Abstract

This document provides a tutorial for the Next Scripting Language NX.

The Next Scripting Language (NX) is a highly flexible, Tcl [Ousterhout 1990] based object oriented scripting language. It is a successor of XOTcl 1 [Neumann and Zdun 2000a] and is based on 10 years of experience with XOTcl in projects containing several hundert thousand lines of code. While XOTcl was the first language designed to provide language support for design patterns, the focus of the Next Scripting Framework and NX are on combining this with Language Oriented Programming. In many respects, NX was designed to ease the learning of the language by novices (by using a more mainstream terminology, higher orthogonality of the methods, less predefined methods), to improve maintainability (remove sources of common errors) and to encourage developer to write better structured programs (to provide interfaces) especially for large projects, where many developers are involved.

The Next Scripting Language is based on the Next Scripting Framework which was developed based on the notion of language oriented programming. The Next Scripting Frameworks provides C-level support for defining and hosting multiple object systems in a single Tcl interpreter. The whole definition of NX is fully scripted (e.g. defined in nx.tcl). The Next Scripting Framework is shipped with three language definitions, containing NX and XOTcl 2. Most of the existing XOTcl 1 programs can be used without modification in the Next Scripting Framework by using XOTcl 2. The Next Scripting Framework requires Tcl 8.5 or newer.

1. NX and its Roots

Object oriented extensions of Tcl have quite a long history. Two of the most prominent early Tcl based OO languages were incr Tcl (abbreviated as itcl) and Object Tcl (OTcl [Wetherall and Lindblad 1995]). While itcl provides a traditional C++/Java-like object system, OTcl was following the CLOS approach and supports a dynamic object system, allowing incremental class and object extensions and re-classing of objects.

Extended Object Tcl (abbreviated as XOTcl [Neumann and Zdun 2000a]) is a successor of OTcl and was the first language providing language support for design patterns. XOTcl extends OTcl by providing namespace support, adding assertions, dynamic object aggregations, slots and by introducing per-object and per-class filters and per-object and per-class mixins.

XOTcl was so far released in more than 30 versions. It is described in its detail in more than 20 papers and serves as a basis for other object systems like TclOO [Donal ???]. The scripting language NX and the Next Scripting Framework [Neumann and Sobernig 2009] extend the basic ideas of XOTcl by providing support for language-oriented programming. The the Next Scripting Framework supports multiple object systems concurrently. Effectively, every object system has different base classes for creating objects and classes. Therefore, these object systems can have different different interfaces and can follow different naming conventions for built-in methods. Currently, the Next Scripting Framework is packaged with three object systems: NX, XOTcl 2.0, and TclCool (the language introduced by TIP#279).

Languages
Figure 1. Language History of the Next Scripting Language

The primary purpose of this document is to introduce NX to beginners. We expect some prior knowledge of programming languages, and some knowledge about Tcl. In the following sections we introduce NX by examples. In later sections we introduce the more advanced concepts of the language. Conceptually, most of the addressed concepts are very similar in XOTcl. Concerning the differences between NX and XOTcl, please refer to the "Migration Guide for the Next Scripting Language".

2. Introductory Overview Example: Stack

A classical programming example is an implementation of a stack, which is most likely familiar to many readers from many introductory programming courses. A stack is a last-in first-out data structure which is manipulated via operations like push (add something to the stack) and pop remove an entry from the stack. These operations are called methods in the context of object oriented programming systems. Primary goals of object orientation are encapsulation and abstraction. Therefore, we define a common unit (a class) that defines and encapsulates the behavior of a stack and provides methods to a user of the data structure that abstract from the actual implementation.

2.1. Define a Class "Stack"

In our first example, we define a class named Stack with the methods push and pop. When an instance of the stack is created (e.g. a concrete stack s1) the stack will contain an instance variable named things initialized with the an empty list.

Listing 2: Class Stack

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
nx::Class create Stack {

   #
   # Stack of Things
   #

   :variable things {}

   :public method push {thing} {
      set :things [linsert ${:things} 0 $thing]
      return $thing
   }

   :public method pop {} {
      set top [lindex ${:things} 0]
      set :things [lrange ${:things} 1 end]
      return $top
   }
}

Typically, classes are defined in NX via nx::Class create followed by the name of the new class (here: Stack). The definition of the stack placed between curly braces and contains here just the method definitions. Methods of the class are defined via :method followed by the name of the method, an argument list and the body of the method, consisting of Tcl and NX statements.

When an instance of Stack is created, it will contain an instance variable named things. If several Stack instances are created, each of the instances will have their own (same-named but different) instance variable. The instance variable things is used in our example as a list for the internal representation of the stack. We define in a next step the methods to access and modify this list structure. A user of the stack using the the provided methods does not have to have any knowledge about the name or the structure of the internal representation (the instance variable things).

The method push receives an argument thing which should be placed on the stack. Note that we do not have to specify the type of the element on the stack, so we can push strings as well as numbers or other kind of things. When an element is pushed, we add this element as the first element to the list things. We insert the element using the Tcl command linsert which receives the list as first element, the position where the element should be added as second and the new element as third argument. To access the value of the instance variable we use Tcl’s the dollar operator followed by the name. The names of instance variables are preceded with a colon :. Since the name contains a non-plain character, Tcl requires us to put braces around the name. The command linsert and its arguments are placed between square brackets. This means that the function is called and returns the new list, where the new element is inserted at the first position (index 0) in the list things. The result of the linsert function is assigned again to the instance variable things which is updated this way. Finally the method push returns the pushed thing using the return statement.

