?

   F U D G I T   Version 2.33

 The three different modes are accessed by commands: 'fmode', 'pmode'
 and 'cmode'. All fmode commands can be abbreviated down to 2
 characters. Vectors have upper case name and scalar variables lower
 case. A mix of both upper and lower cases results in forming a string
 variable. Read the "Intro" help topic for an overview. See the
 "README" help item for complete copyrights.

 Send bugs or comments to <isaac@physics.mcgill.ca>.

?Intro

 FUDGIT is a double-precision multi purpose fitting program.
 It can manipulate complete columns of numbers in the form
 of vector arithmetic. FUDGIT is also an expression language
 interpreter understanding most of C grammar except pointers.
 It supports all functions from the C math library. Finally,
 FUDGIT is a front end for any plotting program supporting
 commands from stdin. It is a nice mathematical complement
 to GNUPLOT, for example.

 The main features of FUDGIT are:
    - Command shell including history;
    - Possible abbreviation of all the ``fitting mode'' commands;
    - Possible plural when it makes sense too;
    - Interactive shell supporting flow control (while,
      if-else-endif, foreach);
    - User definable macros;
    - User definable aliases;
    - On-line help;
    - On-line loadable procedure- or function-objects;
    - On-line selectable plotting program;
    - Fourier transforms;
    - Smoothing;
    - Double-precision built-in calculator;
    - Built-in interpreter supporting most of C language including
      flow control (if, else, while, for, break, continue);
    - User definable functions and procedures;
    - Double-precision vector arithmetic;
    - Access to the complete C math library;
    - Built-in fitting series such as:
        + power series (polynomial);
        + sine series;
        + cosine series;
        + Legendre polynomials;
        + series of Gaussians;
        + series of exponentials;
    - User definable fitting functions;
    - Totally dynamical allocation of variables and parameters;
    - Possible selection of fitting ranges;

 FUDGIT has a collection of fitting routines including:
    - straight line (linear) least squares;
    - straight line (linear) least absolute deviation;
    - general linear least squares using QR decomposition;
    - general linear least squares using singular value decomposition;
    - nonlinear Marquardt-Levenberg method;

 Refer to the ``User's Manual'' for a complete description and a
 tutorial on I/O and fitting.

 See also:
      Modes, C, cmode, fmode, pmode, fit, set, read, save, let

?Modes

 FUDGIT is composed of three different modes. These modes can be
 thought of as a C-shell like interpreter linked with a calculator,
 sharing the same variables in memory, and with a plotting program of
 our choice.
 
 The C-shell like interpreter is called the ``fitting mode''. It is
 the central mode and is the one from which all accesses to the disk
 are done. This mode has a range of commands allowing the user to read
 vectors from or save vectors to a data file, to read a command script,
 save the command history, do a Fourier transform of a vector, make a
 linear or nonlinear least square fit, etc... This mode also allows the
 user to define macros and aliases, and to perform plotting-fitting
 batch processes by using some of the built-in flow control commands
 (while, foreach, if-else-endif). All the commands in the fitting mode
 can be abbreviated. It is worth mentioning that in the fitting mode
 the command line parsing is done by analyzing words separated by one
 or more blanks (space or tab), as in an interactive csh.

 The ``C-calculator mode'' is a language interpreter supporting most of
 C grammar except pointers. It also supports the complete
 double-precision C math library. Thus, recognized keywords cannot be
 abbreviated, and the different tokens need not be separated. Most of
 the C operators and keywords are understood and a few extra operators
 have been added. This mode does essentially all the possible
 calculations on variables or vectors. Functions and procedures can be
 defined. String variables, string comparison, addition, subtraction
 are also supported by C-calculator mode. This mode is accessed by the
 command `cmode'.

 Finally, the ``plotting mode'' is a channel talking directly to the
 plotting program of your choice. Therefore, FUDGIT can serve as a
 front end to any plotting program able to accept input from stdin.
 This way, vectors can be build from the calculator and then plotted by
 your favorite plotting program. The default plotting program is
 GNUPLOT.

?&

 The `&' operator forces FUDGIT to use the built-in following fitting
 mode command and to ignore any existing macro or alias with the same
 name. This can be useful in constructions like:

    macro cd 1
        pmode cd "$1"
        &cd $1                  # The built-in cd
    stop

 See also:
     macro, cd

?\

 If anywhere in the middle of a line, a `\' will indicate
 FUDGIT to take the following character as is. If at the end of a
 line, a `\' indicates that the present line continues on the
 following one, and thus to ignore the following carriage return.

 See also:
     line editing

?!

 Any line beginning with the so-called bang operator `!' will execute
 the system command line with a Bourne shell. Aliased commands as found
 in your interactive C-shell do not hold any more. For example,
 commands like `!rm' will not be interactive (i.e. /bin/rm -i) even
 if you have such an alias in your ".cshrc" file. Be careful!  A
 nice turnaround is to alias rm to ``! rm -i'' in your 
 ".fudgitrc" file and to use the `rm' command directly from
 FUDGIT's shell.

 When used in a macro name or an alias name, the `!' character has
 still another meaning. This tells the parser that characters following
 the `!' are optional. Therefore, if one types the following,
 interactively, (see NOTE)

     set noexpand
     alias da!te !date
     set expand

 then the parser will recognize `da', `dat' and `date' as
 all synonymous to the system command `! date' run through a Bourne
 shell.

 NOTE: In interactive mode, the history functions will try to interpret
 a history substitution if the `!' is not followed by a space. See the
 appendices. To avoid that the line be scanned for a history event
 designator, use the `set noexpand' command. In some cases, it
 might be simpler to use the `system' command.

 Syntax:
     !"command"

 Example:
     ! mail

 See also:
     alias, ls, vi, foreach, system, set expand

?help?

 A question mark will indicate FUDGIT to try to get the possible
 options available to the command presently typed. This kind of help is
 context sensitive and works when an insufficient number of arguments
 is supplied. The question mark also serves as a wild character in
 string subtraction.

 Syntax:
     "command" ?

 Examples:
     ?
     show ?
     set function ?

 See also:
     help, strings

?$

 The `$' operator expands scalar variables or constants (double
 precision numbers from C-calculator mode lookup table) as well as
 string variables or constants. Existing scalar variables can thus be
 expanded as a string in order to serve as a file name or directory
 name, for example. The expansion is done according to the value given
 to the `set vformat' command which initially defaults to
 ``%.3g''. Using the scalar variable expansion operator in
 C-calculator mode is not recommanded since a lot of precision might be
 lost (actually it is a waste!). Scalar variable expansion is
 essentially provided to allow alternative procedures in certain cases,
 such as generating filenames from numbers. Math function `scan'
 can be considered as the complement of scalar variable expansion.

 The `$' character also expands string variables. Expansion is done by
 replacing the $"String-Variable-Name" by the value of the string
 variable. This can be used to replace `scan' in cases where the
 string variable or constant represents a number. For example

     foreach File in echo 2.2 4.4 6.7 8.32
         let x = $File
         .
         .
         .
     end

 In both cases, if the variable name has to be followed by alphanumeric
 characters, then the variable name can be delimited by braces as in
 standard csh.

 Followed by an integer number, the `$' character serves to designate
 the arguments of a macro. Refer to the description of `macro',
 concerning this point.

 Syntax:
     $"name"

 or

     ${"name"}

 See also:
     C, cmode, macro, echo, exit

?_dumplot

 Command `_dumplot' is generally used in a macro to dump vectors
 in the plotting pipe. It is described in more detail under 
 `special' item.

?_killplot

 Command `_killplot' is rarely used. It sends a KILL signal to the
 plotting program. It is described in more detail under `special'
 item.

?adjust

 The `adjust' command is used to specify the parameters to be
 adjusted in the ``least square linear'' and the ``Marquardt-Levenberg
 nonlinear'' fitting methods. Parameters not being adjusted will have
 their standard deviation set to zero.

 Syntax:
     adjust "index-list"

 Example:
     adjust 1 2 4

 See also:
     set parameters, set method, set function, fit, show fit

?alias

 The `alias' command is used to alias a multiple word command to a
 single word. Although macros and aliases are different objects, it is
 not allowed to define a macro and an alias with the same name since
 aliases are always expanded first. Recall that the bang operator
 (`!'), at the beginning of a line is recognized from a macro, an alias
 or a script file so that an alias like

     alias date !date

 is perfectly legal. However, this would have to be typed

     alias date ! date

 at the interactive command line, to avoid that the `!' be interpreted
 by the history functions.

 When called without arguments, `alias' will list all the current
 aliases. For obvious reasons, it is not allowed to `alias' 
 "unalias". `alias' also supports the command abbreviation character
 `!'. To enter a `!' without having it interpreted by the history
 functions, just `set noexpand' for the time entering the command.
 When a `!' is part of the alias name this indicates that the alias
 command name can be abbreviated down to that point. Since the `&'
 operator is used to refer to the native commands, it is therefore
 forbidden to start an alias name by character '&'.

 Syntax:
     alias "command" "command-list"

 Examples:
     alias mv !mv
     alias . quit
     alias da!te !date

 See also:
     !, &, macro, unalias, set expand

?append

 The `append' command can be used to append various things to an
 existing file. If the file does not already exists, it will be
 automatically created.

 The `save' command can be used to save various things to a file.
 If a file with the same name already exists, it will be overwritten
 without any warning.

?append history
?save history

 History can be saved or appended to a file. Any file saved this way
 can later be executed by the `load' command. Note that `append
 history' will silently fail if the file does not exist.

 Syntax:
     append history "filename"
     save history "filename"

 See also:
     load, line editing, fmode

?append macros
?save macros

 All the current macros and aliases can be saved or appended to a file.
 Any file saved this way can be subsequently `load'ed at any time.
 To avoid confusion between data files and script files we recommand
 that you use the ".ft" extension for your script files.

 Syntax:
     append macros "filename"
     save macros "filename"

 See also:
     alias, unalias, load, show, macro, unmacro

?append parameters
?save parameters

 Parameters can be saved into a file at any time. The number output
 format will be the one chosen by the `set format' command. The
 column order will be a parameter followed by its standard deviation.
 All columns are separated by a tab. Therefore, if one has previously
 set parameters, i.e.

     set parameters MYPAR 3
      .
      .
      .
     save parameters myfile

 then there will be 6 columns as follows:

MYPAR[1]    DMYPAR[1]       . . .      MYPAR[3]     DMYPAR[3]

 in file "myfile". Most of the time, the user will desire to
 save parameters along with some variables or constants. This can be
 done by giving the variable or constant (either string or scalar)
 names on the command line. For example,

     let t = 0.23
     set parameters A 2
     .
     .
     .
     save parameters t parfile

 will create a file "parfile" containing the value of scalar
 variable `t', followed by the 2 values of parameters `A',
 alternated with the value of their standard deviations `DA'.
 Note that the given list of variables will be printed first.

 Syntax:
     append parameters "variable-list(optional)" "filename"
     save parameters "variable-list(optional)" "filename"

 See also:
     set format, set parameters, show parameters

?append variables
?save variables

 Any variable or number of variables can be saved to a file at any
 time. Vector elements referenced by an explicit index are considered
 as variables. String variables and constants are recognized as well.

 Syntax:
     append variables "variable-list" "filename" 
     save variables "variable-list" "filename"

 Examples:
     append variables x Y[3] a VECTOR[78] datafile1
     save variables t PARAM[2] DPARAM[2] datafile2

 See also:
     load, cmode, let, C, show, auto

?append vectors
?save vectors

 Any vector or number of vectors can be saved to a file. All the values
 are written in columns separated by a tab. The number format will be
 the one chosen by the `set format' command.

 Syntax:
     append vectors "VECTOR-list" "filename"
     save vectors "VECTOR-list" "filename"

 Examples:
    append vectors X Y ERROR1 TEST2 datafile1
    save vectors TIME TEMP DT datafile2

 See also:
     set format, set data, read, fit, fft, show, auto

?auto

 The `auto' keyword is used to define automatic variables. The type
 of variable can be a scalar variable, a VECTOR or a String, depending
 on the upper-lower case letters in the variable name. The scope of
 auto variables is delimited by braces as in C. All auto variables are
 stored on the stack and are freed when the scope of the variable is
 left. Definition of variables can only be done right after a brace has
 been opened. Only scalar variables can be assigned as the are defined,
 while vectors are assigned to zero, and strings are empty. Contrarily
 to C, automatic scalar variables are set to zero if not assigned.
 `auto' is a C-calculator mode keyword.

 Syntax:
     auto "var-list"

 Examples:
     # Some dummy examples
     set data 100
     cmode
        x = y = 1                # These (x, y) are global
        X = y++                  # As well as vector X
        { auto x=2, X, Y         # All these variables are local...
           X=3; Y=sin(x)
            .
            .
            .
        }                        # ...and stop existing here
        x                        # This x still contains 1
     # An example with a procedure
        proc test(x) {
           auto y=2
           z = x + y++           # This z is global
        }
     fmode

 See also:
     C, cmode, func, proc

?break

 The `break' keyword is used as in C to break C-calculator mode
 `for' or `while' loops. `break' is a C-calculator mode
 command.

