2.4 Parameters and states

To see the parameter values of a star or galaxy, execute the Edit:edit-params command, which has the accelerator key "e". The singen star in the sinMod application has the following parameter screen:

Notice that the frequency parameter is given as an expression, "PI/100" (PI represents the constant ). This section describes the expression language for specifying parameter values.

The parameter screen can be kept open while you experiment with different values of the parameters. Try changing the value "PI/100" to "PI/200". Click "Apply" in the parameter window, and then "GO" in the run control panel. How does this change the display? Clicking "Cancel" in the parameter window will restore the parameter values to the last saved values and dismiss the parameter window. Clicking "Close" will dismiss the parameter window without restoring the parameter values.

This block, however, is a galaxy, not a star. If you look inside (as has been done in figure 2-4), and edit the parameters of the singen block inside modulator, you will see

Notice now that the value of the frequency parameter is a symbolic expression, "freq". This refers to the galaxy parameter "freq". Thus, parameter values can be passed down through the hierarchy. These symbolic references can appear in expressions, which we discuss next.

Parameter expressions

Parameter values set through pigi can be arithmetic expressions. This is particularly useful for propagating values down from a universe parameter to star parameters somewhere down in the hierarchy. An example of a valid parameter expression is:

PI/(2*order)

where order is a parameter defined in the galaxy or universe. The basic arithmetic operators are addition (+), subtraction (-), multiplication (*), division (/), and exponentiation (^). These operators work on integers and floating-point numbers. Currently all intermediate expressions are converted to the type of the parameter being computed. Hence, it is necessary to be very careful when, for example, using floating-point values to compute an integer parameter. In an integer parameter specification, all intermediate expressions will be converted to integers.

Complex-valued parameters

When defining complex values, the basic syntax is

(real, imag)

where real and imag evaluate to integers or floats.

Fixed-point parameters

Fixed-point parameters may be assigned a precision directly. To do this, the parameter is given in the syntax "(value, precision)", where value is an ordinary number and precision is given by either of two syntaxes:

Syntax 1: As a string like "3.2", or more generally "m.n", where m is the number of integer bits (to the left of the binary point) and n is the number of fractional bits (to the right of the binary point). Thus length is m+n.
Syntax 2: A string like "24/32" which means 24 fraction bits from a total length of 32. This format is often more convenient because the word length often remains constant while the number of fraction bits changes with the normalization being used.

In both cases, the sign bit counts as one of the integer bits, so this number must be at least one.

Thus, for example, a fixed-point parameter might be defined as "(0.8, 2/4)." This means that a 4-bit word will be used with two fraction bits. Since the value "0.8" cannot be represented precisely in this precision, the actual value of the parameter will be rounded to "0.75".

A fixed-point parameter can also be given a value without a precision. In this case, the default precision is used. This has a total word length of 24 bits with the number of integer bits set as required to store the value. For example, the number 1.0 creates a fixed-point object with precision 2.22, and a value like 0.5 would create one with precision 1.23.

The precision of internal computations in a star is typically given by a parameter of type precision. A precision parameter has a value specified using either of the two syntaxes above.

2.4.3 Reading Parameter Values From Files

The values of most parameter types can be read from a file. This syntax for this is to use the symbol < as in the following example:

< filename

First, any parameters appearing in the filename in the form of {parameter} are replaced with their values. Then, any references to environment variables or home directories are substituted to generate a complete path name. Finally, the contents of the file are then read and spliced into the parameter expression and reparsed. File inputs can be very useful for array parameters which may require a large amount of data. Other expression may come before or after the <filename syntax (any white space that appears after the < character is ignored).

2.4.4 Inserting Comments in Parameters

Comments are also supported for non-string parameters. A comment is specified with the # symbol. Everything after the # until the end of the line is discarded when the parameter is evaluated. Comments are especially useful in combination with files as they can help remind the user of which galaxy or star parameter the file was written.

For example, a comment could be added to the frequency parameter above:

freq # This is set to the Galaxy parameter

Comments are not supported for the String parameter or String Array parameter types. In fact, when the image processing stars use String states to represent a filename, the # character is used to denote the frame number of the image being processed.

2.4.5 Using Tcl Expressions in Parameters

Arbitrary Tcl expressions can be embedded in a parameter expression by preceding the expression with the ! character as in the following example:

! "expression"

First, parameters in the form of {parameter} appearing in the expression are replaced by their values. Then, the string is sent to the pigiRpc Tcl interpreter for evaluation. Finally, the result is spliced into the parameter expression and reparsed. The pigiRpc Tcl interpreter is the same interpreter that appears as a window when pigi is started by using pigi -console.

This facility is general and supports both numeric and symbolic computing of expressions. Through Tcl, one can access all of its math functions, which generally behave as the ANSI C functions of the same name: abs, acos, asin, atan, atan2, ceil, cos, cosh, double, exp, floor, fmod, hypot, int, log, log10, pow, round, sin, sinh, sqrt, tan, and tanh. So, a parameter expression could be

! "expr sqrt(2.0 / {BitDuration})"

for the amplitude of the oscillators in a binary frequency shift keying system, in which BitDuration is a parameter. The expr command is a Tcl command that treats its arguments as a single mathematical expression that must evaluate to a number.

The Tcl mechanism can be used to return symbolic expressions:

! "join 2*gain1"

Because gain1 is not surrounded by curly braces, its value is not substituted before passing the expression to the Tcl interpreter. The Tcl interpreter will return 2*gain1 which is then evaluated by the parameter parser.