The method pop returns the most recently stacked element and removes it from the stack. Therefore, it takes the first element from the list (using the Tcl command lindex), assigns it to the method-scoped variable top, removes the element from the instance variable things (by using the Tcl command lrange) and returns the value popped element top.

This finishes our first implementation of the the stack, more enhanced versions will follow. Note that the methods push and pop are defined as public; this means that these methods can be used from all other objects in the system. Therefore, these methods provide an interface to the stack implementation.

Listing 3: Using the Stack

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
#!/bin/env tclsh
package require nx

nx::Class create Stack {

   #
   # Stack of Things
   #
   ....
}

Stack create s1
s1 push a
s1 push b
s1 push c
puts [s1 pop]
puts [s1 pop]
s1 destroy

Now we want to use the stack. The code snipped in Listing 3 shows how to use the class Stack in a script. Since NX is based on Tcl, the script will be called with the Tcl shell tclsh. In the Tcl shell we have to require package nx to use the Next Scripting Framework and NX. The next lines contain the definition of the stack as presented before. Of course, it is as well possible to make the definition of the stack an own package, such we could simple say package require stack, or to save the definition of a stack simply in a file and load it via source.

In line 12 we create an instance of the stack, namely the stack object s1. The object s1 has as an instance of the stack access to the methods, which can be invoked by the name of the object followed by the method name. In lines 13-15 we push on the stack the values a, then b, and c. In line 16 we output the result of the pop method using the Tcl command puts. We will see on standard output the value+c+ (the last stacked item). The output of the line 17 is the value b (the previously stacked item). Finally, in line 18 we destroy the object. This is not necessary here, but shows the life cycle of an object. In some respects, destroy is the counterpart of create from line 12.

object-class-appclass.png
Figure 4. Class and Object Diagram

Figure 4 shows the actual class and object structure of the first Stack example. Note that the common root class is nx::Object that contains methods for all objects. Since classes are as well objects in NX, nx::Class is a specialization of nx::Object. nx::Class provides methods for creating objects, such as the method create which is used to create objects (and classes as well).

2.2. Define an Object Named "stack"

The definition of the stack in Listing 2 is following the traditional object oriented approach, found in practically every object oriented programming language: Define a class with some methods, create instances from this class, and use the methods defined in the class in the instances of the class.

In our next example, we introduce generic objects and object specific methods. With NX, we can define generic objects, which are instances of the most generic class nx::Object (sometimes called "common root class"). nx::Object is predefined and contains a minimal set of methods applicable to all NX objects.

In our second example, we will define a generic object named stack and provide methods for this object. The methods defined in our first example were methods provided by a class for objects. Now we defined object specific methods, which are methods applicable only to the object for which they are defined.

Listing 5: Object stack

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
nx::Object create stack {

   :variable things {}

   :public method push {thing} {
      set :things [linsert ${:things} 0 $thing]
      return $thing
   }

   :public method pop {} {
      set top [lindex ${:things} 0]
      set :things [lrange ${:things} 1 end]
      return $top
   }
}

The example in Listing 5 defines the object stack in a very similar way as the class Stack. But the following points are different.

  • First, we use nx::Object instead of nx::Class to denote that we want to create a generic object, not a class.

  • As in the example above, we use :variable to define the instance variable things for this object (the object stack).

The definition for the methods push and pop are the same as before, but this times they are object specify. All methods defined on an object are object-specific. In order to use the stack, we can use directly the object stack in the same way as we have used the object s1 in Listing 3 (e.g. stack push a). this the object stack.

object-stack.png
Figure 6. Object stack

A reader might wonder when to use a class Stack or rather an object stack. A big difference is certainly that one can define easily multiple instances of a class, while the object is actually a singleton. The concept of the object stack is similar to a module providing a certain functionality via a common interface without providing the functionality to create multiple instances. The reuse of methods provided by the class to objects is as well a difference. If the methods of the class are updated, all instances of the class well immediately get the modified behavior. But this does not mean that there is no reuse for the methods of stack possible. NX allows for example to copy objects (similar to prototype based languages) or to reuse methods via e.g. aliases (more about this later).

Note that we use capitalized names for classes and lowercase names for instances. This is not required and a pure convention making it easier to understand scripts without much analysis.

2.3. Implementing Features using Mixin Classes

So far, the definition of the stack methods was pretty minimal. Suppose, we want to define "safe stacks" that protect e.g. against stack under-runs (a stack under-run happens, when more pop than push operations are issued on a stack). Safety checking can be implemented mostly independent from the implementation details of the stack (usage of internal data structures). There are as well different ways of checking the safety. Therefore we say that safety checking is orthogonal to the stack core implementation.

With NX we can define stack-safety as a separate class using methods with the same names as the implementations before, and "mix" this behavior into classes or objects. The implementation of Safety in stack under-runs and to issue error messages, when this happens.

Listing 7: Class Safety

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
nx::Class create Safety {

  #
  # Implement stack safety by defining an additional
  # instance variable named "count" that keeps track of
  # the number of stacked elements. The methods of
  # this class have the same names and argument lists
  # as the methods of Stack; these methods "shadow"
  # the methods of class Stack.
  #

  :variable count 0

  :public method push {thing} {
    incr :count
    next
  }

  :public method pop {} {
    if {${:count} == 0} then { error "Stack empty!" }
    incr :count -1
    next
  }
}

Note that the methods of the class Safety all end with next. This command is a primitive command of NX, that will call the same-named method with the same argument list as the current invocation.

