(x,y,xf [,tension])
 Interpolate closed curves using splines under tension.
.lll=
. <x, y>  = input points. (1D arrays)
. <y>  = input dependent variable (1D array)
. <xf> = independent variable at which to interpolate, in units of curve
        length (ranging from 0 to 1) (1D array)
. <tension> = amount of tension (default: 1), must not be less than 1.
..
 Returns a 2D array containing x and y coordinates of interpolated points.
TIFFWRITE(array,file,[ppi])
 Writes a very plain tiff file, 8 bit deep. The input data in "array" must be a
 2-D byte array, "file" is a string containing the output file name, and
 "ppi" is an optional parameter which is the pixels per inch to use for
 programs that look at this tag in the TIFF file (most do). The default
 value is 100 pixels/inch. This value does not affect the image data stored
 in the file. There is no color associated with the image.
 See also: "TIFFCOLORWRITE"
TIFFCOLORWRITE(array,colormap,file,[ppi])
 This differs from TIFFWRITE only in the addition of a color map to support
 simple color images. Only 256 colors are supported, it does not write
 full color images. The input "array" must be a 2-D byte array representing
 the image. The colormap must be a (256,3) array containing the color
 translations. It is internally converted to a 16 bit value and the range
 of colormap values is 0 to 2^16-1. The output file name is in the string
 "file" and "ppi" is an optional parameter which is the pixels per inch to use
 for programs that look at this tag in the TIFF file (most do). The default
 value is 100 pixels/inch.
TOTAL(x [,mode])
 Returns the F*4 total of all values in x.  If <mode> is defined and
 a nonnegative scalar, then totals are determined along the axis
 indicated by <mode>, i.e. "TOTAL(X,0)" yields "X(+,*,*,...)", and
 "TOTAL(X,1)" yields "X(*,+,*,*,...)".  If <mode> is an array of the
 same size as x, then its elements indicate the class of each
 element of x and the total for each class is determined.
 See also:  "MEAN, DSUM"
TRACE[,level][,/FILE,/LOOP,/BLOCK,/ROUTINE,/ALL,/SHOWSTATS]
 Starts tracing.  When tracing, all statements are displayed just before
 execution, complete with line numbers in the innermost enclosing
 subroutine, function, block routine, or file, if the current execution
 level is lower than or equal to the current threshold level.
 Mode keywords:  FILE (1) - to trace into @files;  LOOP (2) - to trace
 into any kind of loop (repeat, for, etc.);  BLOCK (4) - to trace into
 {}-delimited blocks;  ROUTINE (8) - to trace into user-defined
 subroutines, functions, and block routines;  ALL (16) - to trace
 all of the above;  SHOWSTATS (32) - to also display SHOWSTATS information.
 See also: "ATOMIZE, DUMP, SHOW, ECHO, STEP, SHOWSTATS"
 Reverse: " NOTRACE"
TREND(x [,n])
 <n>th-degree polynomial fit to x
TV,image [,x,y,window]
 Display <image> at position <x,y> (default: 0,0) in specified <window> 
 (default: last accessed one, or 0), using the ANA colormap.
 See also: " XSETACTION, XCOPY, XTVREAD, TVRAW, TVMAP"
 Alias: "    XTV"
TVLAB,3,4, ana_xlabel,
TVLCT,r,g,b
 Loads a color table specified in 1D <r>, <g>, and <b>, and scale them to
 the available range.
TVMAP,image [,x,y,window]
 For a non-BYTE array, equal to "TV".  For a "BYTE" array, the array values
 are taken to be color indices, are mapped to the appropriate pixel values
 and displayed.
 See also: " TV, TVRAW, COLORPIXEL"
TVRAW,image [,x,y,window]
 For a non-"BYTE" array, equal to "TV".  For a "BYTE" array, the array values
 are taken to be pixel values and they are dumped raw on the screen,
 and may therefore be displayed with unexpected colors
 See also: " TV, TVMAP"
TY,x [,...]
 Displays the contents of symbol x on the screen.  Up to 32 arguments
 may be specified.
 See also: "PRINTF"
 Alias: "PRINT, T, TYPE"
TYPE,x [,...]
 Displays the contents of symbol x on the screen.  Up to 32 arguments
 may be specified.
 See also: "PRINTF"
 Alias: "PRINT, T, TY"
TYPENAME(type)
 Returns the name of the numerical type.
 See also:  "SYMDTYPE, CLASSNAME"
ULIB[,string]
 Sets or displays the user library directory name.
UNLOAD,1,1, ana_unload,
UPCASE,s
 All letters in <s> converted to upper case.
 Alias: "UPPER"
 See also: "LOWER, LOWCASE"
UPPER,s
 All letters in <s> converted to upper case.
 Alias: "UPCASE"
 See also: "LOWER, LOWCASE"
VARNAME(symbol)
 Returns the name of the symbol.