 Syntax:
     break

 See also:
     C, continue, cmode, for, while

?C

 The following gives a brief description of the supported C-calculator
 syntax and differences with standard C.

 The following operators are recognized, in order of precedence:

    ++, --                          (post and pre) increment-decrement
    -, !                            unary minus and logical NOT
    ^                               exponentiation, right associative
    /, *, %                         division, multiplication, modulo
    +, -                            addition, subtraction
    >, >=, <, <=, ==, !=            relational operators
    &&                              logical AND
    ||                              logical OR
    =, +=, -=, /=, *=               assignments, right associative

 All operators are left associatives except those specified.
 They are all common to C except for the exponentiation operator.

 The following keywords are reserved tokens:  `auto, if, else, while,
 for, break, continue', and `return', plus two extra keywords 
 `proc, func'. They roughly obey the same syntax as in C so that statements
 like:

     if ("conditions")
         "cmode-line-statement"

 or

     if ("conditions") "cmode-line-statement"

 or

     if ("conditions") {
          "cmode-statements"
     }

 The same thing is true for the else constructions `else' of which
 some examples follow:

     if ("conditions") 
         "cmode-line-statement"
     else
         "cmode-line-statement"

 or

     if ("conditions") {
         "cmode-statements"
     } else {
         "cmode-statements"
     }

 Here "cmode-line-statement" means any semicolon separated list
 of C-calculator mode statements typed on the same line. Since
 semicolons are separators and not terminators, empty statements are
 defined by empty braces `{ }'.

 The `return' keyword must have parentheses when returning a value
 from a function as in `return(x * sin(y))'. A single `return'
 will only be recognized from within a procedure.

 To avoid potential confusion with variables, keywords cannot be
 abbreviated.

 As opposed to C, there exists no integer in the C-calculator mode. All
 scalar variables and numbers are double precision. This means that
 logical true is 1.0 and false is 0.0. As in C, one must be careful
 with comparison operators. The C `switch' syntax is not supported
 (would require integers).

 As an extension, string comparison is possible with the equality
 operators `==' and `!='. This will return true or false if the string
 variables (or constants) are identical or not. Assignments of string
 variables actually copies all characters of the string on the RHS to
 the string variable on the LHS. String additions and subtractions are
 also possible.

 Function and procedure definitions are defined with prototypes, i.e.,
 a list of variables representing the proper kind of variable. At
 run-time, the arguments of the function are checked for type
 compatibility and for their number.

 All variables are global except automatic variables defined using the
 `auto' keyword. 

 See also:
     cmode, let, math, scan, strings, auto

?cd

 The `cd' command changes the working directory. Called with no
 argument, `cd' will bring you to your $HOME directory. Note that
 `cd' changes the current working directory of FUDGIT only.
 Therefore, your plotting program will still be in the previous
 directory. To get around this difficulty, you only have to define a
 macro as follows, if your plotting program supports `cd':

     macro Cd 1
         pmode cd "$1"
         &cd $1
     stop
     alias cd Cd

 Syntax:
     cd "filename(optional)"

 Examples:
     cd
     cd /nazgul/users/fulano

 See also:
     &, pwd, alias

?cmode

 The `cmode' command allows you to go in the C-calculator mode. The
 only way to come back to the main fitting mode is by using the 
 `fmode' command or to type ^D in interactive mode. Commands cannot
 be abbreviated in `cmode'. Parallel to the `cmode' command,
 the `let' command can be used to pass one single command, or
 command line to mathematical parser. To be consistent with `pmode'
 command, `cmode' also accepts arguments in which case it is
 equivalent to the `let' command. It is not an error to call 
 `cmode' from the C-calculator mode. A warning message will be given
 though.

 Syntax:
     cmode "command-list(optional)"

 The C-calculator mode supports most of C syntax (see item C), and most
 of the C math library. Thus, the following functions are supported:

    trigo:    hyperbolic:  expo:    special:   conversion:  random:

    cos()     cosh()       ln()      besy0()    trunc()     srand()
    cot()     coth()       log()     besy1()    floor()     rand()
    csc()     csch()       exp()     besj0()    ceil()
    sec()     sech()       sqrt()    besj1()    rint()
    sin()     sinh()       cbrt()    besjn()    abs()
    tan()     tanh()                 besyn()    int()
    acos()   acosh()                 erf()      scan()
    asin()   asinh()                 erfc()     min()
    atan()   atanh()                 lgamma()   max()
    atan2()                          interp()   sum()
                                                vread()

 Any upper case variable (possibly including `_') possibly mixed with
 digits will be recognized as a vector, e.g., `TEMP_2, TEST, D',
 etc. Any lower case name will be taken as a scalar variable, e.g.,
 `x, t4', etc. There are two predefined constants, `pi' =pi
 and `e'=e, which should not be unlocked and modified. As well,
 the built-in constant `data' contains the current size of the
 vectors and can be modified through the `set data' command, by the
 `read'/`exec' commands, or by `unlock'ing the constant and
 modifying it directly. The built-in constant `chi2' contains the
 value of chi^2 as obtained from the latest fit. And finally, the
 built-in scalar constant `param' contains the number of parameters
 as defined by `set parameters'.

 A mix of upper case and lower case letters will serve to indicate a
 string variable. Strings values are indicated by double quotes as in
 C. Unlike C, FUDGIT considers strings as self-contained objects that
 can be added, subtracted, and checked for (in)equality. Thus, string
 objects (i.e. string variables, string constants and string values)
 can: serve as argument to `scan' function; be part of string
 assignment statements or of a truth statement involving (in)equality
 operator; be added (concatenated using the `+' operator) one with
 another; be subtracted (remove string termination using the `-'
 operator) one with another; and finally be argument of string
 functions.
 
 A predefined string constant called `Tmp' contains the string
 "/tmp/fudgitPID" where PID is the process id number of the current
 process. This file, and any file belonging to you, whose name starts
 with the same string, will be erased automatically by the `exit'
 or `quit' commands. This string is typically used by the 
 `gnuplot' macro in order to pass data to the GNUPLOT plotting program
 which cannot read data from standard input. Another predefined string
 constant is `ReadFile' which contains the last data filename that
 has been loaded. Finally, the string constant `Cwd' is made
 available in order to get the current working directory.

 The following table contains all the built-in constants.

    chi2               Value of chi^2 from the last fit;
    data               Length of all vectors (< samples) as set by set data;
    e                  Neperian number;
    param              Number of parameters as set by set parameters;
    pi                 Guess this one;

    Cwd                Current working directory;
    ReadFile           The last file (program) read by read (exec);
    Tmp                A temporary filename "/tmp/fudgitPID";

 Constants (either strings and scalars) can also be created by 
 `lock'ing a variable. In the same manner, a constant can be modified
 directly if it has been `unlock'ed.

 The algebraic operations applicable to scalar variables can be applied
 to vectors. Vector algebra can be mixed with scalar variable algebra
 in which case the user has to take the implied loop into account. For
 example, although the following operation is not standard C
 programming:

 NOTE: In order to show that some commands can be typed from both
 C-calculator mode and the fitting fmode, the following examples shows
 the typing mode from the first line. However, one can always type the
 same C-calculator mode command from the fitting mode by using the 
 `let' command (or `cmode' command).

     cmode
         x = 0
         X = x++

 will define a vector X of size `data' (see `set data')
 ranging from X[1] to X[data] and taking values from 0 to
 data-1. Multiple commands can be given with the separator `;', for
 example, another version of the previous command could be written

     cmode
         x=0;X=++x

 in which case a vector X taking values from 1 to data will be
 created. (Note that the latter uses a pre-increment whereas the
 former uses a post-increment operator on `x': results are thus
 different). Vector elements can be referenced by elements using
 standard C grammar. Therefore, the same vector could be created by
 using a `while' construction as in:

     fmode
     set data 1000
     let X=0;i=0
     cmode
         while (i++ <= data)
            X[i] = i
     fmode

 or, using a `for' loop, 

     cmode
        for (x=0;x<=data;x++) {
            X[x] = x
        }
     fmode

 Noninteger variables will be truncated to the nearest lower integer to
 form a vector index.

     cmode
         y= 2.01
         x=2.23; X[2]=Z[y]+5^x 

 Assigning a vector to a constant will assign all the elements to that
 constant.

     fmode
     let X = pi
     let Z2 = 0

 The C-calculator checks for undefined variables on the RHS of any
 assignment. From C-calculator mode, variables values can be seen by
 typing the variable name by itself or by using the `print'
 command, if the output is selected to be "stdout". From the
 fitting mode, contents of constants and variables (either strings or
 scalars) is displayed using `show variables' command, or by using
 the `$' expansion operator. However, vectors can be only be seen from
 the fitting mode by using the `show vector' command.

 Each unknown vector name given on the command line allocates a vector
 of `sample' size.

 To be a calculator as such, the C-calculator prints the value of the
 expression given on the command line. Thus, the statement

     cmode
         x + 2

 will print the value of x + 2.0. The contents of many variables can
 be displayed at the same time by giving a coma separated list such as
 in

     cmode
         x,"temperature", t

 where the string "temperature" will be printed between the values
 of variables x and t. Note that the C-calculator mode recognizes
 strings by double quotes. Special characters such as '\n' are
 also legal in a string.

 We conclude by giving some examples involving string variables:

     fmode
     let String = "new.file"
     let x = (String == "new.file")
     let y = ("file1" == "file2")
     let Bing = "aaa"
     let Here = Cwd        # Store the value of the current working directory
     let Input = Read()                       # Read from stdin
     let Test = FileName(ReadFile) - ".data"
     let Dir = DirName(InputFile)
     let y = scan(Read(), "%lf")
     let File = "STRING_23.4"
     let number = scan("%*[_A-Z]%lf", File)
     let Message = "A tab t and a newlinen"

 where the truth statement could be legally used as a condition for an
 `if', a `while', or a `for'.

 See also:
     let, C, data, func, proc, print, fmode, math, while, for, return,
     auto, if, break, samples, quotes, strings

?comments

 By default, anything following a ``#'' will be treated as a
 comment and ignored. This holds for data files as well as for command
 script files loaded with the `load' command. This default can be
 changed with the `set comment' command. Sometimes a comment
 character needs to be taken literally in a script file. The comment
 character will be accepted as data if it follows the `\' escape
 operator, i.e. `\#', or, in the fitting mode only, whenever the
 comment character is somewhere inside quotes or parentheses. The
 comment character is always accepted literally when typed on the
 interactive command line.

 See also:
     set comment, read, load, show comment, exec

?continue

 The `continue' keyword has the same usage it has in C for sending
 the control to the next iteration of a `for' or `while' loop.
 `continue' is a C-calculator mode command.

 Syntax:
     continue

 See also:
     for, while, cmode, C

?datafiles

 Files containing data are loaded by specifying the name of the data
 file to the `read' command. Data files should contain one data
 point per line. A data point can be a 256 dimensional object. By
 default, anything following character ``#'' will be treated as
 comment and ignored. In all cases, the numbers on each line of a data
 file must be separated by any number of blank spaces or tabs. These
 blanks divide each line into columns. Thus, FUDGIT can handle up to
 256 columns per line. Warning will be given if a line has a different
 number of columns. Strings such as "NaN" or "Infinity" are
 recognized and refused. The default compilation gives a maximum line
 size of 1024 characters.

 See also:
     read, exec, set comment

?echo

 The `echo' command allows the user to print a string to the
 standard output. If no argument is given `echo' will only print a
 newline. This command can be used to display a message or, when
 coupled with the variable expansion operator `$', to see the value of
 a printable (either string or scalar) variable defined in the
 C-calculator.

 Syntax:
     echo "string-list"

 Examples:
      echo Starting the fit
      echo $Mydir

 See also:
     cmode, $

?else

 The `else' keyword is used in `if' constructions, both in
 C-calculator anf fitting modes. Refer to the `if' entries for a
 complete description.

?end

 The `end' command is used to complete a `foreach' loop or a
 `while' loop. Keyword `end' is also used to tell `read'
 that we are finished writing data to stdin. This command should always
 be found on a line by itself (comments are allowed though).

 See also:     foreach, while, read, stop

?endif

 The `endif' command is used to complete an `if' construction
 in fitting mode. Keyword `endif' must always be used on a line by
 itself (comments are allowed though). Refer to the `if' entries
 for a the complete description.

?environment

 FUDGIT is sensitive to the following environment variables:

   =>  PAGER for the program called to format long listings.

   =>  HOME for the directory to which `cd' defaults.

   =>  SHELL for the shell called by `system' when this
          latter is called without arguments.

 If not defined, the default pager is "/usr/?/more" (path depends
 on system) and the default shell "/bin/csh".