Assume we safe the definition of the class Stack in a file named Stack.tcl and the definition of the class Safety in a file named Safety.tcl in the current directory. When we load the classes Stack and Safety into the same script (see the terminal dialog in e.g. a certain stack s2 as a safe stack, while all other stacks (such as s1) might be still "unsafe". This can be achieved via the option -mixin at the object creation time (see line 9 in option -mixin mixes the class Safety into the new instance s2.

Listing 8: Using the Class Safety

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
% package require nx
2.0
% source Stack.tcl
::Stack
% source Safety.tcl
::Safety
% Stack create s1
::s1
% Stack create s2 -mixin Safety
::s2
% s2 push a
a
% s2 pop
a
% s2 pop
Stack empty!

% s1 info precedence
::Stack ::nx::Object

% s2 info precedence
::Safety ::Stack ::nx::Object

When the method push of s2 is called, first the method of the mixin class Safety will be invoked that increments the counter and continues with next to call the shadowed method, here the method push of the Stack implementation that actually pushes the item. The same happens, when s2 pop is invoked, first the method of Safety is called, then the method of the Stack. When the stack is empty (the value of count reaches 0), and pop is invoked, the mixin class Safety generates an error message (raises an exception), and does not invoke the method of the Stack.

The last two commands in Listing 8 use introspection to query for the objects s1 and s2 the order in which the classes are processed. This order is called the precedence order and is obtained via info precedence. We see that the mixin class Safety is only in use for s2, and takes there precedence over Stack. The common root class nx::Object is for both s1 and s2 the base class.

per-object-mixin.png
Figure 9. Per-object Mixin

Note that the class Safety is only mixed into a single object (here s2), therefore we refer to this case as a per-object mixin. Figure 9 shows the class diagram, where the class Safety is used as a per-object mixin for s2.

The class Safety can be used as well in other ways, such as e.g. for defining classes for safe stacks <<xmp-class-safestack,

Listing 10: Class SafeStack

  1
  2
  3
  4
  5
  6
  7
#
# Create a safe stack class by using Stack and mixin
# Safety
#
Class create SafeStack -superclass Stack -mixin Safety

SafeStack create s3

The difference to the case with the per-object mixin is that now, Safety is mixed into the definition of SafeStack. Therefore, all instances of the class SafeStack (here the instance s3) will be using the safety definitions. for this definition.

per-class-mixin.png
Figure 11. Per-class Mixin

Note that we could use Safety as well as a per-class mixin on Stack. In this case, all stacks would be safe stacks and we could not provide a selective feature selection (which might be perfectly fine).

2.4. Define Different Kinds of Stacks

The definition of Stack is generic and allows all kind of elements to be stacked. Suppose, we want to use the generic stack definition, but a certain stack (say, stack s4) should be a stack for integers only. This behavior can be achieved by the same means as introduced already in Listing 5, namely object-specific methods.

Listing 12: Object Integer Stack

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
Stack create s4 {

  #
  # Create a stack with a object-specific method
  # to check the type of entries
  #

  :public method push {thing:integer} {
    next
  }
}

The program snippet in Listing 12 defines an instance s4 of the class Stack and provides an object specific method for push to implement an integer stack. The method pull is the same for the integer stack as for all other stacks, so it will be reused as usual from the class Stack. The object-specific method push of s4 has a value constraint in its argument list (thing:integer) that makes sure, that only integers can be stacked. In case the argument is not an integer, an exception will be raised. Of course, one could perform the value constraint checking as well in the body of the method proc by accepting an generic argument and by performing the test for the value in the body of the method. In the case, the passed value is an integer, the push method of Listing 12 calls next, and therefore calls the shadowed generic definition of push as provided by Stack.

Listing 13: Class IntegerStack

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
nx::Class create IntegerStack -superclass Stack {

  #
  # Create a Stack accepting only integers
  #

  :public method push {thing:integer} {
    next
  }
}

An alternative approach is shown in Listing 13, where the class IntegerStack is defined, using again the method definition use for s4, this time on the class level.

2.5. Define Class Specific Methods

In our previous examples we defined methods provided by classes (applicable for its instances) and object-specific methods (methods defined on objects, only applicable for these objects). In this section, we introduce methods defined on classes, which are only applicable for the class objects. Such methods are sometimes called class methods or "static methods".

In NX classes are objects with certain properties (providing methods for instances, managing object life-cycles; we will come to this later in more detail). Since classes are objects, we can define as well object-specific methods for the class objects. However, since :method applied on classes defines methods for instances, we have to use the method-modifier class to denote methods to be applied on the class itself. Note that class methods are not inherited to instances. These methods defined on the class object are actually exactly same as the object-specific methods in the examples above.

Listing 14: Class Stack2

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
nx::Class create Stack2 {

   :public class method available_stacks {} {
      return [llength [:info instances]]
   }

   :variable things {}

   :public method push {thing} {
      set :things [linsert ${:things} 0 $thing]
      return $thing
   }

   :public method pop {} {
      set top [lindex ${:things} 0]
      set :things [lrange ${:things} 1 end]
      return $top
   }
}

Stack create s1
Stack create s2

puts [Stack available_stacks]

The class Stack2 in Listing 14 consists of the the earlier definition of the class Stack extended by the class-specific method available_stacks, that returns the current number of instances of the stack. The final command puts (line 26) prints 2 to the console.

stack2.png
Figure 15. Stack2

The class diagram in Figure 15 shows the diagrammatical representation of the class object-specific method available_stacks. We omit here the common root class. Since every class is a specialization of the common root class nx::Object, the common root class is often omitted from the class diagrams.