VOIGT(a,v)
 Returns the Voigt function (usually written as H(a,v)) for
 Voigt parameter <a> (a scalar, 0 selects a Doppler profile)
 and distance to line center <v> (an array, in units of the
 Doppler width).
WAIT,x
 Wait x seconds (integer)
WAIT_FOR_MENU [,erase_prev]
 Waits for a user menu selection.  If <erase_prev> is set and unequal to
 zero, then pending menu selections are discarded before waiting for
 a new user selection.  Otherwise, the oldest pending selection is returned.
 The results are stored in "!LAST_MENU" (menu number), "!MENU_ITEM"
 (item number in menu, starting at zero), "!XMENU", "!YMENU" (position
 of mouse relative to menu's upper left hand corner;  in units of one
 character or one line, respectively)
 See also: " MENU, CHECKMENU"
WAIT_FOR_TAPE,1,1, ana_wait_for_tape,
WATERSHED(image [,mode])
 Interprets <image> as an altitude map and determines to which nearest
 neighbor water at each pixel location would flow.  If <mode> is zero,
 then the index offset of the visited neighbor is returned for each
 element.  If <mode> is non-zero, then the index of the sink where the
 water would finally collect is returned for each element.
WEOF,n
 Write "n" end-of-file markers on tape.
WINDOW[,xmin,xmax,ymin,ymax,zmin,zmax]
 Sets or displays the plot window (cube) boundaries.  The next plot is
 drawn between the specified device-independent coordinates.  If no
 arguments are specified, then the current settings are displayed.
 The values of "!WXB, !WXT, !WYB, !WYT, !WZB," and "!WZT" are modified
 accordingly. 
 See also:  "!WXB, !WXT, !WYB, !WYT, !WZB, !WZT"
WMAP(x)
 Maps x byte for byte into a word (I*2) array.
 See also: "SMAP, BMAP, LMAP, FMAP, DMAP"
WORD(x)
 Return a word (I*2) version of x.
 See also: "BYTE, LONG, FLOAT, DOUBLE"
 Alias: "INT"
WRITEU,lun,x [,...]
 Writes the contents of symbol x to the file opened on logical unit
 <lun>.  Up to 32 arguments may be specified.
 See also: "READU, PRINTF"
XCOPY,window1,window2 [,x1,y1,width,height,x2,y2]
 Copies the part of the contents of <window1> starting at position <x1,y1>
 (default: 0,0) and with <width> and <height>, to <window2> at position
 <x2,y2>.  Positive window numbers denote windows, negatives pixmaps.
 See also: "XTV, XSETACTION"
XDELETE,port
 Deletes window <port>.
 See also: "XPORT"
XEVENT
 Gives information on the events currently in the event buffer.
XFLUSH
 Forces execution of all pending X display commands.
 See also: "XPURGE"
XFONT,fontname [,window]
 Selects font <fontname> for <window> (which defaults to the last accessed
 one, or 0).
 See also: "XPORT"
XLABEL,label,x,y [,window]
 Puts <label> (a string) at position <x,y> in <window> (which defaults
 to the last accessed one, or 0).
 See also: "XPORT, XTV"
XLABELWIDTH(label)
 Returns the width of the <label>.
 See also: "XLABEL"
XLOOP
 Window event manager.  Waits for an event of the kind registered with
 "XREGISTER" and returns the event type in "!EVENT_TYPE".  The event itself
 is kept in the queue and can be read with type-specific routines.
 If no specific event source was selected with "XREGISTER", then all
 sources (ANA and non-ANA) are monitored.  If no specific event type was
 selected, then "XLOOP" returns immediately.
 See also: "XREGISTER, WAIT_FOR_MENU, XQUERY, CHECKMENU, XPLACE,
 !NOEVENTFLUSH"
XOPEN,display
 Sets display that image output will be sent to, to <display> (a string).
 See also: "XPORT"
XPLACE
 Waits for a key or button press in an ANA window.  Returns pressed button
 in "!BUTTON", mouse coordinates in "!IX" and "!IY", and the pressed
 key in "!KEY".
 See also: "XPORT, XQUERY"
XPORT [port, width, height, xpos, ypos, windowtitle, icontitle]
 Creates a window.  Number <port> (default: 0, or last accessed one);
 dimensions <width>, <height>, position <xpos>, <ypos>.  If
 <windowtitle> or <icontitle> start with a minus sign, then no "ana
 xx:" string is prepended.  If the window already exist and no sizes
 were specified, or the specified sizes are equal to the window's
 sizes, then the window is merely selected, not recreated.
XPURGE
 Executes all pending X display commands and empties the X event buffer.