 See also:
     cd, system, show vectors, help

?exec

 The command `exec' executes a program and reads data from it. It
 supports the same syntax `read' does except that the program name
 replaces the file name. A program is a program name or anything that
 can be typed in a shell. If the command line has more than one string,
 it must be glued with quotes. On a successful call, `exec' will
 set the string constant `ReadFile' to the name of the program
 which generated the data.

 Syntax:
     exec "commands" "assignment[range](optional)" ...

 Examples:
     exec simulate X:1 Y:2[0:200]
     exec "cat data | myfilter -g" X1:1[0:*] X2:2 X3:4

 See also:
     read, comments

?exit

 The commands `exit' and `quit' will exit FUDGIT.
 See details under item `quit'.

 Syntax:
     exit

 See also:
     quit, cmode

?fft

 The `fft' command will take the Fourier transform of the specified
 vectors and put the real part in a vector specified by the third
 argument. The imaginary part will be put in a vector specified by the
 fourth argument. Input vectors can be used for output. The resulting
 vectors will contain frequencies ranging from 0 to N/2 followed by
 -(N/2 - 1) to -1 in units of 1/(N*Delta) where Delta is the
 sampling rate. If a real vector is transformed h(t) -> H(f), we
 should have H(-f) = H^*(f). Therefore, with H = R + iI and
 H^* = R - iI be the transformed vectors, we should have
 R(-f) = R(f) and I(-f) = -I(f), where f is discrete and ranges
 as mentioned above. In terms of vector indices, these relations
 become R[i] = R[N-i+2] and I[i] = -I[N-i+2] for 1 < i < N/2 in
 addition to the fact that I[1] = I[(N/2)+1] = 0. Therefore, because
 the negative frequency part is the mirror image of the positive one,
 it is common to plot only the positive frequencies of the Fourier
 transform of a real vector. This can be done by reducing `data' to
 half its value.

 Because of the use of a FFT algorithm, the number of data points must
 be an integer power of 2. If not, the user should pad the vector with
 zeros up to the next largest power of two. Each transform is
 normalized by the factor sqrt(N) so that `fft RE IM T_RE T_IM'
 followed by `invfft T_RE T_IMA RE2 IMA2' will not introduce a
 factor N in vectors `RE2' and `IMA2' (i.e., `RE' = `RE2'
 and `IM' = `IM2'). At his choice, the user can use the
 C-calculator functionality in order to implement windowing.

 The power spectrum can be obtained from:

      fft RE IMA T_RE T_IMA
      let POW = T_RE^2 + T_IMA^2

 where POW[i] will contain the power value associated with frequency f,
 which goes from 0 to N/2 followed by -(N/2 - 1) to -1 (in units
 of 1/(N*Delta)) as i goes from 1 to N.

 Syntax:
     fft "real-VECTOR" "ima-VECTOR" "real-VECTOR" "ima-VECTOR"

 Examples:
      # real vector X
      let IM=0
      # re-use IM vector for output
      fft X IM Z IM
      # complex vectors X+iY where i = sqrt(-1) transformed in V+iW  
      fft X Y V W

 See also:
     invfft, smooth, cmode, let, read, math, data

?fit

 The `fit' command is used to fit a function, chosen by `set
 function', to a pair of vectors containing the independent and
 dependent variables. Depending on the type of fit, selected by the
 `set method' command, a third vector containing the standard
 deviation might be required. `fit' allocates a vector having the
 name of the dependent variable appended with the string `FIT'.
 This vector contains the computed values of the function for the given
 independent vector. Depending on the method, the built-in constant
 `chi2' will contain the value of the mean square deviation
 weighted by vector "sigma-VECTOR" or the mean absolute deviation.

 Syntax:
     fit "independent-VECTOR" "dependent-VECTOR" "sigma-VECTOR"

 Example:
     fit X Y DY

 will create a vector `YFIT' containing the value of the fitted
 function for each of the values of the independent vector `X'.
 Note that the standard deviation is required for most fitting routines
 since it is used to weigh the value of local square deviation from the
 fit (in fact, this is the definition of chi^2). If 
 "sigma-VECTOR" is unavailable just use

     let DY=1

 using the previous example. This simply gives the same weight to all
 data points.

 See also:
     set method, set function, show fit, show parameters, append

?fmode

 The `fmode' command allows you to return to the fitting mode, when
 the program is in one of the C-calculator or plotting modes. The
 fitting mode, is the main mode of the program. The two other modes are
 the C-calculator mode, accessed by the `cmode' command, and the
 plotting mode, accessed by the `pmode' command. When used
 interactively, ^D returns to the fitting mode from either of the
 C-calculator mode or from the plotting mode. It is not an error to
 call `fmode' from the fitting mode. A warning message will be
 given though.

 Syntax:
     fmode

 See also:
     cmode, pmode, let

?for

 The `for' command is a C-calculator mode command. It behaves
 roughly like a standard C `for' construction. In interactive
 mode, any new input line will be prompted with a ``n{... n\t''
 where `n' stands for the nesting level and `\t' for a tab.
 Keyword `for' is a C-calculator mode command.

 Syntax:
     for ("init-expressions"; "cond-expressions"; "loop-expressions") 
           "cmode-line-statement"

 or 

     for ("init-expressions"; "cond-expressions"; "loop-expressions") {
           "cmode-statements"
     }

 Examples:
     cmode
         for (i=1,j=2;i+j <= data; i+=2,j+=3) A[i] = X[j]
     fmode
     # Another example:
     # A macro to remove point x in a vector. Syntax: delete "vector" "index"
     macro delete 2
         cmode
             for(i=$2;i<data;i++) {
                 $1[i] = $1[i+1]
             }
         fmode
         unlock data
         let data--
         lock data
     stop

 See also:
     C, break, continue, cmode, if, set data, func, proc, if, lock, math

?foreach

 The `foreach' command loops through the strings obtained from a
 given UNiX command. Wild card characters are allowed since
 everything following the `in' keyword is passed to a Bourne shell
 for execution. Strings can be obtained from any program including the
 easiest cases "echo, ls" and "cat". The variable name must be
 of string type, i.e., consisting of both upper case and lower case
 letters (and possibly _'s and digits).

 Syntax:
     foreach "StringVarName" in "UNiX-command"
          "body of the loop"
     end

 Example:
     #convert columns 2 and 3 of the following files in log-log format
     foreach Fname in ls data*.7[0-9] datatest.42 data*.8[4-7]
         echo $Fname ...
         read $Fname X:2[0.001:*] Y:3[0.001:*]
         let X = log(X)
         let Y = log(Y)
         save vectors X Y $Fname.log
     end

 See also:
     for, math function scan, while, macro

?free

 The command `free' is made available for memory management. It is
 used to free vectors, variables, functions, procedures, and numbers
 that were allocated in the C-calculator mode.

 When called with the special argument ``@all'', `free' will erase
 all the user vectors, numbers and variables, as well as all active
 functions and procedures (not macros and aliases). Otherwise, `free'
 will free the specified vector(s) or variable(s). Constants (either
 scalar or string) cannot be removed without first unlocking them.

 Syntax:
     free "VECTOR- or variable-list" 
     free @all

 Examples:
      free @all
      free X y TEMP

 See also:
     unlock, C, cmode, show table, show memory, samples, let

?func

 The `func' command defines a function. A function is distinct from
 a procedure from the fact that a function must return a value whereas
 a procedure must not. Arguments are given in the definition with any
 name prototype representative of the data type. As in C, the argument
 list must be comma separated when calling the function (after having
 defined it). An example follows. `func' is a C-calculator mode
 command.

 The prototype list defines the type of variable to be used. Although
 all global variables are accessible from within the function,
 variables are always searched for from the prototype list first, then
 from the local list (`auto' variables), and finally from the
 global list. All scalar variables are passed by value: thus any scalar
 expression is legal as scalar argument. String arguments and vector
 arguments are passed by pointer: thus string and vector arguments must
 refer to a variable explicitly. The `show table' can be used to
 list all the installed objects at a given time.

 Syntax:
     func "functionname"("proto-list(optional)") "cmode-line-statement"; return("value")

 or

     func "functionname"("proto-list(optional)") {
          "cmode-statements"
          return("value")
     }

 Examples:
    # The following example will print the factorial of all integers up to 120.
    cmode
        func fac(x) {  # This `x' is a prototype: it does not exist.
            if (x <= 0)  {
               return(1)
            } else {
               return(x * fac(--x))
            }
       }
       x=1  # This `x' is a global scalar variable
       while (x<120) {
           fac(x++)
       }
    fmode
    # The following calculates the average of a vector
    cmode
        func avg(X) {
            auto i,x

            for (x=0,i=1;i<=data;i++) {
                x += X[i]
            }
            return(x/data)
        }
    fmode

 See also:
     C, return, for, while, cmode, math, free, proc, show table, install

?help

 The `help' command displays on-line help. To specify information on a
 particular topic use the syntax:

     help "topic"

 If "topic" is not specified, a short message is displayed about
 FUDGIT. Topic names can be abbreviated down to the shortest
 unambiguous string. In case of doubt, `help' will print out all
 possible completions. Thus, `help f' will print all help topics
 starting with the letter `f'. After help for the requested topic has
 been given, help for a subtopic may be requested by typing the
 subtopic name, extending the help request. After that subtopic help
 has been displayed, the request may be extended again, or pressing
 return will return one level back to the previous topic. Eventually,
 the fitting mode prompt will return.

 See also:
     help?

?history

 The `history' command lists all the previous command lines, along
 with a number. History lines can be called using the !"string"
 construction or the !"number". History is only available in
 interactive mode. See Appendix A for more details.

 Syntax:     history

 See also:
     append history, line editing

?if

 There are two kinds of `if' constructions available in FUDGIT,
 one in the fitting mode and the other in the C-calculator mode.

?if cmode_style

 In C-calculator mode, `if' and `else' are reserved keywords.
 C-calculator mode `if' construction is similar to the one in
 standard C. Note that "cmode-statements" refers to any sequence of
 C-calculator mode commands and that "cmode-line-statement" refers
 to a semicolon separated list of C-calculator mode commands typed on
 the same line.

 Syntax:
     if ("conditions") "cmode-line-statement"

 or

     if ("conditions")
         "cmode-line-statement"

 or

     if ("conditions") {
          "cmode-statements"
     }

 or, using the `else' constructions,

     if ("conditions")
          "cmode-line-statement"
     else
          "cmode-line-statement"

 or, for statements on more than one line,

     if ("conditions") {
          "cmode-statements"
     } else if ("conditions") {
          "cmode-statements"
     } else {
          "cmode-statements"
     }

 See also:
     C, cmode

?if fmode-style

 Fitting mode `if' has a syntax very similar to the one in C-shell.
 It requires the keywords `then' and `endif' and supports 
 `else' constructions. The difference resides in the fact that the
 conditional statement has to follow C-calculator mode grammar and
 syntax and thus has a richer set of operators. The `$' expansion
 operator is therefore not needed in the conditional statement, as it
 is in C-shell conditional statements. All active variables, constants,
 and their string counterparts, are directly available to the
 conditional statement. Note that the "fmode-statements" can also
 contain C-calculator mode commands (even possibly including
 C-calculator mode `if''s!).

 Syntax:
     if ("conditions") then
          "fmode-statements"
     endif

 or, using the `else' constructions,

     if ("conditions") then
          "fmode-statements"
     else if ("conditions") then
          "fmode-statements"
     else
          "fmode-statements"
     endif

 See also:
     while, foreach, macro

?in

 The `in' keyword is required in `foreach' constructions in
 fitting mode. Refer to the latter for details.

?install

 The `install' command dynamically loads defined routines from an
 object file. The user decides on the internal name of the routine but
 the internal name must consist of lower case letters only. The object
 file is an object compiled by the C or FORTRAN compiler. On IRIX, the
 object must be compiled with the option `-G 0' given to either the
 C or FORTRAN compiler. The "rtn-name" is the name of one of the
 procedure(s) or function(s) the user wants to install from the
 object file. The routine will be installed as "name" and as a
 procedure (not returning value) or as a function (returning value)
 depending on the name separator being a `:' (colon) or a `='
 (equal sign) respectively. (See example below).

 NOTE: This option is only available on IRIX and SUNOS for the moment.

 The external routine must expect pointers to double for all its
 arguments. Thus, all arguments are passed by pointers except that
 pointers to variables do not point to the variables as such but to a
 temporary copy of them. This allows us to have expressions like

 `f = mycall(X, sin(x) + 1, data)'

 for which the value `sin(x) + 1' must necessarily be a temporary copy.
 In this example, the prototype is a function `mycall(VEC, expr, expr)'.
 We shall consider an example in more detail below. All arguments are
 strongly typed as vector, parameter, expression or string. Prototyping
 is done using the uppercase-lowercase convention. Parameters are prototyped
 using the word `PARAM' or less (e.g. `PAR').

 On IRIX, the linker will create a binary file built from the module
 name and with the extension "file".`ld'. This binary is the
 one that will be loaded in memory. Time stamps are included so that
 `ld(1)' will not be called if not necessary. These files are not
 erased at exit, since they are reusable and prevent the linker to be
 called if nothing changed between two sessions of FUDGIT.