3. Basic Language Features of NX

3.1. Variables and Properties

In general, NX does not need variable declarations. It allows to create, modify or variables on the fly by using e.g. the Tcl commands set and unset. Depending on the variable name (or more precisely, depending on the prefix consisting of colons :), a variable is either local to a method, or an instance variable, or a global variable. The rules are:

  • A variable without any colon prefix refers typically to a method scoped variable (the variable is created at the begin of the invocation of the method and deleted, when the method ends); In the example below, the variable a is method scoped.

  • A variable with a single colon prefix refers to an instance variable (the variable is part of the object, when the object is destroyed, the variable is deleted as well. In the example below, the variable b is an instance variable.

  • a variable with two leading colons refers to a global variable (the lifespan ends when te variable is explicitly unset or the script ends). Also variables in placed in Tcl namespaces are global variables. In the example below, the variable c is a global variable.

Listing 16: Scopes of Variables

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
Class create Foo {

  :method foo args {...}
    # Method scoped variable a
    set a 1
    # Instance variable b
    set :b 2
    # Global variable/namespaced variable c
    set ::c 3
  }
}

Listing 16 shows a method foo of some class Foo referring to differently scoped variables.

3.1.1. Properties: Instance Variables with Accessors

So, in general, there is no need to define or declare instance variables in NX. However, in some cases, a definition is useful. For example, one can define properties on classes, which are inherited to subclasses, and which are used during object initialization. Consider the following example:

Listing 17: Properties

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
#
# Define a class Person with properties "name"
# and "birthday"
#
nx::Class create Person {
  :property name:required
  :property birthday
}

#
# Define a class Student as specialization of Person
# with additional properties
#
nx::Class create Student -superclass Person {
  :property matnr:required
  :property {oncampus:boolean true}
}

#
# Create instances using object parameters
# for the initialization
#
Person create p1 -name Bob
Student create s1 -name Susan -matnr 4711

# Access property value via accessor method
puts "The name of s1 is [s1 name]"

We define here a class named Person with two properties, namely name and birthday.

We refer with the term property to an instance variable with accessors, where the property definition might carry as well value-constraints and a default value.

When the class Person is defined, NX provides as well automatically accessors. Accessors are methods named like the variables, which are used to access (to read and write) the underlying instance variables. Therefore, in our example, the class Person has two methods implied by the property definition, namely the method name and the method birthday.

The class Student is defined as a specialization of Person with two additional properties, namely matnr and oncampus. The property matnr is required (it has to be provided, when an instance of this class is created), and the property oncampus is boolean, and is per default set to true. Note that the class Student inherits the properties of Person, therefore it will have 4 properties in total.

The property definitions are also used for providing object parameters. These are non-positional parameters provided during object creation to supply values to the instance variables. In our listing, we create an instance of Person using the object parameter name and provide the value of Bob to the instance variable name. Similarly, we create an instance of Student using the two object parameters name and matnr. Finally we use the accessor method name to obtain the value of the instance variable name of object s1.

3.1.2. Instance Variables without Accessors

In some cases one would like to define instance variables without accessors, e.g. for keeping the internal state of an object. For this purpose, one can use the predefined method variable, which is in many respects similar to property. One difference is, that property uses the same syntax as for method parameters, and variable receives the default value as a separate argument (similar to the variable command in Tcl. The introductory Stack example in in Listing 2 used already the method variable.

Listing 18: Declaring Variables

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
Class create Base {
  :variable x 1
  # ...
}

Class create Derived -superclass Base {
  :variable y 2
  # ...
}

# Create instance of the class Derived
Derived create d1

# Object d1 has instance variables
# x == 1 and y == 2

Note that the variable definitions are inherited in the same way as properties. The example in Listing 18 shows a class Derived that inherits from Base. When an instance d1 is created, it will contain the two instance variables x and y.

Listing 19: Setting Variables in the Constructor

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
Class create Base2 {
 # ...
 :method init {} {
   set :x 1
   # ....
 }
}

Class create Derived2 -superclass Base2 {
 # ...
 :method init {} {
   set :y 2
   next
   # ....
 }
}

# Create instance of the class Derived2
Derived2 create d2

In many other object oriented languages, the instance variables are initialized solely by the constructor, similar to class Derived2 in Listing 19 . This approach is certainly as well possible in NX. Note however, that the approach using constructors requires an explicit method chaining between the constructors and is less declarative.

3.2. Method Definitions

3.2.1. Scripted Methods

3.2.2. C-implemented Methods

3.2.3. Method-Aliases

3.3. Method Protection

3.4. Scopes of Methods

3.4.1. Inherited Methods

3.4.2. Class Methods

3.4.3. Object Methods

3.5. Parameters

NX provides a generalized mechanism for passing values to either methods (we refer to these as method parameters) or to objects (these are called object parameters). Both kind of parameters might have different features, such as:

  • Positional and non-positional parameters

  • Required and non-required parameters

  • Default values for parameters

  • Value-checking for parameters

  • Multiplicity of parameters

TODO: complete list above and provide a short summary of the section

Before we discuss method and object parameters in more detail, we describe the parameter features in the subsequent sections based on method parameters.