Note that whitespace between ! and " is permitted in numeric parameters, but not in string parameters: to get a Tcl call to be recognized in a string parameter you must write:

!"list /users/ptolemy/myfile"

There are several Tcl commands embedded in pigiRpc that help support parameter calculations. They are: listApplyExpression, max, min, range, rangeApplyExpression, and sign. For example,

! "min [max 1 2 3] [sign -2]"

first evaluates to min 3 -1 and then to -1. The procedure range returns a consecutive sequence of numbers:

! "range 0 5"

returns 0 1 2 3 4 5. The rangeApplyExpression procedure generates a sequence of values by applying a consecutive sequence of numbers to a Tcl expression that is a function of i. For example, you can generate the taps of an FIR filter that is a sampled sinusoid by using

! "rangeApplyExpression { cos(2*{PI}*$i/5) } 0 4"

generates one period of sinusoidal function and returns

1.0 0.309042 -0.808986 -0.809064 0.308916

The listApplyExpression is similar to rangeApplyExpression except that it only takes two arguments: the second argument is a list of numbers to substitute for i in the expression. The command

! "listApplyExpression { cos(2*{PI}*$i/5) } [range 0 4]"

is equivalent to the previous example of the rangeApplyExpression function.

If you are running Tycho TclShell from within pigi or pigi -console, you can receive help on the new Tcl procedures listApplyExpression, max, min, range, rangeApplyExpression, and sign, by typing

help sign

at the prompt. To start Tycho from within pigi, type a y while the mouse is over a vem facet or palette.

The Tycho TclShell and the pigiRpc console includes the Ptolemy interpreter (ptcl) which defines the help mechanism. Help is available on all of the commands we have added to the Tcl language.

2.4.6 Using Matlab and Mathematica to Compute Parameters

Since Tcl can be used to compute parameters as described in the previous section, Ptolemy's Tcl interface to Matlab [Han96] and Mathematica [Wol91][Bla92] can be used to compute parameters. This allows even more expressiveness, but the drawback is that demonstrations relying on Matlab and Mathematica will only work at sites that have Matlab and Mathematica installed. For example, we can use Matlab to design an 32-order FIR half-band filter using the Parks-McClellan optimal equiripple FIR filter design algorithm:

! "matlab getpairs c {c=remez(32, [0 0.4 0.6 1], [1 1 0 0])}"

! "mathematica get c {c=Integrate[A x, {x, 0, 1}]}"

This command returns the symbolic expression A/2 which is reparsed by Ptolemy. Matlab and Mathematica can be used to keep track of how parameter values are computed. Mathematica can also be used to return symbolic expressions that can be used in conjunction with higher-order functions to define scalable systems [Eva95].

The Ptolemy interface to Matlab and Mathematica can also be accessed from the pigiRpc console window, and the Tycho editor offers console windows that mimic the Matlab and Mathematica teletype (tty) interfaces. More information about the options of the Tcl commands matlab and mathematica can be found by using the help facility described above.

2.4.7 Array parameters

When defining arrays of integers, floats, complex numbers, fixed-point numbers, or strings, the basic syntax is a simple list separated by spaces. For example,

1 2 3 4 5

defines an integer array with five elements. The elements can be expressions if they are surrounded by parentheses:

1 2 PI (2*PI)

Repetition can be indicated using the following syntax:

value[n]

where n evaluates to an integer. An array or portion of an array can be input from a file using the symbol < as in the following example:

1 2 < filename 3 4

Here the first two elements of the array will be 1 and 2, the next elements will be read from file filename, and the last two elements will be 3 and 4. This latter capability can be used in combination with the WaveForm star to read a signal from a file.

2.4.8 String Parameters

There is a bit of complication when one wishes to set a string parameter or string array parameter equal to the value of a galaxy or universe parameter. This is because a distinction must be made between a sequence of characters that give the name of a symbol and a sequence of characters to be interpreted literally. The syntax to use is explained in the example:

This string has the word {word} taken from another parameter

Here {word} represents the value of a string universe or galaxy parameter. This capability is especially useful for constructing labels for output plots. When using string states to specify options for a Unix command, as in the options parameter in Xgraph stars, you can use either double quotes or single quotes to include white space within a single word:

-0 'original signal' -1 'estimated signal'

String arrays have a few more special restrictions. Each word (separated by white space) is a separate entry in the array. To include white space in an element of the array, use quotation marks. Thus, the following string array

first "the second element" third

has three elements in it. The string array

repeat[10]

has ten separate copies of the string "repeat" in 10 separate entries in the array. Curly braces are used to substitute in values from galaxy parameters. Thus, in

{paramname}

paramname must be the name of either a string array or a scalar-valued parameter (an integer, float or complex array, for example, is not permitted). If it is a string array, then each element of paramname becomes an element of the parameter. If it is some other kind of parameter the value becomes a single element of the string array.

To use one of [, ], {, or } literally, quote them with double quotes. To turn off the special meaning of a double quote, precede it with a backslash: \". Similarly, use \\ to get a single backslash.

String array values may also be read from files using the < symbol. For details on how to use file references, see section 2.4.3 above. Note that for string arrays, the filename can be a
literal string such as

< $PTOLEMY/data/filename

as well as a string that refers to parameters such as

< $PTOLEMY/{data_dir}/data_file

in which case the value of the parameter data_dir would be substituted. Ptolemy does not perform expansion of filenames such as file.{1,2} into file1 file2 as a Unix shell might do.