 Successive calls of `install' with the same module should not be
 done unless the same functions and procedures are reinstalled. If this
 is the case, the user should then reinstall the same modules (that
 could have been modified and recompiled in the mean time) using the
 `reinstall' command. If a module is reinstalled with different
 routines or function or procedure names, the previously defined
 functions or procedures might not be properly installed anymore and
 calling them might result in an undefined behavior.

 The file "fudgit.h" describes the functions user-defined programs
 can linked with. Among other things, these functions allow the user to
 have elegant error handling and exit.

 The `show table' command can be used to list all the installed
 objects at a given time.

 A file having the same base name of the module but with the extension
 "libs" can be put in the same directory in order to include extra
 libraries while loading the module. On IRIX, these extra libraries
 must all contain objects compiled with the flag "-G 0" (see 
 `cc(1)').  (For example, some IRIX systems have a -lm_G0 math
 library.) User-defined libraries can be specified along with system
 libraries. A typical example could be a line like:

     /home/myname/myproject/libmyG0.a /usr/lib/libmG0.a

 for linking with user's library "/home/myname/myproject/libmyG0.a".
 Equivalently, for non-IRIX systems, loading a FORTRAN object might
 require something like this:

     /usr/lib/libF77.a /usr/lib/libm.a

 Library names can be on multiple lines. However, the file cannot have
 more than 1024 bytes. A ``#'' found anywhere in this file will make
 the rest of the file to be ignored.

 Note that when loading FORTRAN code, the user must append an underscore
 to the routine name so that `install' or `reinstall' can find it.

 The IRIX version does not fully support incremental linking, i.e., to
 use, in an object to be installed, symbols that were defined in
 previously `install'ed objects. However, all the symbols contained
 in the original FUDGIT executable remain at all time available to
 all linked routines. Therefore, IRIX users should make sure that
 external objects are self-contained and only reference to external
 routines that are intrinsic to FUDGIT or come directly (and once)
 from linked libraries at installation time.

 Syntax:
     install "object-file" "rtn-name"[:|=]"name"(arg-list)...

 Example:
    hostname: cat mymodule.c
        #include <math.h>
        #include "fudgit.h"
        
        /* An example of a user-defined routine inversing the order of an even
         * vector. Typical call would be:
         *  myproc(A_VEC, data)
         * from C-calculator mode. NOTE that both VEC and expr are pointers.
         * To make things explicit, fudgit.h contains a few typedef's.
         */
        
        void myproc(X, dn)
        VEC X;
        expr dn;
        {
            int i, half_n;
            int n = (int)*dn;  /* note that dn is a pointer to a double */
            double tmp;
    
            if (n%2 == 1)   /* report error if odd number (Why not?)*/
                Ft_matherror("%s: Called with an odd number %d.", "myproc", n);

            /* You have full use of math and stdio libraries too!!! */
            fprintf(stderr,
            "BTW, Did you know that %lf is the sqrt(pi)?n", sqrt(M_PI));

            half_n = n >>1;  /* half of n */
            for (i=0;i<half_n;i++) { /* Standard C: indices from 0 to data-1 */
                tmp = X[i];
                X[i] = X[n-i];
                X[n-i] = tmp;
            }
        }
        
        /* 
         * Another example involving a function. The following calculates the
         * non-normalized correlation between vectors A and B as defined by 
         * corr(A, B) = <A*B> - <A> * <B>
         *
         */
        
        double myfunc(A, B, dn)
        VEC A, B;
        expr dn;
        {
            int i, n = (int)*dn;  /* Again, dn is a pointer to a double */
            double sumA, sumB, sumAB;
        
            sumA = sumB = sumAB = 0.0;
            /* sum up the values of interest */    
            for (i=0;i<n; i++) {  /* indices go from 0 to data-1 */
                sumA += A[i];
                sumB += B[i];
                sumAB += A[i] * B[i];
            }
            /* leave it simple */
            sumA /= *dn;
            sumB /= *dn;
            sumAB /= *dn;
        
            return (sumAB - sumA*sumB);
        }
    hostname: cc -G 0 -O -c mymodule.c
    hostname: cat loadex.ft
        # This is an example for loading
        # Install function myfunc as corr() and procedure myproc as inverse()
        # Prototypes are made from any name representing the proper type:
        install mymodule.o myproc:inverse(V, n) myfunc=corr(V, V, n)
        set data 24
        let x=1;X=x++
        let Y=sin(X)
        cmode
            # Inverse order of vector X
            inverse(X, data)
            # Calculate correlation between X and Y
            y=corr(X, Y, data)
            # Print its value
            "correlation:", y
        fmode
    hostname: fudgit loadex.ft
    install: myproc installed as procedure inverse.
    install: myfunc installed as function corr.
    BTW, Did you know that 1.772454 is the sqrt(pi)?
    correlation: 9.20717026e-01 

 When linking FORTRAN functions or subroutines, the user must append an
 underscore after every function or subroutine name. All argument
 variables and vectors have to be defined `double precision' as
 well as returning functions. Typical examples are included in the
 distribution in the "tools" directory.

 See also:
     C, cmode, show table, func, proc

?invfft

 Command `invfft' performs the inverse Fourier transform of the
 given vectors. It assumes that the frequencies are ordered from 0 to
 N/2 followed by negative frequencies ranging from -(N-1)/2 to -1
 in units of 1/(N*Delta) where Delta is the sampling interval.
 The results are normalized by a factor 1/sqrt(N) so that a
 transform followed by an inverse transform should give the original
 vector. The resulting vectors are stored in the third and fourth
 arguments. Thus, `invfft X Y V W' inverse transforms X+iY into
 V+iW. Input vectors can be used as output vectors. See `fft' for
 more details.

 Syntax:
     invfft "real-VECTOR" "ima-VECTOR" "real-VECTOR" "ima-VECTOR"

 See also:
     fft, smooth, cmode, let, read, math, data

?let

 The `let' command opens the door to the C-calculator mode from the
 fitting mode, but leaves the program in fitting mode. All the 
 `let' commands can always be typed directly from the C-calculator mode
 without having to prepend with the `let' keyword. The converse is
 also true; all the commands given in C-calculator mode could be typed
 from the fitting mode by prepending them with the `let' command.
 Although `let' is typed from the fitting mode, the remainder of
 the line is parsed according to C-calculator mode rules, and thus
 quotes are no longer swallowed. Variable expansion operator `$' is
 still recognized, but its use is not recommanded for C-calculator
 statements. See `$' for more details on this point.

 Syntax:
     let "C-calculator-mode-commands"

 Examples:
      # generate the zero order first kind bessel
      # function between (0, 2*pi]
      fmode
      set data 2000
      let x=1; X=x++
      let tmp = 2*pi/data     # compute sequence only once
      let X *= tmp
      let Y = besj0(X)

 See also:
     cmode, C, math

?editing
?line

 The command shell supports line editing and history. The editing
 commands are based on the basic EMACS commands. A short summary
 follows but a more complete description can be found in Appendix B.

 Line editing:

   =>  ^B moves back a single character.

   =>  ^F moves forward a single character.

   =>  ^A moves to the beginning of the line.

   =>  ^E moves to the end of the line.

   =>  ^H and DEL delete the previous character.

   =>  ^D deletes the current character.

   =>  ^K deletes from current position to the end of line.

   =>  ^L,^R redraws line in case it gets trashed.

   =>  ^U deletes the entire line.

   =>  ^W deletes the last word.

 History:

   =>  ^P moves back through history.

   =>  ^N moves forward through history.

   =>  !!   previous command.

   =>  !$  previous command last argument.

   =>  !"string"  last command starting with "string".

 Completion:

   =>  tab complete command if first arg, filename otherwise.

   =>  esc-? or double tab list possible completions.

 Each line of input must be smaller than 1024 bytes which is more than
 sufficient for most applications. Lines can be continued on several
 lines provided carriage returns follow a `\' (as in standard
 shells).

 See also:
     append history, $, history

?load
?source

 The `load' command executes each line of the specified input file
 as if it had been typed in interactively. Files created by the 
 `save history' command can be `load'ed directly. Text files
 containing valid commands can be created and then executed by the 
 `load' command. Files being `load'ed may themselves contain 
 `load' commands. See `comment' for information about comments in
 command scripts. The `load' command is recursive so it can be
 nested. The only limitation is the I/O stack which has a default
 capacity of 32. This value can be easily changed at compilation time
 of the program.

 The current working directory always returns to the value in effect
 before the loaded script was called. This is valid for nested 
 `load' commands too.

 In order to avoid confusion between data files and script files we
 strongly recommand you to stick to the conventional ".ft" extension
 for your script files.

 Syntax:
     load "filename.ft"

 A `load' command is also performed implicitly on any filenames
 given as arguments to `fudgit', when called from your UNiX
 session. These are loaded and executed in the order specified, and
 then FUDGIT exits.

 See also:
     set comment, exec, startup, append history, append macros

?lock
?constant

 Variables can be turned into constants using the `lock' command.
 Once a variable is `lock'ed, any assignment trying to change its
 value will result in a parsing error. This is valid for both scalar
 and string variables. It is not an error to try to lock a constant. A
 warning message will be given though. However, trying to lock an
 unexisting variable or something else than a constant or variable will
 result in an error.

 Syntax:
     lock "var-list"

 See also:
     C, cmode, unlock

?ls

 The command `ls' calls ``/bin/ls -FC''. If any arguments are
 given, those are passed to ``/bin/ls -FC''. Wild card characters are
 possible since expansion is done by a Bourne shell.

 Syntax:
     ls "ls-argument-list"

 Examples:
      ls p* test?
      ls -l datafile
      ls -l *.data

 See also:
     system, alias

?macro

 The `macro' command allows the user to define macros. Macros can
 be embedded, but another macro cannot be defined from within a macro,
 mainly because of their common way to refer to arguments. The name of
 the macro followed by the number of arguments required must be given.
 The maximum number of arguments a macro can have is 16. An exclamation
 mark in the macro name will indicate that the macro name can be
 abbreviated and that the characters following the exclamation point
 are optional. Macros are only recognized in the fitting mode. The
 total length of each macro is limited to 2048 bytes in size. Macros
 can be nested to a maximum of 32. Macros are only recognized from the
 fitting mode.

 Syntax:
     macro "macroname" "argument-number"
         "body of the macro"
     stop

 Example:
     # define a macro named fpl!ot (o, t, are optional)
     # requiring 3 arguments . Uses the plotting program gnuplot.
     # Syntax: fplot X Y YFIT
     # plot X Y with data points and X YFIT with solid line
     macro fpl!ot 3
         # save vectors in temp file (will be automatically removed on exit)
         save vec $1 $2 $3 $Tmp.fplot
         # plot second column with points and third with line
         pmode plot '$Tmp.fplot' us 1:2 wi point, 
            '$Tmp.fplot' us 1:3 wi line
     stop

 See also:
     append macros, show macros, load, startup, unmacro, alias, unalias

?math

 The C-calculator mode math functions found in FUDGIT are very close
 to the corresponding functions found in the UNiX math library. Some
 other functions, not found in the math library, are also part of
 FUDGIT. Most of the numerically unstable functions (i.e. ln, log,
 exp,...) check for both an argument out of range and a value out of
 domain at each call. All math functions are double precision and can
 only be called from the C-calculator mode, or by using the `let'
 command from the fitting mode. These functions are also available in
 the conditional statements of the fitting mode `if' and 
 `while', since these statements are C-calculator mode statements,
 although part of fitting mode constructions.

?math abs
?abs

 The `abs()' function returns the absolute value of its argument. 

?math acos
?acos

 The `acos()' function returns the arc cosine (inverse cosine) of
 its argument. `acos()' returns its argument in radians.

?math acosh
?acosh

 The `acosh()' function returns the positive (principal) hyperbolic
 arc cosine (inverse cosine) of its argument.

?math asin
?asin

 The `asin()' function returns the arc sine (inverse sine) of its
 argument. `asin()' returns its argument in radians.

?math asinh
?asinh

 The `asinh()' function returns the hyperbolic arc sine (inverse
 sine) of its argument.

?math atan
?atan

 The `atan()' function returns the arc tangent (inverse tangent) of
 its argument. `atan()' returns its argument in radians.

?math atan2
?atan2

 The `atan2(y, x)' function returns the arc tangent (inverse
 tangent) of the ratio of its arguments (y/x). `atan2()' returns
 its argument in radians. The signs of y and x are used to
 determine the quadrant.

?math atanh
?atanh

 The `atanh()' function returns the hyperbolic arc tangent (inverse
 tangent) of its argument.

?math besj0
?besj0

 The `besj0()' function returns the j0th Bessel function of its
 argument, i.e it returns the zero^th order Bessel function of the
 first kind. `besj0()' expects its argument to be in radians.

?math besj1
?besj1

 The `besj1()' function returns the j1st Bessel function of its
 argument, i.e it returns the first order Bessel function of the first
 kind. `besj1()' expects its argument to be in radians.