3.5.1. Positional and Non-Positional Parameters

If the position of a parameter in the list of formal arguments (e.g. passed to a function) is significant for its meaning, this is a positional parameter. If the meaning of the parameter is independent of its postion, this is a non-positional parameter. When we call a method with positional parameters, the meaning of the parameters (the association with the argument in the argument list of the method) is determined by its position. When we call a method with non-positional parameters, their meaning is determined via a name passed with the argument during invocation.

Listing 20: Positional and Non-Positional Method Parameters

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
nx::Object create o1 {

  #
  # Method foo has positional parameters:
  #
  :public method foo {x y} {
    puts "x=$x y=$y"
  }

  #
  # Method bar has non-positional parameters:
  #
  :public method bar {-x -y} {
    puts "x=$x y=$y"
  }

  #
  # Method baz has non-positional and
  # positional parameters:
  #
  :public method baz {-x -y a} {
    puts "x? [info exists x] y? [info exists y] a=$a"
  }
}

# invoke foo (positional parameters)
o1 foo 1 2

# invoke bar (non-positional parameters)
o1 bar -y 3 -x 1
o1 bar -x 1 -y 3

# invoke baz (positional and non-positional parameters)
o1 baz -x 1 100
o1 baz 200
o1 baz -- -y

Consider the example in Listing 20. The method foo has the argument list x y. This means that the first argument is passed in an invocation like o1 foo 1 2 to x (here, the value 1), and the second argument is passed to y (here the value 2). Method bar has in contrary just with non-positional arguments. Here we pass the names of the parameter together with the values. In the invocation o1 bar -y 3 -x 1 the names of the parameters are prefixed with a dash ("-"). No matter whether in which order we write the non-positional parameters in the invocation (see line 30 and 31 in Listing 20) in both cases the variables x and y in the body of the method bar get the same values assigned (x becomes 1, y becomes 3).

It is certainly possible to combine positional and non-positional arguments. Method baz provides two non-positional parameter (-y and -y) and one positional parameter (namely a). The invocation in line 34 passes the value of 1 to x and the value of 100 to a. There is no value passed to y, therefore value of y will be undefined in the body of baz, info exists y checks for the existence of the variable y and returns 0.

The invocation in line 35 passes only a value to the positional parameter. A more tricky case is in line 36, where we want to pass -y as a value to the positional parameter a. The case is more tricky since syntactically the argument parser might consider -y as the name of one of the non-positional parameter. Therefore we use -- (double dash) to indicate the end of the block of the non-positional parameters and therefore the value of -y is passed to a.

3.5.2. Optional and Required Parameters

Per default positional parameters are required, and non-positional parameters are optional (they can be left out). By using parameter options, we can as well define positional parameters, which are optional, and non-positional parameters, which are required.

Listing 21: Optional and Required Method Parameters

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
nx::Object create o2 {

  #
  # Method foo has one required and one optional
  # positional parameter:
  #
  :public method foo {x:required y:optional} {
    puts "x=$x y? [info exists y]"
  }

  #
  # Method bar has one required and one optional
  # non-positional parameter:
  #
  :public method bar {-x:required -y:optional} {
    puts "x=$x y? [info exists y]"
  }
}

# invoke foo (one optional positional parameter is missing)
o2 foo 1

The example in Listing 21 defined method foo with one required and one optional positional parameter. For this purpose we use the parameter options required and optional. The parameter options are separated from the parameter name by a colon. If there are multiple parameter options, these are separated by commas (we show this in later examples).

The parameter definition x:required for method foo is equivalent to x without any parameter options (see e.g. previous example), since positional parameters are per default required. The invocation in line 21 of Listing 21 will lead to an undefined variable y in method foo, because no value us passed to the optional parameter. Note that only trailing positional parameters might be optional. If we would call method foo of Listing 20 with only one argument, the system would raise an exception.

Similarly, we define method bar in Listing 21 with one required and one optional non-positional parameter. The parameter definition -y:optional is equivalent to -y, since non-positional parameter are per default optional. However, the non-positional parameter -x:required is required. If we invoke bar without it, the system will raise an exception.

3.5.3. Default Values for Parameters

Optional parameters might have a default value, which will be used, when not value is provided for this parameter. Default values can be specified for positional and non-positional parameters.

Listing 22: Method Parameters with Default Values

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
nx::Object create o3 {

  #
  # Positional parameter with default value:
  #
  :public method foo {x:required {y 101}} {
    puts "x=$x y? [info exists y]"
  }

  #
  # Non-positional parameter with default value:
  #
  :public method bar {{-x 10} {-y 20}} {
    puts "x=$x y? [info exists y]"
  }
}

# use default values
o3 foo
o3 bar

In order to define a default value, the parameter specification must be of the form of a 2 element list, where the second argument is the default value. See for an example in Listing 22.

3.5.4. Value Constraints

NX provides value constraints for all kind of parameters. By specifying value constraints a developer can restrict the permissible values for a parameter and document the expected values in the source code. Value checking in NX is conditional, it can be turned on or off in general or on a per-usage level (more about this later). The same mechanisms can be used not only for input value checking, but as well for return value checking (we will address this point as well later).