 See also: "XFLUSH"
XQUERY [,window]   or  XQUERY([window])
 Returns the current mouse position and system time:
.ll-
. "!IX,!IY"         -  mouse position relative to current window
. "!ROOT_X,!ROOT_Y" -  mouse position relative to root window
. "!XTIME"          -  system time
..
 The function form returns 1 if the mouse was in the specified window,
 0 otherwise.  <window> defaults to the last window ("!WINDOW").
 See also: "XPORT, XPLACE"
XREGISTER[,event,WINDOW,MENU][,/KEYPRESS,/BUTTONPRESS,/BUTTONRELEASE,/POINTERMOTION,
 "/ENTERWINDOW,/LEAVEWINDOW,/ALLEVENTS,/ALLWINDOWS,/ALLMENUS,/DESELECT]"
 Register, unregister, or display registered event types for "XLOOP".
 Without arguments, "XREGISTER" displays currently registered event types.
 With a positive argument, the corresponding event types are registered.
 With a negative argument, the corresponding event types are unregistered.
 With a zero argument, all event types are unregistered.
 <event> has bits to set for each event type and also has bits to
 select specific event sources:
.rrl-
.   Bit - Or-Mask - Event Type
.     0 -     1   - Press of a key
.     1 -     2   - Release of a key
.     2 -     4   - Press of a mouse button
.     3 -     8   - Release of a mouse button
.     4 -    16   - Motion of the mouse pointer
.     5 -    32   - Mouse enters a window or menu item
.     6 -    64   - Mouse leaves a window or menu item
.     7 -   128   - Select ANA windows for event source
.     8 -   256   - Select ANA menus for event sources
..
 See also: " XLOOP"
XSETACTION [,action,window]
 Sets the copy action associated with the window.  <window> defaults
 to "!WINDOW", <action> to 1.  Choose <action> from:
.rll-
. code - name           -  destination (dst) is replaced with
.   0  - copy           -  src
.   1  - xor            -  src xor dst
.   2  - equiv          -  not(src) xor dst
.   3  - clear          -  0
.   4  - set            -  1
.   5  - noop           -  dst
.   6  - invert         -  not(dst)
.   7  - and            -  src and dst
.   8  - and_reverse    -  src and not(dst)
.   9  - and_inverted   -  not(src) and dst
.  10  - or             -  src or dst
.  11  - or_reverse     -  src or not(dst)
.  12  - or_inverted    -  not(src) or dst
.  13  - nand           -  not(src) or not(dst)
.  14  - nor            -  not(src) and not(dst)
.  15  - copy_inverted  -  not(src)
..
 Note that actions 1 and 2 are reversible.  These are suitable for display
 of sprites:  objects that change the image but may be removed without
 a trace by drawing them again.  I use them for simultaneous display of
 multiple pseudo-cursors.
 See also: " XTV"
XTV,image [,x,y,window]
 Display <image> at position <x,y> (default: 0,0) in specified <window> 
 (default: last accessed one, or 0).
 See also: " XACTION, XPORT"
 Alias: "    TV"
XTVREAD([ix,iy,nx,ny] [,window])
 Returns part of window <window>, starting at <ix, iy> and of dimensions
 <nx, ny>.  The window part defaults to the whole window, <window> to the
 last selected window.
 See also: " XPORT, XTV"
XYMOV,x,y[,mode][,/DEV,/DVI,/IMG,/PLT,/RIM,/RPL,/BOUNDINGBOX,/MOVEFIRST]
 Draws a line to coordinates (x,<y>) with plot mode <mode>.  All
 three arguments may be arrays.  Coordinate system keywords are
 available.  </BOUNDINGBOX> sets bounding box updating.  </MOVEFIRST>
 forces a move (i.e. not draw) to the first point, and is selected by
 default. 
 See also:  "COORDINATE SYSTEMS, !BB_UPDATE"
ZAP,var [,var2,...]
 Completely removes the listed variables (except for function pointers).  
 Afterward, it is as if they never existed.  If a (function) pointer is
 "ZAP"ped, then the corresponding variable, subroutine, function, or
 blockroutine, and all associated variables and executable code, get
 zapped.  To zap a (function) pointer itself, add the "/POINTER" key
 to the "ZAP" call.
ZERO,var [,var2,...]  or  ZERO(var)
.ll-
. Subroutine -  Fills specified variables with zeroes (up to 32 variables may
              be specified).
. Function   -  Returns zeroed copy of <var>.
..
ZINV(var)
 Zero-protected inverse: returns "1/x" for all nonzero elements "x" of <var>,
 and zero for all zero elements.
[x [,...]]
 Concatenation function.  Calls:
.ll-
. "Z = [x]"       - x is returned with one more dimensions (equal to 1)
. "Z = [x,y,...]" - x and <y> and up to 30 more scalars or arrays are
                    concatenated and the result is returned.
..
 All but one of the arguments may be undefined.
 Alias: "CONCAT"
ZZZZZZZ last line of file