?math besjn
?besjn

 The `besjn(n, x)' function returns the jnst Bessel function of its
 argument, i.e it returns the n^th order Bessel function of the
 first kind. `besjn()' expects its second argument to be in
 radians.

?math besy0
?besy0

 The `besy0()' function returns the y0th Bessel function of its
 argument, i.e it returns the zero^th order Bessel function of the
 second kind. `besy0()' expects its argument to be in radians.

?math besy1
?besy1

 The `besy1()' function returns the y1st Bessel function of its
 argument, i.e it returns the first order Bessel function of the second
 kind. `besy1()' expects its argument to be in radians.

?math besyn
?besyn

 The `besyn(n, x)' function returns the ynst Bessel function of its
 argument, i.e it returns the n^th order Bessel function of the
 second kind. `besyn()' expects its second argument to be in
 radians.

?math cbrt
?cbrt

 The `cbrt()' function returns the cubic root of its argument.

?math ceil
?ceil

 The `ceil()' function returns the smallest integer that is not
 less than its argument.

?math cos
?cos

 The `cos()' function returns the cosine of its argument. 
 `cos()' expects its argument to be in radians.

?math cosh
?cosh

 The `cosh()' function returns the hyperbolic cosine of its
 argument.

?math cot
?cot

 The `cot()' function returns the cotangent of its argument. 
 `cot()' expects its argument to be in radians.

?math coth
?coth

 The `coth()' function returns the hyperbolic cotangent of its
 argument.

?math csc
?csc

 The `csc()' function returns the cosecant of its argument. 
 `csc()' expects its argument to be in radians.

?math csch
?csch

 The `csch()' function returns the hyperbolic cosecant of its
 argument.

?math erf
?erf

 The `erf()' function returns the error function of its argument.
 The error function is defined as

    2/sqrt(pi) * integral from 0 to x of  exp(-t^2) dt 

?math erfc
?erfc

 The `erfc()' function returns `1 - erf()' where `erf()' is
 the error function of its argument. It is provided because of the
 extreme loss of relative accuracy if `erf(x)' is called for large
 x and the result subtracted from 1.0 (e.g., for x = 10, 12 places
 are lost).

?math exp
?exp

 The `exp()' function returns the exponential function of its
 argument (e raised to the power of its argument). Overflow is
 checked on all `exp()' operations.

?math floor
?floor

 The `floor()' function returns the largest integer not greater
 than its argument.

?math hypoth
?hypot

 The `hypot(x, y)' function returns sqrt(x*x+y*y) computed in such
 a way that underflow will not happen, and overflow occurs only if the
 final result deserves it.

?math int
?int

 The `int()' function returns the integer part of its argument,
 truncated toward zero. The returned value is still a double. This
 function is equivalent to trunc() is is kept for compatibility.

?math interp
?interp

 The `interp()' function returns an interpolated value of the
 function at the value of its argument. The functional relation is
 previously initialized using the fitting mode command `spline'.
 The interpolation is obtained from cubic splines. "Natural" (i.e.,
 the second derivative of the interpolating function at either or both
 the first and last point of the original data equal zero) cubic spline
 or specific first derivatives at the extreme points of the original
 data set are specified while initializing the process using `spline'
 command.

 See also:
     spline

?math lgamma
?lgamma

 The `lgamma()' function returns the natural logarithm of the gamma
 function of its argument. For an integer `n, lgamma(n+1) =
 ln(fac(n))' where fac is a factorial function.

?math ln
?ln

 The `ln()' function returns the natural logarithm (base e ) of
 its argument. Illegal argument is checked for.

?math log
?log

 The `log()' function returns the logarithm (base 10) of its
 argument.

?math max
?max

 The built-in function `max(x, y)' returns the maximum value of
 x and y. `max(x, max(y, z))' obviously returns the largest value
 of x, y, and z.

?math min
?min

 The built-in function `min(x, y)' returns the minimum value of
 x and y. `min(x, min(y, z))' obviously returns the smallest value
 of x, y, and z.

?math rand
?rand

 The `rand()' function returns a random number between [0,1).
 Depending on the machine on which it is compiled, it might use the
 extended 48 bits random number generator or less.

?math rint
?rint

 The `rint()' function returns the value of its argument rounded to
 the nearest integer.

?math scan
?scan

 This math function is a bit different from others in the fact that it
 handles strings and returns a number. In fact, the `scan(
 "String, Format")' function returns a double precision number as
 extracted from string "String" and according to string format 
 "Format". The format is built with the same rules `sscanf' uses.
 See man pages on `scanf(3)'. Note that the format must contain
 one active "%lf". An example might be of some help here, especially
 to show how to use `scan' in conjunction with C-calculator mode
 defined strings. `scan' is particularly helpful to extract numbers
 from filenames. Recall that strings are defined by double quotes as
 in standard C.

 At this point, it might be useful for you to know the "%[ ]" scanf
 construction. Let's go through some examples: `"%*[a-zA-Z]"'
 means to ignore the longest string matched so that it is composed of
 any letter; `"%*[^0-9]"' means to ignore the longest string
 matched so that it is NOT composed of any digit; `"%*[^_.]"'
 means to ignore the longest string matched so that it is not composed
 of characters `_' or `.'.

 Examples:
     # define a string called Testname
     let Testname = "dummy25.dat"
     # let y be the Neperian log of the number contained in that string
     let y = ln(scan(Testname, "%*[^0-9]%lf.dat"))
     # The following reads a number from stdin
     let input = scan(Read(), "%lf")

 See also:
     $, strings, C, cmode, quotes

?math sec
?sec

 The `sec()' function returns the secant of its argument. 
 `sec()' expects its argument to be in radians.

?math sech
?sech

 The `sech()' function returns the hyperbolic secant of its
 argument.

?math sin
?sin

 The `sin()' function returns the sine of its argument. 
 `sin()' expects its argument to be in radians.

?math sinh
?sinh

 The `sinh()' function returns the hyperbolic sine of its
 argument. `sinh()' expects its argument to be in radians.

?math sqrt
?sqrt

 The `sqrt()' function returns the square root of its argument.

?math srand
?srand

 The `srand()' function sets the seed of the random number
 generator. Its argument will always be truncated to an integer
 towards zero. `srand()' returns the truncated value.

?math sum
?sum

 The `sum' function returns the sum of the elements of the vector
 passed as an argument. Recall that vector are passed by pointers so
 that `y = sum(X^2)' is not legal. Instead, on must explicitly
 calculate

 # Given vector X, the following calculates the sum of X^2
 let X2 = X^2
 let y = sum(X2)

 in order to evaluate the sum. The `sum' function can be used to
 calculate basic statistics (mean, standard deviation, correlation, ...)
 and to do basic integration together with a spline-interp algorithm
 if the points are distant and the function smooth enough.

 See also:
     interp, spline

?math tan
?tan

 The `tan()' function returns the tangent of its argument. 
 `tan()' expects its argument to be in radians.

?math tanh
?tanh

 The `tanh()' function returns the hyperbolic tangent of its
 argument. `tanh()' expects its argument to be in radians.

?math trunc
?trunc

 The `trunc()' function returns the value of the argument when
 truncated towards zero.

?pause
?wait

The `pause' command displays any text associated with the command
and then waits a specified amount of time or until a carriage return is
pressed if the given time value is a negative integer. The `pause'
command is especially useful in conjunction with `load'ed files.

 Syntax:
     pause "value" "string(optional)"

 Examples:
     pause -1
     pause 3
     pause -1  Hit return to continue
     pause 10  This fits equation 4 to file $ReadFile.

 See also:
     echo, load

?plot

 There exists no plot command as such. However two macros are
 predefined. One is `gnu!plot' to use with GNUPLOT and 
 `sgi!plot' to use with SGIPLOT. As they currently are, only two
 vectors can be passed to these macros. They serve like examples for
 building your own macros as well. See `show macros' to see the
 contents of the predefined macros of your site.

 See also:
     set plotting, special, macro, show macros

?pmode

 The `pmode' command talks directly to the plotting program chosen
 with the `set plotting' command. Any command usually typed to the
 plotting routine is now valid. Furthermore, all the current variables,
 constants and their string counterparts can be expanded in the
 plotting mode. The fitting macros and aliases are not recognized in
 this mode. The command `fmode' permits the user to return from the
 plotting mode as does ^D when typed interactively. If `pmode'
 is called with trailing arguments, the remainder of the line will be
 passed to the plotting program while remaining in fitting mode. It is
 not an error to call `pmode' from the plotting mode. An warning
 message will be given though.

 If the plotting program is defined as a null string (`set plotting ""')
 then all command lines given in `pmode' will be ignored and
 warning messages will be given accordingly.

 Syntax:
     pmode "command(optional)"

 Examples:
     pmode
     pmode set nokey
     pmode plot "fudgfile" with lines

 See also:
     set plotting, set prompt-pm, special

?print

 The `print' command is a C-calculator mode command that writes the
 value of a valid mathematical expression to a file selected by 
 `set output'. The default is "stdout". If there is more than one
 variable, a coma separated list must be given in which case each
 expression value will printed on the same line and separated by a tab.
 As with other number output commands, the output format is the one
 selected by the `set format' command. The default is "% 10.8e".
 The `print' command differs from `show variables' as follows:

   =>  `print' accepts any expression for indexing vector elements;

   =>  `print' requires a comma separated list;

   =>  `print' can be part of a function or procedure;

   =>  `print' can print strings provided they are in double
          quotes. This includes characters '\n', '\t', '\a', ...;

   =>  `print' does not append a newline.

   =>  `print' can print any mathemetical expression.

   =>  `print' is a C-calculator mode command.

 A simpler way to print variables to "stdout" from the C-calculator
 mode is to use the feature that any variable or coma separated list of
 variables given on the command line will be displayed, separated by tabs
 and appended with a newline character. Thus the construction @ifhelp

     set output stdout
     cmode
         print x, y, "n"

 is equivalent to

     cmode
         x, y

 typed in C-calculator mode (it becomes `let x,y' in fitting mode).
 The only difference between `print' and the automatic printing
 feature of C-calculator mode is that (1) `set output' only affects
 `print' command, and that (2) `print' does not automatically
 append a new line character.

 Syntax:
     print "coma-separated-var-list"

 Examples:
     cmode
         print x+2
         print String, x, y, z
         print "Warning aaa", "x = ", x, "n"

 See also:
     cmode, func, C, show table, show variable, math functions, quotes,
           set format, set output

?proc

 The `proc' command is a C-calculator command used to define
 procedures. Procedures differs from functions in the fact that they do
 not return any value. The procedure arguments are passed and referred
 to the same way they are in functions. Keyword `proc' is a
 C-calculator mode command. The `show table' command can be used
 to list all the installed objects at a given time.

 Syntax:
     proc "procedurename"("proto-list(optional)") "cmode-line-statement"

 or

     proc "procedurename"("proto-list(optional)") {
          "cmode-statements"
     }

 Examples:
     # The following example will print the Fibonacci numbers lower than 1000
     cmode
         proc fib(x) { 
             a = 0 
             b = 1 
             while (b < x) { 
                 print b
                 c = b 
                 b += a 
                 a = c 
             } 
             print "n"
         } 
         # The following 'for' loop is equivalent to the preceding fib()
         proc fib2(x) {
             auto a,b,c        # This proc creates no global variable

             for(a=0,b=1;b<x;c=b,b+=a,a=c) {
                 print b
             }
             print "n"
         }
         fib(1000)  # A procedure as called from C-calculator mode.
    fmode
    let fib2(1000)  # A procedure as called from fitting mode.
    # A short example involving a vector
    set data 10
    let proc init(X, x) X=x
    let b=3
    let init(Y, 2/4 + b) # Shows that scalar can also be expressions.

 See also:
     return, cmode, C, func, auto, math, show table, install

?pwd
?cwd

 The `pwd' command prints the name of the working directory on
 the screen.

 Syntax:
     pwd

 See also:
     cd, ls

?quit

 The commands `exit' and `quit' are equivalent and both will
 exit FUDGIT. On exit, all temporary files "/tmp/fudgitPID*" (note
 the wild card) will be erased. Here PID is the current process number.
 Moreover, if a plotting process is active, it will be sent a KILL
 signal. It is therefore a good habit to use the `$Tmp' string
 variable to build your temporary files.

 Syntax:
     quit

 See also:
     cmode, exit

?quotes

 In the fitting mode, single and double quotes serve to indicate that
 all the characters between quotes should be taken as only one word,
 even if there are some blanks (tab or space) among them. The
 difference between single and double quotes is that within the former
 variable expansion (using `$') does not take place whereas it does in
 the latter. Quotes are not recognized between parentheses.

 In C-calculator mode, double quotes serve to indicate a string and
 parsing is done accordingly. As in C, double quotes can be included in
 a string using the `\' operator. Note that C special characters
 as '\n' for a newline, '\a' for a bell, '\t' for
 a tab, and so on, are recognized in a string. Single quotes have no
 special meanings. The only way to pass a `$' without expanding the
 following name is to escape the `$' with a `\'.