Built-in Value Constraints

NX comes with a set of built-in value constraints, which can be extended on the scripting level. The built-in checkers are either the native checkers provided directly by the Next Scripting Framework (the most efficient checkers) or the value checkers provided by Tcl through string is …. The built-in checkers have as well the advantage that they can be used also at any time during bootstrap of an object system, at a time, when e.g. no objects or methods are defined. The same checkers are used as well for all C-implemented primitives of NX and the Next Scripting Framework.

value-checkers.png
Figure 23. General Applicable Value Checkers in NX

Figure 23 shows the built-in general applicable value checkers available in NX, which can be used for all method and object parameters. In the next step, we show how to use these value-checkers for checking permissible values for method parameters. Then we will show, how to provide more detailed value constraints.

Listing 24: Method Parameters with Value Constraints

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
nx::Object create o4 {

  #
  # Positional parameter with value constraints:
  #
  :public method foo {x:integer o:object,optional} {
    puts "x=$x o? [info exists o]"
  }

  #
  # Non-positional parameter with value constraints:
  #
  :public method bar {{-x:integer 10} {-verbose:boolean false}} {
    puts "x=$x y=$y"
  }
}

# The following invocation raises an exception
o4 foo a

Value contraints are specified as parameter options in the parameter specifications. The parameter specification x:integer defines x as a required positional parmeter which value is constraint to an integer. The parameter specification o:object,optional shows how to combine multiple parameter options. The parameter o is an optional positional parameter, its value must be an object (see Listing 24). Value constraints are specified exactly the same way for non-positional parameters (see method bar in Listing 24).

Listing 25: Parameterized Value Constraints

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
#
# Create classes for Person and Project
#
Class create Person
Class create Project

nx::Object create o5 {
  #
  # Parameterized value constraints
  #
  :public method work {
     -person:object,type=Person
     -project:object,type=Project
   } {
    # ...
  }
}

#
# Create a Person and a Project instance
#
Person create gustaf
Project create nx

#
# Use method with value constraints
#
o5 work -person gustaf -project nx

The native checkers object, class, metaclass and baseclass can be further specialized with the parameter option type to restrict the permissible values to instances of certain classes. We can use for example the native value constraint object either for testing whether an argument is some object (without further constraints, as in Listing 22, method foo), or we can constrain the value further to some type (direct or indirect instance of a class). This is shown by method work in Listing 25 which requires the parameter -person to be an instance of class Person and the parameter -project to be an instance of class Project.

Scripted Value Constraints

The set of predefined value checkers can be extended by application programs via defining methods following certain conventions. The user defined value checkers are defined as methods of the class nx::Slot or of one of its subclasses or instances. We will address such cases in the next sections. In the following example we define two new value checkers on class nx::Slot. The first value checker is called groupsize, the second one is called choice.

Listing 26: Scripted Value Checker for Method Parameters

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
#
# Value checker named "groupsize"
#
::nx::Slot method type=groupsize {name value} {
  if {$value < 1 || $value > 6} {
    error "Value '$value' of parameter $name is not between 1 and 6"
  }
}

#
# Value checker named "choice" with extra argument
#
::nx::Slot method type=choice {name value arg} {
  if {$value ni [split $arg |]} {
    error "Value '$value' of parameter $name not in permissible values $arg"
  }
}

#
# Create an application class D
# using the new value checkers
#
Class create D {
  :public method foo {a:groupsize} {
    # ...
  }
  :public method bar {a:choice,arg=red|yellow|green b:choice,arg=good|bad} {
    # ...
  }
}

D create d1

# testing "groupsize"
d1 foo 2
d1 foo 10

# testing "choice"
d1 bar green good
d1 bar pink bad

In order to define a checker groupsize a method of the name type=groupsize is defined. This method receives two arguments, name and value. The first argument is the name of the parameter (mostly used for the error message) and the second parameter is provided value. The value checker simply tests whether the provided value is between 1 and 3 and raises an exception if this is not the case (invocation in line 36 in Listing 26).

The checker groupsize has the permissible values defined in its method’s body. It is as well possible to define more generic checkers that can be parameterized. For this parameterization, one can pass an argument to the checker method (last argument). The checker choice can be used for restricting the values to a set of predefined constants. This set is defined in the parameter specification. The parameter a of method bar in Listing 26 is restricted to the values red, yellow or green, and the parameter b is restricted to good or bad. Note that the syntax of the permissible values is solely defined by the definition of the value checker in lines 13 to 17. The invocation in line 39 will be ok, the invocation in line 40 will raise an exception, since pink is not allowed.

If the same checks are used in many places in the program, defining names for the value checker will be the better choice since it improves maintainability. For seldomly used kind of checks, the parameterized value checkers might be more convenient.

3.5.5. Multiplicity

Multiplicity is used to define whether a parameter should receive single or multiple values.

A multiplicity specification has a lower and an upper bound. A lower bound of 0 means that the value might be empty. A lower bound of 1 means that the the parameter needs at least one value. The upper bound might be 1 or n (or synonymously *). While the upper bound of 1 states that at most one value has to be passed, the upper bound of n says that multiple values are permitted. Other kinds of multiplicity are currently not allowed.

The multiplicity is written as parameter option in the parameter specification in the form lower-bound..upper-bound. If no multiplicity is defined the default multiplicity is 1..1, which means: provide exactly one (atomic) value (this was the case in the previous examples).