 Thus, a null string is given by `''' or `""' in the fitting
 mode and by `""' only in C-calculator mode.

 In pmode, both single and double quotes are freely passed to the
 plotting program. This is valid when trailing commands are are passed
 to `pmode', although FUDGIT implicitly stays in the fitting
 mode. Once again, expansion of a `$' followed by a string can be
 avoided using the escape character, i.e., by typing `\$'.

 See also:
     exec, set plotting, math function scan, print

?read

 The `read' command is used to read data points from a file or from
 standard input. Each column is assigned to a given vector. Vectors not
 already allocated will automatically be. Range of values can be
 specified on any variable using the ["low":"high"] syntax. A
 ``*'' replacing a value will be taken as unexistent. Range of
 lines can be specified on any variable using the {"low":"high"}
 syntax. The last line range given will be the only one in effect. If
 the file name specified is `-' data will be read from the current
 standard input until the keyword `end' is found on a line by
 itself. The `read' - and the `load' commands are recursive
 functions so they can be nested insofar as you can understand what is
 going on. An assignment consists in a vector name and a column number
 separated by a colon. After a file has been successfully read, 
 `read' will put the name of the data file in string constant 
 `ReadFile'.

 Syntax:
     read "filename" "assignment[range](optional){linerange}(optional)" ...

 Examples:
     read file1 X:1[0:*] Y:2
     read file2 TIME:2{100:400}
     read - T:1 VALUE:2
     1    2.3
     2    4.7
     .    .
     .    .
     .    .
     end

 The first form will read positive values of the first column in vector
 X and corresponding values of the second in vector Y. The second
 will read the second column of file "file2" from line 100 to line
 400. The third will read T and "VALUE" from stdin. The
 assignment does not need to be in increasing order of column. Also
 note that the first column is 1.

 See also:
     exec, data

?reinstall

 The `reinstall' command is used to perform the dynamical loading
 of a module that was already loaded. Typically, this is done after a
 module has been modified and recompiled. Refer to `install' for
 more detail.

?return

 The C-calculator `return' keyword is used as in standard C to
 return from a function or a procedure. Contrary to C, `return'
 requires parentheses when returning a value from a function. It is an
 error to return a value from a procedure or to not return anything
 from a function. `return' is a C-calculator mode command.

 Syntax:
     return("expression")
     return
 See also:
     C, cmode, func, proc, auto

?save

 Look under `append' command description.

?set

 The `set' command sets a lot of options, as follows.

?set comment

 The `set comment' command selects the character which will cause
 the rest of the line to be ignored. The default value is ``#''.
 Note that the effect of a comment character will be void if: (1) found
 somewhere between single quotes in fitting mode or (2) escaped with a
 `\'.

 Syntax:
     set comment "character"

 Example:
     set comment ?

 See also:
     show comment, comments

?set data
?data

 The `set data' command changes the effective size of vectors. All
 the vector arithmetic checks for index boundaries. The `data'
 constant is the higher bound of the check and necessarily the size of
 all vectors. Changing the `data' value does not change the values
 nor the capacity of vectors. It only changes the upper bound on the
 value the index can take. The `data' constant is also changed by
 the commands `read' and `exec', which set it to the number of
 valid data points read. Because the upper bound can never be higher
 than the effective capacity of vectors, a `data' value higher than
 the current `samples' value will be refused. See `set
 samples'. Typically, `set data' is used when one wants the
 C-calculator to generate (and plot) vectors. The `read' and 
 `exec' commands take care of adjusting it. `data' constant can also
 be changed from the C-calculator mode if the constant is 
 `unlock'ed. However, no check is made to ensure the given value is not
 higher than `sample' size, in which case a segmentation fault will
 crash the whole program. It is always safer to use `set data'.

 Syntax:
     set data "number"

 Example:
     set data 300

 See also:
     lock, unlock, read, exec, cmode

?set debug
?debug

 The `set debug' command puts the reading of `load'ed files in
 verbose mode, so that debugging is more easily done. All the commands,
 expanded macros and/or string variables are echoed as they are
 executed. There are some different debug levels at the present time:

   =>  0 clear all the debugging states.

   =>  1 echo the expanded lines as they are read. The command is
          parsed and comments are stripped out. This is most useful
          for debugging script files. History substitutions are shown.

   =>  2 display all command lines as they are read from the script.

   =>  3 display the line numbers of the ignored lines as they are
          read from datafiles.

   =>  4 echo command lines as they are passed to the math parser.

   =>  5 turn the math parser debugger on. To use this, the program
          must have been compiled with the YYDEBUG preprocessor
          variable on.

   =>  6 trace the flow of fitting mode `if' constructions.

 Debugging values are not exclusive so that more than one level
 can be turned on. Levels are subject to change.

 Syntax:
     set debug "value-list"

 Example:
     set debug 0 1 3

 See also:
     load

?set error
?error

 FUDGIT allows the user to select among different possible error
 checks to be made on each single mathematical operations. The
 `set error' command will set computational error checks as follows:

   =>  0: clear all computational error check bits.

   =>  1: check for `infinity' values.

   =>  2: check for `not a number' values.

   =>  3: check for `out of domain' math function errors.

   =>  4: check for `out of range' math function errors.

 Error checks are not exclusive and more than one can be specified on
 the command line. The default status has all error check levels
 activated (1 2 3 4).

 It is sometimes desirable to disable one of the checks. For example,
 the operation y = 1/sinh(x) will give a `out of range' error for
 large x ( > 709 on most machines), although y is in fact 0.
 If one uses `set error 0 1 2 3', then no error will be reported
 and y will be set to zero accordingly.

 Syntax:
     set debug "value-list"

 Example:
     set error 0 2 3

 See also:
     C, cmode

?set expand
?expand

 In interactive mode, history expansion and substitution will occur
 only if the `expand' variable is set. It is disabled using 
 `set noexpand'. The default is on.

 Syntax:
     set expand

 See also:
     set noexpand, history, line editing

?set format
?format

 The command `set format' will set the printf format for variables.
 Use only if you are sure of what you are doing. It defaults to
 ``% 10.8e". See `man printf(3)' if in doubt.

 Syntax:
     set format "string"

 Examples:
      set format %6.2lf
      set format "% .8g"

 See also:
     show, append

?set function
?function

 The `set function' command is perhaps the most crucial command in
 data fitting. It is used to select a built-in fitting function or to
 enter a user-defined function. The following fitting functions are
 available:

     NAME          DESCRIPTION             PARAMETERS REQUIRED
     ----          -----------             -------------------
     straight      Straight line           (2 parameters)
     sine          Sine series             (N parameters)
     cosine        Cosine series           (N parameters)
     legendre      Legendre series         (N parameters)
     polynomial    Power series            (N parameters)
     gauss         Gaussian series         (3N parameters)
     expo          Exponential series      (2N parameters)
     user          User-defined function   (N parameters)

 Assume a variable vector X and a parameter vector A then,
 the nonlinear gauss fitting function is a series of gaussians where

    f(X,A) = SUM (i=1,4,7,...,N) of A[i] * exp(-((X - A[i+1])/A[i+2])^ 2).


 The nonlinear expo function is a series of exponentials where

    f(X,A) = SUM (1=1,3,...,N) of A[i] * exp(X*A[i+1]).


 For a user-defined function, the `set function user' will prompt
 for more input. The following input is related to the variable to
 fit. For purposes of clarity, let's say that we have to fit vectors
 `X Y DY'. This requires a fit function `YFIT' (the name is
 made from the dependent variable appended with `FIT') and all the
 partial derivatives `DYFITD1, DYFITD2, ..., DYFITDN' taken with
 respect to the parameters n=1,... N. All these functions are
 defined one per line as in the case of a macro until a `stop' is
 entered. Temporary variables are permitted. `set function user'
 actually defines a C-calculator mode macro that will be executed
 before each iteration of the fit. Therefore the complete C-calculator
 mode grammar is fully supported here. Temporary vectors can thus be
 used to speed up the calculation.

 The C-calculator macro can be a simple call to a predefined procedure.
 When defined so, the parsing does not have to be done at each iteration,
 and a slightly faster process should result.

 Example:
    # read column 1, 2 and 3 of file "file"
    read file T:1 R:2 DR:3
    # make a three parameter fit
    set parameter K 3
    # this is a linear fit; use singular value decomposition
    set method svd_fit
    # enter my function
    set function user
        RFIT = K[1] + K[2]*T^0.5 + K[3]*T^1.5
        DRFITD1 = 1
        DRFITD2 = T^0.5
        DRFITD3 = T^1.5
    stop
    fit T R DR

 The vector `RFIT' will contain the fitted function. The difference
 between the fit and real data can be obtained right away by defining a
 vector

   let RDIFF = R - RFIT

that can be plotted with respect to `T'.

 The same thing is done for nonlinear fit with the exception that the
 partial derivatives of the function with respect to the parameters
 will contain reference to some parameter(s). (This is precisely the
 meaning of nonlinear here).

 There is virtually no restriction on the number of parameters (memory
 is the sole limitation: `set parameter' command allocates a
 matrix of `parameters' X `samples' ). The only conditions are
 that a linear regression must have 2 parameters defined (this is
 obvious) and the built-in nonlinear functions must be modulo 3 for the
 series of gaussians and modulo 2 for the series of exponentials.

 See also:
     fit, set method, adjust, proc, auto

?set input
?input

 The `set input' command selects the file for the input of the
 C-calculator mode `Read' and `vread' command. The string 
 "stdin" is valid as a filename. If the selected file does not exist or
 cannot be read, an error message will be given and the value will go
 back to the default value, which is "stdin".

 Syntax:
     set input "filename"

 See also:
     Read, vread

?set iteration
?iteration

 The `set iteration' command permits the user to change the
 iteration number for the Marquardt-Levenberg nonlinear fitting method.
 See `set function'. The default value is 10. However, the fitting
 process will stop if there is no difference in chi^2 for two
 consecutive iterations. However, a negative value will force to
 iterate up to the absolute value of that number, without checking for
 convergence.

 Syntax:
     set iteration "value"

 Example:
     set iteration 3

 See also:
     fit, set method, set function

?set method
?method

 The `set method' command allows the user to select the fitting
 method to be used when calling the `fit' command. The following
 methods are available:

     NAME         DESCRIPTION
     ----         -----------
     ls_reg       least square linear regression (2 parameters)
     lad_reg      least absolute deviation linear regression (2 parameters)
     ls_fit       general least square linear fit using QR decomposition
     svd_fit      general least square linear fit using singular value
                   decomposition
     ml_fit       general least square nonlinear fit using
                   Marquardt-Levenberg method

 Among them, only `ml_fit' and ls_fit depends on `iteration'
 and `adjust'.

 For all methods except `lad_reg', the value of chi^2 will be put
 in the scalar constant `chi2'. In the case of `lad_reg', %chi^2
 will contain the average absolute deviation.

 Syntax:
     set method "method"

 Example:
     set method svd

 See also:
     fit, set iteration, set function

?set noexpand
?noexpand

 The `set noexpand' command disallows history expansion on the
 interactive command line.

 Syntax:
     set noexpand

 See also:
     set expand

?set output
?output

 The `set output' command selects the file for the output of the
 C-calculator mode `print' command. The strings "stdout" and
 "stderr" are both valid as a filename. If the selected file
 already exists, it will be overwritten with no warning. The default
 value is "stdout".

 Syntax:
     set output "filename"

 See also:
     print

?set pager
?pager

 The `set pager' command allows the user to select a pager. A pager
 is the program that is called when the structure to be displayed has
 more than 24 elements. The default pager is (1) the environment
 variable PAGER if it exists or (2) "/usr/?/more" (path depends on
 system) if not. If `pager' is defined to a null string (`""'),
 then no pager will be used.

 Syntax:
     set pager "string"

 Example:
     set pager "more -c"

 See also:
     show, show pager

?set parameters
?parameters

 The command `set parameters' will fix the parameter name and
 size. Since the set of parameters is a kind of vector, parameter name
 cannot contain lower case letters. Parameters are initialized to
 zero. A built-in scalar constant called `param' contains the
 number of parameters at all time.

 Syntax:
     set parameters "parameter-name" "size"

 Example:
     # set the vector D of size 3 to be determined by the fit.
     set parameters D 3

 See also:
     show parameters, show setup

?set plotting
?plotting

 The `set plotting' command changes the default plotting program
 used by the plotting mode. The default is GNUPLOT but this can be
 changed to any plotting program that can be driven from stdin. A
 maximum of 16 arguments can be passed when the program is first
 called. Changing the plotting program will send a KILL signal to the
 existing plotting program (if any). If the plotting program is set to
 a null string (`""'), FUDGIT will ignore all the plotting
 commands and warning messages will be given. Setting the plotting
 program to a file that cannot be found or executed will result in an
 error at the first `pmode' call.

 Syntax:
     set plotting "command"

 Examples:
     set plotting "/usr/local/bin/sgiplot -p"
     set plotting /usr/local/bin/gnuplot

 See also:
     show plotting

?set prompts
?prompts

 All three FUDGIT prompts can be changed by the `set' command.
 The name of the prompts are:

   =>  `prompt-cm' for the C-calculator mode prompt
          (default: "cmode> ";

   =>  `prompt-fm' for the fitting mode prompt
          (default: "fudgit> ";

   =>  `prompt-pm' for the plotting mode prompt
          (default: "pmode> ".

 A null string `""' (i.e., two consecutive quotes) can be given to
 any of these.

 Syntax:
     set prompt-cm "string"
     set prompt-fm "string"
     set prompt-pm "string"

 See also:
     show prompts

?set samples
?samples

 The command `set samples' changes the current capacity of the
 fitting program. Typically, `samples' is set at the beginning of a
 session since all the existing vectors and variables are erased on
 this call. The default setting is 4000 points.

 Syntax:
     set samples "value"

 Example:
     set samples 6000

 See also:
     set data, cmode, let, lock

?set vformat
?vformat

 The command `set vformat' will set the sprintf format used for the
 expansion of scalar variables by the expansion operator `$'. Use
 only if you are sure of what you are doing. It defaults to ``%.3lg''.
 See `man printf(3)' if in doubt.

 Syntax:
     set vformat "string"

 Examples:
     set vformat %6.2lf
     set vformat "%.4lg"

 See also:
     $, cmode, C

?shell

 The `shell' command starts a shell according to your SHELL
 environment variable. It is equivalent to `system' command. Refer
 to the latter for details.

?show

 The `show' command is used to see the chosen options or to look at
 any defined vectors, parameters or variables.

 See also:
     set, echo

?show comment

 The `show comment' command echoes the current comment escape
 character.

 Syntax:
     show comment

 See also:
     set comment, comments

?show data

 The `show data' command displays the current value of `data'
 constant. Left for compatibility.

 Syntax:
     show data

 See also:
     set data, lock, unlock, set samples

?show debug

 The `show debug' command displays the current value of the 
 `debug' variable. The value is displayed in octal since the
 `set debug' n command turns on the n^th bit of this number.

 Syntax:
     show debug

 See also:
     set debug

?show error

 The `show error' command displays the current value of the 
 `error' computational check variable. The value is displayed in octal
 since the `set error' n command turns on the n^th bit of
 this number.

 Syntax:
     show error

 See also:
     set error

?show input

 The `show input' command shows the filename selected for the
 input of the C-calculator mode `Read' and `vread' command.
 The default value is "stdin"

 Syntax:
     show input

 See also:
     set input, Read, vread

?show iterations

 The `show iteration' command displays the current value of
 `iteration' variable.

 Syntax:
     show iteration

 See also:
     set iteration, set method

?show fit

 The `show fit' command displays the different quantities relevant
 to the current fitting method. Typical examples are chi^2, the 
 covariance matrix, the curvature matrix, correlation factor, etc...

 Syntax:
     show fit

 See also:
     fit, set parameters, set function, set method

?show format

 The `show format' command displays the current value of
 `format' variable. The `format' string is used when
 displaying any number on the screen. Refer to printf(3) of
 the UNiX manual.

 Syntax:
     show format

 See also:
     set format, show

?show function

 The command `show function' displays the current function type. If
 the function type is `user', then the user-defined function will
 be displayed.

 Syntax:
     show function

 See also:
     set function, show setup, fit, math

?show macros

 If called with an argument, the `show macros' command will display
 the specified macro. Otherwise, all currently defined macros will be
 displayed. The selected `pager' is called if the command is given
 in interactive mode (at the command line prompt).

 Syntax:
     show macros "macroname"(optional)

 See also:
     set pager, save macros, alias

?show memory
?memory

 The `show memory' function will display the current state of
 memory consumption of the program. All sizes are given in bytes. It
 uses a direct call to mallinfo(3). The arena is the size of memory
 requested by the process to the kernel. It is then split in different
 blocks shared among the internal matrices and user's vectors, macros,
 functions, procedures, variables and history.

 Syntax:
     show memory

 See also:
     free, show table

?show method

 The `show method' command displays the current value of the
 fitting `method'. It contains "none" by default.

 Syntax:
     show method

 See also:
     set method, fit, set function

?show output

 The `show output' command shows the filename selected
 for the output of the C-calculator mode `print' command. 
 The default value is "stdout"

 Syntax:
     show output

 See also:
     set output, print

?show pager

 The `show pager' command displays the current value of the
 `pager' program. 

 Syntax:
     show pager

 See also:
     set pager, environment, show

?show parameters

 The command `show parameters' will display the parameter values on
 the screen. If the number of parameters is larger than 24, then the
 selected `pager' will be called if the command is given in
 interactive mode (at the command line prompt). As with `append'
 and `save parameters', `show parameter' can accept optional
 variable or constant (either string or scalar) list of names, in which
 case the value of the given variables will be displayed along with the
 parameter values.

 Syntax:
     show parameters "variable-list(optional)"

 See also:
     set pager, set parameters, save parameters, show fit

?show plotting

 The `show plotting' command displays the current value of the
 `plotting' program. 

 Syntax:
     show plotting

 See also:
     set plotting, startup, pmode

?show prompts

 The `show prompts' command displays the current values of the
 different mode `prompts'. 

 Syntax:
     show prompt-cm
     show prompt-fm
     show prompt-pm

 See also:
     set prompt, startup

?show samples

 The `show samples' command displays the current value of the 
 `samples' variable. Recall that although `data' is responsible for
 the visible part of all vectors, vectors all have a fixed allocated
 length of `samples' long. Any change to `samples' through 
 `set samples' frees all the existing vectors.

 Syntax:
     show samples

 See also:
     set samples, set data, cmode

?show setup
?setup

 The command `show setup' will show some values of the program,
 such as the last data filename read, the number of data points, 
 current capacity, current comment character, current iteration number,
 current plotting program, etc. Left for compatibility.

 Syntax:
     show setup

 See also:
     set comments

?show table
?table

 The command `show table' displays the current lookup table of the
 C-calculator mode parser. It shows all current variables, numbers,
 vectors and functions included in the internal table. It also shows
 the state of the internal machine (C interpreter), stack and frame
 used in the C-calculator. This is used mainly for debugging or to
 prevent stack or machine code overflow.

 Syntax:
     show table

 See also:
     free, show memory, cmode

?show variables

 Any constants or variables can be displayed on the screen. 
 The `show variable' command differs from `print' as follows:

   =>  `show variables' only accepts integers for indexing
          vector elements;

   =>  `show variables' requires a blank separated list;

   =>  `show variables' cannot be part of a function or procedure.

 As it has been mentioned previously, this is due to the different
 types of parsing between the C-calculator and fitting modes. As with
 all other number displaying commands, the printing format is
 always the one selected by the `set format' command.

 Syntax:
     show variables "variable-list"

 Example:
     show variables x X[2] Y[2] DY[2] time

 See also:
     print, save variables, show table, show vectors, cmode

?show vectors

 Any vector or number of vectors can be seen on the screen. If the size
 of vectors is larger than 24, the selected `pager' will be called
 if the command is given in interactive mode (at the command line
 prompt).

 Syntax:
     show vectors "VECTOR-list"

 Example:
     show vectors X Y DY

 See also:
     set pager, append vectors, read, cmode, let

?show vformat
?vformat

 The command `show vformat' will display the printf format used
 for the expansion of scalar variables by the expansion operator `$'.
 Refer to the printf(3) description in the UNiX manual for more details.

 Syntax:
     show vformat

 See also:
     $, cmode, C, set vformat

?smooth

 The `smooth' command uses a gaussian windowing function (low-pass
 filter) on a Fourier transform loop in order to smooth the given
 vector. The windowing function is exp(-(f/(sigma X f_max))^2)
 where f_max is equal to half of the smallest power of 2 larger
 than the number of data points `data'. Variable f is the
 frequency that ranges from 0 to f_max. More likely, the smoothing
 factor is a non null positive real number from the (0, 1] interval. A
 smoothing factor sigma >= 1 leaves the vector unchanged.

 The number of data points `data' needs not to be a power of 2.

 To be used with discernment!

 Syntax:
     smooth "sigma" "in-VECTOR" "out-VECTOR"

 See also:
     fft, invfft, cmode, C

?special

 The following special commands are left for debugging or macro purposes.
 They start with an underscore to avoid mistakes and remind of their
 special character.

 `_killplot' will kill the current plotting program.

 Syntax:
     _killplot

 `_dumplot' will send the following vectors in the plotting
 program pipe. This is only useful if the current plotting program
 accept data from its stdin. `_dumplot' can accept up to 16
 arguments.

 Syntax:
     _dumplot "VECTOR-list"

 Example:
     _dumplot X Y DY

 See also:
     macro, show macros, plot

?spline

 The `spline' function initializes the internal table for the
 calculation of interpolated values using cubic spline method.
 Interpolated values are obtained from calls to the C-calculator math
 function `interp()'. The value of the first derivative at the
 first and last data points can be specified by optional arguments. If
 not specified, or if one of the optional arguments is an asterisk `*',
 then a "natural" cubic spline is assumed in which case the
 interpolated curve is such that the second derivative at the extreme
 points (or one of them) is null. The asterisk is more likely to be used
 in cases where the user would like to specify the first derivative at
 the last point only. The independent vector must be such that its
 value increases monotonically.

 Syntax:
     spline "indep-VECTOR" "dep-VECTOR" "y1(optional)" "yn(optional)"

 Example:
     # Read vectors having a functional relation Y = F(X) from file "datafile"
     read datafile X:1 Y:2
     # Initialize the spline (as being natural)
     spline X Y
     # Save extreme values
     let from = X[1]; to = X[data]
     # Say there were data=10 points and you want 100
     set data 100
     # Rebuild X vector
     # First build X ranging [0, 1]
     let x=0; X=x++; tmp=data-1; X/=tmp
     # Then from 'from' to 'to': from + (to - from)*X
     let tmp=(to-from); X = from + X*tmp
     # Rebuild Y vector possibly containing original values as a subset
     let Y = interp(X)
     # Note that any value can be asked for
     let interp(2.34*pi)

 See also:
     math interp

?startup

 If a file ".fudgitrc" exists in your home directory, it will be
 automatically loaded at startup time of the program. This is useful if
 one wants to include his own macros or have his own preferences loaded
 to FUDGIT. This file is loaded for both interactive use (`fudgit')
 and batch use (`fudgit "script1" "script2"...').

 Examples:
     set plotting /usr/local/bin/sgiplot
     set prompt-pm ""
     set comment ?
     set samples 10000

 A file called ".hist_fudgit" is will be created in your home
 directory in order to keep history between calls of FUDGIT. The
 number of events is determined at compilation time and defaults to 52.

 See also:
     environment, alias, set plotting, set prompt

?stop

 The command `stop' is used to terminate a macro or a fitting
 function defined by the user. However, it can also be used in a script
 file in order to stop execution at a certain point. In this case, an
 warning message will report that `stop' is being used outside a
 macro or function and the file from which the command was found will
 be considered as at the end of file (EOF).

 See also:
     macro, set function

?strings
?Strings

 FUDGIT has a set of functions returning string objects. These are
 made available to deal with filename construction, or to read from
 standard input. To be consistent with string type, string functions
 are named with both lower case and upper case letters.

 Strings can be added or subtracted in the C-calculator mode. String
 addition "s1" + "s2" simply concatenates strings "s2" to
 string "s1". String subtraction "s1" - "s2" removes 
 "s2" from the end of "s1". Note that the wild card `?' is supported
 in string subtractions.

?Strings DirName
?strings DirName
?DirName
 The string function `DirName' returns the directory name as extracted
 from the filename given as an argument.

 Syntax:
     Dirname("String")

 See also:
     string functions FileName, Scan, Read

?strings FileName
?Strings FileName
?FileName

 The string function `FileName' strips the leading directory names
 of the filename given as an argument. Note that the UNiX command:

     basename "File" "Extension"

 is equivalent to the FUDGIT command:

     FileName("File") - "Extension"

 so that filename constructions can be made in `foreach' loop
 for example.

 Syntax:
     FileName("String")

 Examples:
     foreach File in ls /usr/machin/data/*.32
         read $File X:1 Y:2{2:23}
         # Some commands
         .
         .
         # let File be the filename only, less the ".32" extension
         let File = FileName(File) - ".32"
         # And let Dir be the directory name
         let Dir = DirName(File)
     end

 See also:
     foreach, string functions, cmode, $

?strings Read
?Strings Read
?Read

 The `Read' function read a line from the file chosen by the 
 `set input' function, strips the newline character and returns the
 resulting string. If the input is `stdin', the user will be
 prompted by a "?" and the program will stop until a non-null
 string is entered. This is most likely to be used in macros requiring
 some input during run time. The `Read()' function can be used to
 read numbers with the help of `scan()'. See the example below.