Listing 27: Method Parameters with Explicit Multiplicity

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
nx::Object create o6 {

  #
  # Positional parameter with an possibly empty
  # single value
  #
  :public method foo {x:integer,0..1} {
    puts "x=$x"
  }

  #
  # Positional parameter with an possibly empty
  # list of values value
  #
  :public method bar {x:integer,0..n} {
    puts "x=$x"
  }

  #
  # Positional parameter with a non-empty
  # list of values
  #
  :public method baz {x:integer,1..n} {
    puts "x=$x"
  }
}

Listing 27 contains three examples for positional parameters with different multiplicities. Multiplicity is often combined with value constraints. A parameter specification of the form x:integer,0..n means that the parameter x receives a list of integers, which might be empty. Note that the value constraints are applied to every single element of the list.

The parameter specification x:integer,0..1 means that x might be an integer or it might be empty. This is one style of specifying that no explicit value is passed for a certain parameter. Another style is to use required or optional parameters. NX does not enforce any particular style for handling unspecified values.

All the examples in Listing 27 are for single positional parameters. Certainly, multiplicity is fully orthogonal with the other parameter features and can be used as well for multiple parameters, non-positional parameter, default values, etc.

4. Advanced Language Features

4.1. Objects, Classes and Meta-Classes

4.2. Details on Method and Object Parameters

The parameter specifications are used in NX for the following purposes. They are used for

  • the specification of input arguments of methods and commands, for

  • the specification of return values of methods and commands, and for

  • the specification for the initialization of objects.

We refer to the first two as method parameters and the last one as object parameters. The examples in the previous sections all parameter specification were specifications of method parameters.

Method parameters specify properties about permissible values passed to methods.

The method parameter specify how methods are invoked, how the actual arguments are passed to local variables of the invoked method and what kind of checks should be performed on these.

Object parameters are parameters that specify, with what values instance variables of objects are initialized and how these objects could be parameterized.

Syntactically, object parameters and method parameters are the same, although there are certain differences (e.g. some parameter options are only applicable for objects parameters, the list of object parameters is computed dynamically from the class structures, object parameters are often used in combination with special setter methods, etc.). Consider the following example, where we define the two application classes Person and Student with a few properties.

Listing 28: Object Parameters

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
#
# Define a class Person with properties "name"
# and "birthday"
#
nx::Class create Person {
  :property name:required
  :property birthday
}

#
# Define a class Student as specialization of Person
# with and additional property
#
nx::Class create Student -superclass Person {
  :property matnr:required
  :property {oncampus:boolean true}
}

#
# Create instances using object parameters
# for the initialization
#
Person create p1 -name Bob
Student create s1 -name Susan -matnr 4711

# Access property value via accessor method
puts "The name of s1 is [s1 name]"

The class Person has two properties name and birthday, where the property name is required, the property birthday is not. The class Student is a subclass of Person with the additional required property matnr and an optional property oncampus with the default value true (see Listing 28). The class diagram below visualizes these definitions.

object-parameter.png
Figure 29. System and Application Classes

In NX, these definitions imply that instances of the class of Person have the properties name and birthday as non-positional object parameters. Furthermore it implies that instances of Student will have the object parameters of Person augmented with the object parameters from Student (namely matnr and oncampus). Based on these object parameters, we can create a Person named Bob and a Student named Susan with the matriculation number 4711 (see line 23 and 24 in Listing 28). After the object s1 is created it has the instance variables name, matnr and oncampus (the latter is initialized with the default value).

4.2.1. Object Parameters for all NX Objects

The object parameters are not limited to the application defined properties, also NX provides some predefined definitions. Since Person is a subclass of nx::Object also the object parameters of nx::Object are inherited. In the introductory stack example, we used -mixin applied to an object to denote per-object mixins (see Listing 8). Since mixin is defined as a parameter on nx::Object it can be used as an object parameter -mixin for all objects in NX. To put it in other words, every object can be configured to have per-object mixins. If we would remove this definition, this feature would be removed as well.

As shown in the introductory examples, every object can be configured via a scripted initialization block (the optional scripted block specified at object creation as last argument; see Listing 5 or Listing 12). The scripted block and its meaning are as well defined by the means of object parameters. However, this object parameter is positional (last argument) and optional (it can be omitted). The following listing shows (simplified) the object parameters of Person p1 and Student s1.

Listing 30: Computed Actual Object Parameter (simplified)

  1
  2
  3
  4
  5
  6
  7
  8
Object parameter for Person p1:
   -name:required -birthday ...
   -mixin:mixinreg,alias,1..n -filter:filterreg,alias,1..n _:initcmd,optional

Object parameter for Student s1:
   -matnr:required {-oncampus:boolean true}
   -name:required -birthday ...
   -mixin:mixinreg,alias,1..n -filter:filterreg,alias,1..n _:initcmd,optional

The actual values can be optained via introspection via Person info parameter definition. The object parameter types mixinreg and filterreg are for converting definitions of filters and mixins. The object parameter type initcmd says that the content of this variable will be executed in the context of the object being created (before the constructor init is called). More about the object parameter types later.

4.2.2. Object Parameters for all NX Classes

Since classes are certain kind of objects, classes are parameterized in the same way as objects. A typical parameter for a class definition is the relation of the class to its superclass.In our example, we have specified, that Student has Person as superclass via the non-positional object parameter -superclass. If no superclass is specified for a class, the default superclass is nx::Object. Therefore nx::Object is the default value for the parameter superclass.

Another frequently used parameter for classes is -mixin to denote per-class mixins (see e.g. the introductory Stack example in Listing 10), which is defined in the same way.

Since Student is an instance of the meta-class nx::Class it inherits the object parameters from nx::Class (see class diagram Figure 29). Therefore, one can use e.g. -superclass in the definition of classes.