 `Read' can also be used to build vectors by taking one every n
 points. This can be done by two imbedded `for' loops.

 Note: The newline character is not passed to the string.

 Examples:
    # Read a string from stdin (the default)
    set input stdin
    let String = Read()
    # How to get a value out of a string: equivalent to vread()
    let value = scan(Read(), "%lf")
    # How to skip lines in a file
    # Read say file project/numbers.data
    set input project/numbers.data
    cmode
    for (i=1; i<=top; i++) {
        Line = Read()                            # Read one line
        X[i] = scan(Line, "%lf");                # get first column
        Y[i] = scan(Line, "%*lf %*lf %lf");        # get third column
        for (j=1; j<n; j++) {
            Line = Read()                        # Read n-1 lines
        }
    }
    fmode
    set input stdin

 See also:
     set input, math function scan, string functions, $

?strings Scan
?Scan

 `Scan("String, Format")' function returns a string as
 extracted from string "String" and according to string format 
 "Format". The format is built with the same rules `sscanf' uses.
 See man pages on `scanf(3)'. Note that the format must contain one
 active "%s" or "%[]" construction. An example might be of some help
 here, especially to show how to use `Scan' in conjunction with
 C-calculator mode defined strings. `Scan' is particularly helpful
 to extract parts of filenames. Recall that strings are defined by
 double quotes as in standard C.

 Knowing about the "%[ ]" scanf(3) construction might be useful at
 this point. Consider the following few examples: `"%[a-zA-Z]"'
 means to read the longest string matched so that it is composed of any
 letter; `"%[^0-9]"' means to read the longest string matched
 so that it is NOT composed of any digit; `"%[^_.]"' means to
 read the longest string matched so that it is not composed of
 characters `_' or `.'.

 Examples:
     # define a string called Testname
     let Testname = "dummy25.dat"
     # Read until a point is encountered
     let Base = Scan(Testname, "%[^.]"))

 See also:
     $, scan, string functions Read, DirName, FileName, C, cmode, quotes

?system

 When called with arguments, the `system' command is equivalent to
 the `!' bang operator, so the remainder of the line will be given to a
 Bourne shell for execution. If `system' has no argument, a shell
 (depending on environment variable SHELL) will be started.

 Syntax:
     system "shell-commands(optional)"

 See also:
     environment, !, shell

?then

 The `then' keyword is required in the fitting mode `if'
 constructions. Refer to the latter for details.

?unalias

 The `unalias' command unaliases any alias previously assigned
 by the `alias' command.

 Syntax:
     unalias "alias_name"

 Examples:
     unalias date
     unalias gnuplot

 See also:
     &, alias, macro, unmacro, append, show

?unlock

 The `unlock' command changes a constant into a variable and thus
 allows the user to change its value. This is particularly useful in
 functions and procedures needing to change the value of the `data'
 constant. Unlocking `data' gives the user complete freedom on the
 effective size of vectors. No check is done on `data'
 assignments, and therefore assigning a value to `data' that is
 superior to `samples' will result in a program crash. For this
 reason, it is always safer to change `data' using the `set
 data' command. It is not an error to unlock a variable. A warning
 message will be given though. However, trying to unlock something else
 than a constant or variable will result in an error.

 See also:
     lock, set data, set samples, cmode

?unmacro

 The `unmacro' command is the counterpart of `macro'. It is
 used to undefine macros. As does `free', `unmacro' accepts the
 ``@all'' string in which case all the macros will be erased and freed
 from memory.

 Syntax:
     unmacro "macro-list" 
     unmacro @all

 Examples:
     unmacro myplot
     unmacro @all

 See also:
     alias, unalias, append, show

?version

 The `version' command displays the version number and the
 welcoming message of FUDGIT.

?vi
?editor

 The command `vi' calls the editor. It is equivalent to
 `!vi filename'. Note that wild cards are also recognized and
 expanded.

 Syntax:
     vi "argument-list"

 Examples:
     vi file
     vi test.*

 See also:
     !, system, alias, shell

?while
?loop

 The `while' command allows the user to construct controlled loops
 on a series of operations. However, FUDGIT supports two kinds of
 `while' constructions, one in the fitting mode and the other in
 the C-calculator mode.

?while cmode_style

 The C-calculator mode while construction has a syntax similar to
 that of standard C. In interactive mode, any new input line will be
 prompted with a ``n{... n\t'' where `n' stands for the nesting
 level and `\t' for a tab. Recall that "cmode-line-statement"
 means a string of semicolon separated C-calculator mode commands.

 Syntax:
     while ("conditions") "cmode-line-statement"

 or

     while ("conditions")
         "cmode-line-statement"

 or
     while ("conditions") {
          "cmode-statements"
     }

 Example:
     read file X:1 Y:2
     let x=12;sum1=0;sum2=0
     while (x++<=100) {
         sum1 += X[x]+Y[x]
         sum2 += X[x]^2 + Y[x]^2
     }

 See also:
     C, break, continue, cmode, for, func, proc, auto, math

?while fmode_style

 The fitting mode `while' is very similar to the C-shell
 `while'. As for the `if' construction, the difference remains
 in a broader range of operators available to the conditional
 statement and the fact that the variable expansion operator `$' is
 not required.

 As for the `foreach' construction, a `end' keyword is
 required to indicate the end of the loop. Note that
 "fmode-statements" can also contain C-calculator mode commands
 (including cmode `while' loops!). Recall that the conditional
 statement is a C-calculator mode expression.

 Syntax:
     while("conditions")
          "fmode-statements"
     end

 See also:
     foreach, if, cmode

?Credits

 Parts of FUDGIT were built from some existing facilities. I would
 like to give full credits for ideas or even segments of code that
 have been taken from the following sources.

   =>  For parts of the user interface:
 
    The help facility and the line editor were taken and adapted
    respectively from GNUPLOT, and READLINE. READLINE was written by
    Brian Fox. The help facility is originally from John D. Johnson.

   =>  For the C-calculator:
 
    The calculator is inspired from HOC calculator which was debugged
    and largely augmented to support vector algebra, memory management
    and extra operators. The source of the basic program is reproduced
    for educational purposes in "The Unix Programming Environment"
    by Brian W. Kernighan and Rob Pike, Prentice-Hall (1984).

   =>  For the fitting functions:

    Some of the included fitting routines are based on the algorithms
    found in chapter 14 of "Numerical Recipes in C" by W. H. Press,
    B. P. Flannery, S. A. Teukolsky and W. T. Vettering, Cambridge
    University Press (1988), of which some were in turn adapted from
    LINPACK. I had to adapt all the algorithms to include elegant error
    recovery and to perform all calculations on vectors outside the
    fitting loops, since their implementation would not permit the use
    of run-time user selectable functions. FUDGIT would not have been
    possible without the valuable help of that book. Since I strongly
    recommand that you have a copy of a fitting book, I strongly
    suggest you have a copy of this one. Not only this book will
    describe all the methods used in FUDGIT but it will also give you
    unvaluable insights to get to the state of the art of fitting.
    These routines are copyrighted and cannot be separated from FUDGIT.

    Copyright (C) 1987, 1988 Numerical Recipes Software. Reproduced by
    permission from the book "Numerical Recipes: The Art of
    Scientific Computing" published by Cambridge University Press.

   =>  For the fft routine:
    
    The fft routine was first derived from an original Pascal version
    written in "Simple Calculations with Complex Numbers" by David
    Clark in DDJ 10/84 and then translated to C by R. Hellman
    (02/21/86). I rewrote the C version to use vectors of alternated
    double real and imaginary instead of the original (slow) vector of
    pointers to complex structures. I also merged all functions in one
    to prevent unnecessary function calls. I was astonished to see that
    the resulting version was almost identical to the one found in
    Numerical Recipes with the exception of a trigonometric recursion.
    In conclusion, given an algorithm, I think that is is a hard task
    to try to write a code much better than the one found there. This
    is normal since the space of code possibilities gets narrower as
    the constraints (optimization) acting on an implementation in a
    given language are increased. The one included is from N.R. which
    was adapted from N. Brenner.

    Copyright (C) 1987, 1988 Numerical Recipes Software. Reproduced by
    permission from the book "Numerical Recipes: The Art of
    Scientific Computing" published by Cambridge University Press.

   =>  For the help file:

    I would like to thank Ross Thompson for proofreading part of the
    documentation and also for giving me constructive feedback in
    course of the program development.

   =>  For the IRIX supported dynamic loading package:

    The `dl' Dynamic Loading package is from Jack Jansen
    <Jack.Jansen@cwi.nl> from the "Centrum voor Wiskunde en
    Informatica". This package only works on IRIX for now.

   =>  For the SUNOS, ULTRIX supported dynamic loading package:

    The implementation of `dl' for SUNOS and ULTRIX is from Guido
    van Rossum <guido@cwi.nl> from the "Centrum voor Wiskunde en
    Informatica". I modified part of it to allow multiple routine
    loading from the same object. The other part of the puzzle is the
    `dld' loader from Wilson Ho <how@cs.ucdavis.edu> which is based
    on GNU ld(1) and is under GNU license.

   =>  For compilers not having `alloca()':

    The included public domain version of alloca is from D. A. Gwyn.

   =>  For systems not having `putenv()':

    The version of putenv was adapted from Dave Taylor's elm who
    adapted it from cnews.

   =>  For a lot of ideas:

    Many thanks to Steve Hornes for stimulating e-mail discussions.
    Steve is responsible for the idea of implementing dynamic loading
    in the final development of FUDGIT.

   =>  For the port to linux:

    The port to linux was made by Thomas Koenig
    <ig25@rz.uni-karlsruhe.de>.  Thanks a lot Thomas!

   =>  For the rest of the code:

    Copyright (C) 1993   Martin-D. Lacasse
  
    See the Copyrights file for more detail, or the `README' help topic.

    Permission to use, copy, and distribute this software and its
    documentation for any peaceful purpose and without fee is hereby
    granted, provided that the above notices appear in all copies and
    that both those notices and this permission notice appear in
    supporting documentation. No part of this can be used for commercial
    purposes.

    Send bugs, comments or suggestions to

            <isaac@physics.mcgill.ca>.

 Disclaimer:

    This software is provided "as is" without express or implied warranty.


?README

 The eclectic nature of FUDGIT makes it a borderline program.
 I would like to include the following comments on copyrights.

 Some Definitions
 ================

 Scientific Community: By "Scientific Community" is meant the whole of
 researchers in industries, universities and government agencies,
 students as well as individuals making research on their own.

 Software Developers: By "Software Developers" is meant all the
 people writing or selling softwares with commercial purposes.

 Publicly Available Sources: By "Publicly Available Sources" is meant
 the pool of all source codes, algorithms and ideas that can be found
 in publicly available scientific journals, publicly available
 educational books, public domain source code, algorithms and ideas, or
 source code, algorithms and ideas aimed at helping the Scientific
 Community. That is, Publicly Available Sources are publicly available
 material that can be part of higher education programs.

 Progressive Copyrights
 ======================

 People involved in the Scientific Community need access to the most
 recent developments to continue their own research. At the same time,
 the same people share their most recent discoveries to the rest of the
 Scientific Community. Standard Copyrights apply to the commercial
 applications of their results.

 Source code raised the problem that for the first time, the research
 product is deeply involved in the research process. For the first
 time, restrictions are being put on a fundamental research tool. This
 is why the traditional way of thinking about Copyrights is obsolete.

 Copyrights were implemented with the idea of protecting the commercial
 interests of the owner. Thus, standard Copyrights give the owner the
 exclusive rights of having commercial applications of his/her
 implementation of an idea. However, some copyrights are much too
 restrictive concerning scientific applications. Sometimes, old ideas
 suddenly become protected because a group of Software Developers
 claim Copyrights for pieces of code that freely existed for a long
 time already, or are the straightforward representation of a given idea
 in a given language. This simply does not make sense, and for this
 reason, more realistic copyright procedures are required.

 Progressive Copyrights should give the author the full commercial
 rights but leaves the Scientific Community the right to use the source
 code, the algorithms and the ideas for their own purposes. It does not
 allow Software Developers to use the source code for commercial
 applications, but they are free to use the ideas included in the
 source code. In other words, Software Developers cannot make money
 from the hard work someone else did. If they want to do so, they have
 to agree with the people owning the Progressive Copyrights.

 Progressive Copyrights should apply to all Publicly Available Sources.
 From the moment a source code is published and becomes publicly
 available, the people from the whole Scientific Community have the
 right to use it for their own purposes.

 A small group of Software Developers cannot claim Copyrights for
 source code, algorithms and ideas that have been around for a while,
 tested, improved, and commented by the whole Scientific Community.

 A University student cannot start in life with a Bachelor degree and
 have part of it copyrighted!

 Progressive Copyrights are more a way of thinking than a legal
 matter. I just hope that this way of thinking will spread in the
 Scientific Community. After all, Progressive Copyrights are just
 common sense! Despite Software Developers lobby.