Since nx::Class is a subclass of nx::Object, the meta-class nx::Class inherits the parameter definitions from the most general class nx::Object. Therefore, every class might as well be configured with a scripted initialization block the same way as objects can be configured. We used actually this scripted initialization block in most examples for defining the methods of the class. The following listing shows (simplified) the parameters applicable for Class Student.

Listing 31: Computed Parameters for a Class (simplified)

  1
  2
  3
Object parameter for Class Student:
   -mixin:mixinreg,alias,1..n -filter:filterreg,alias,1..n ...
   {-superclass:class,alias,1..n ::nx::Object} ...

The actual values can be obtained via introspection via nx::Class info parameter definition.

4.2.3. User defined Parameter Types

More detailed definition of the object parameter types comes here.

4.2.4. Slot Classes and Slot Objects

In one of the previous sections, we defined scripted (application defined) checker methods on a class named nx::Slot. In general NX offers the possibility to define value checkers not only for all usages of parameters but as well differently for method parameters or object parameters

slots.png
Figure 32. Slot Classes and Objects

4.2.5. Attribute Slots

Still Missing

  • return value checking

  • switch

  • initcmd …

  • subst rules

  • converter

  • incremental slots

5. Miscellaneous

5.1. Profiling

5.2. Unknown Handlers

NX provides two kinds of unknown handlers:

  • Unknown handlers for methods

  • Unknown handlers for objects and classes

5.2.1. Unknown Handlers for Methods

Object and classes might be equipped with a method unknown which is called in cases, where an unknown method is called. The method unknown receives as first argument the called method followed by the provided arguments

Listing 33: Unknown Method Handler

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
::nx::Object create o {
  :method unknown {called_method args} {
    puts "Unknown method '$called_method' called"
  }
}

# Invoke an unknown method for object o:
o foo 1 2 3

# Output will be: "Unknown method 'foo' called"

Without any provision of an unknown method handler, an error will be raised, when an unknown method is called.

5.2.2. Unknown Handlers for Objects and Classes

The next scripting framework provides in addition to unknown method handlers also a means to dynamically create objects and classes, when these are referenced. This happens e.g. when superclasses, mixins, or parent objects are referenced. This mechanism can be used to implement e.g. lazy loading of these classes. Nsf allows to register multiple unknown handlers, each identified by a key (a unique name, different from the keys of other unknown handlers).

Listing 34: Unknown Class Handler

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
::nx::Class public class method __unknown {name} {
  # A very simple unknown handler, showing just how
  # the mechanism works.
  puts "***** __unknown called with <$name>"
  ::nx::Class create $name
}

# Register an unknown handler as a method of ::nx::Class
::nsf::object::unknown::add nx {::nx::Class __unknown}

::nx::Object create o {
  # The class M is unknown at this point

  :mixin add M
  # The line above has triggered the unknown class handler,
  # class M is now defined

  puts [:info mixin classes]
  # The output will be:
  #     ***** __unknown called with <::M>
  #     ::M
}

The Next Scripting Framework allows to add, query, delete and list unknown handlers.

Listing 35: Unknown Handler registration

  1
  2
  3
  4
  5
# Interface for unknown handlers:
# nsf::object::unknown::add /key/ /handler/
# nsf::object::unknown::get /key/
# nsf::object::unknown::delete /key/
# nsf::object::unknown::keys
References
  • Gustaf Neumann, Stefan Sobernig: XOTcl 2.0 - A Ten-Year Retrospective and Outlook, in: Proceedings of the Sixteenth Annual Tcl/Tk Conference, Portland, Oregon, October, 2009

  • U. Zdun, M. Strembeck, G. Neumann: Object-Based and Class-Based Composition of Transitive Mixins, Information and Software Technology, 49(8) 2007 .

  • G. Neumann and U. Zdun: Filters as a language support for design patterns in object-oriented scripting languages. In Proceedings of COOTS’99, 5th Conference on Object-Oriented Technologies and Systems, San Diego, May 1999.

  • G. Neumann and U. Zdun: Implementing object-specific design patterns using per-object mixins. In Proc. of NOSA`99, Second Nordic Workshop on Software Architecture, Ronneby, Sweden, August 1999.

  • G. Neumann and U. Zdun: Enhancing object-based system composition through per-object mixins. In Proceedings of Asia-Pacific Software Engineering Conference (APSEC), Takamatsu, Japan, December 1999.

  • G. Neumann and U. Zdun: XOTCL, an object-oriented scripting language. In Proceedings of Tcl2k: The 7th USENIX Tcl/Tk Conference, Austin, Texas, February 2000.

  • G. Neumann and U. Zdun: Towards the Usage of Dynamic Object Aggregations as a Form of Composition In: Proceedings of Symposium of Applied Computing (SAC’00), Como, Italy, Mar 19-21, 2000.

  • G. Neumann, S. Sobernig: XOTcl 2.0 - A Ten-Year Retrospective and Outlook, in: Proceedings of the Sixteenth Annual Tcl/Tk Conference, Portland, Oregon, October, 2009.

  • J. K. Ousterhout: Tcl: An embeddable command language. In Proc. of the 1990 Winter USENIX Conference, January 1990.

  • J. K. Ousterhout: Scripting: Higher Level Programming for the 21st Century, IEEE Computer 31(3), March 1998.

  • D. Wetherall and C. J. Lindblad: Extending Tcl for Dynamic Object-Oriented Programming. Proc. of the Tcl/Tk Workshop '95, July 1995.