Topic variable

Topic variable
You are encouraged to solve this task according to the task description, using any language you may know.

Several programming languages offer syntax shortcuts to deal with the notion of "current" or "topic" variable. A topic variable is a special variable with a very short name which can also often be omitted.

Demonstrate the utilization and behaviour of the topic variable within the language and explain or demonstrate how the topic variable behaves under different levels of nesting or scope, if this applies, within the language.

For instance you can (but you don't have to) show how the topic variable can be used by assigning the number $3$ to it and then computing its square and square root.

Axe

In Axe, evaluated expressions can be "remembered" until the next expression is evaluated. <lang axe>3 Disp *3▶Dec,i</lang>

Prints:

    9

However, attempting to use the result now would result in garbage due to the ▶Dec and Disp commands overwriting the previous result.

Clojure

The Clojure REPL has *1 (and also *2 and *3 for the 2nd or 3rd most recent) <lang clojure> user=> 3 3 user=> (Math/pow *1 2) 9.0 user=> (Math/pow *2 0.5) 1.7320508075688772 </lang>

Erlang

There are no global variables, so Erlang uses a function, v(N), to get old results. Either the absolute command result, if N is positive, or, if N is negative, the return value of the Nth previous command.

Output:

7> 1 + 2.
3
8> v(-1).
3

Forth

In a stack oriented language like Forth the definition of variables is minimized as much as possible. The closest thing to a topic variable is the use of R@. This gets the top item from the return stack, which by the way is also used for flow control. It is up to the programmer to keep the stack balanced. In some Forth dialects R@ and I are identical. I is used as a loop index, e.g.

<lang forth>: myloop 11 1 do i . loop cr ; myloop</lang>

Which will print all numbers from 1 to 10. A typical use of R@ is illustrated here:

<lang forth>: ^2 dup * ;

sqrt 0 tuck ?do 1+ dup 2* 1+ +loop ;

topic >r r@ ^2 . r@ sqrt . r> drop ;

23 topic</lang>

The word >R places the item on the return stack and the word R> retrieves it from the return stack - an experienced Forth programmer would optimize this definition even further. Note that for technical reasons all words listed cannot be used outside definitions, so it may be argued that Forth doesn't have topic variables.

J

With this new definition of topic variables, the closest thing J has to a topic variable is probably dummy variables used in function definitions, because we always omit declaring which variables they are, and because we may omit them entirely. But, we still need to place the function in the context of the value it's being used on.

Thus, for example (entirely eliminating the variable representing the argument):

<lang J> example=: *:, %: NB. *: is square, %: is square root

  example 3

9 1.73205</lang>

Or, if we want to see the dummy variable in place (though still not declared, because there is no point to that):

<lang J> Example=: verb def '(*: y), (%: y)'

  Example 3

9 1.73205</lang>

Or course, if it's crucial to the concept of topic variables that they not be constrained to definitions of things like functions, then it might be argued that J does not have them.

On the third hand, note that "definitions of functions" do not actually need to be associated with names. At worst, some definitions might need to be enclosed in parenthesis:

<lang J> (*:, %:) 3 9 1.73205</lang>

And if we were to insist on leaving out functions, it's not clear that there would be much of anything left of the language to be doing anything with. See also Henry Rich's writeup on Tacit Programs.

jq

In jq, . (a period) generally refers to the "current input", which is always a JSON entity. For example, the jq program consisting of . alone is a filter, like "cat" in the Unix tradition.

The "." referring to the current input can often be omitted altogether, for example the expression ". | exp" for computing e^x can always be written simply as "exp".

There are some special character combinations involving ".":

• .[] is a filter which expects the input to be an array or a JSON object, and converts the input to be a stream of the constituent values;
• if i as an integer then .[i] is a filter which expects the input to be an array, and which extracts its i-th element (counting from 0).

Multiple references to the current input are allowed, e.g. [.,.] constructs an array of length two, each component being equal to the current input; thus the compound expression "1 | [.,.]" yields [1,1].

Mathematica

This depends on the Mathematica REPL. For the examples, I will use the textual format that Mathematica provides. Here's some basic examples:

In[1]:= 3

Out[1]= 3

In[2]:= %1^2

Out[2]= 9

In[3]:= Sqrt[%%]

Out[3]= Sqrt[3]

In[4]:= N[Out[-1]] (* for floating point *)

Out[4]= 1.73205

In this, I use 3 different forms. Here's a list of them:

• %<n>: Does the same as Out[n].
• %, %%, %%%, etc.: Does the same as Out[-<number of %s>].
• Out[n]: Returns the output of the nth cell.
• Out[-n]: Returns the output of the nth to last cell, excluding the one that it is in. (Out[-1] gives last output, -2 gives second to last, etc.)
• In[n]: Reevaluates and returns the input of the nth cell.
• In[-n]: Reevaluates and returns the input of the nth to last cell.

When an input is reevaluated, it also reassigns all relative Ins and Outs. Look at this for an example of its strange effects:

In[1]:= In[2]

Out[1]= In[2]

In[2]:= In[1]

$IterationLimit::itlim: Iteration limit of 4096 exceeded.

Out[2]= Hold[In[1]]

In it, it gets stuck in an infinite loop between In[1] and In[2], which evaluate to each other.

Oforth

Oforth does not have global variables, topic or not.

The closest thing Oforth has to a topic variable definition is the top of its data stack which always holds the last result without the need to assign it to a (local) variable but I don't think it is the philosophy of this task.

This will push 3 on the stack and compute sq and sqrt : <lang Oforth>3 dup sq swap sqrt</lang>

PARI/GP

gp is a REPL for GP, within which % can be used to refer to the last result. <lang parigp>3 [sqrt(%),%^2]</lang>

Output:

%1 = 3
%2 = [1.7320508075688772935274463415058723670, 9]

Perl

In Perl the topic variable is $_. It is the default argument for many functions, including the square root. It is also the default parameter for loops:

<lang Perl>my $_ = 3; print for $_**2, "\n", sqrt;</lang>

The topic parameter is lexical, so its use can be nested into several lexical scopes. However, assignment to the topic variable at loop declaration is not lexical, and a my function or local keyword is needed to enable the loops to nest:

<lang perl>for ($_ = 0; $_ <= 9; $_++) {

 print "Outer";
 print "$_\n";
 # The inner loop will not nest properly unless
 # it is preceded by a my statement
 my $_;    # This is required to nest the inner loop
 for ($_ = 0; $_ <= 9; $_++) {
   print "Inner";
   print "$_\n";
 }
 # Alternatively we can use a local keyword in the
 # inner loop declaration instead of a my statement
 for (local $_ = 0; $_ <= 9; $_++) {
   print "Alternative";
   print "$_\n";
 }

}</lang>

Perl 6

As in previous versions of Perl, in Perl6 the topic variable is $_. In addition to a direct affectation, it can also be set with the 'given' keyword. A method can be called from it with an implicit call:

<lang Perl 6>given 3 {

   .say for $_**2, .sqrt;

}</lang> The scope of the $_ variable is always lexical in Perl 6, though a function can access its caller's topic if it asks for it specially via CALLER::<$_>.

PicoLisp

<lang PicoLisp>PicoLisp sets the value of the variable (symbol) '@' to the result of conditional and controlling expressions in flow- and logic-functions (cond, if, and, when, while, etc.).

Within a function or method '@' behaves like a local variable, i.e. its value is automatically saved upon function entry and restored at exit.

For example, to read the current input channel until EOF, and print the square of every item which is a number:</lang> Test: <lang PicoLisp>(while (read)

  (when (num? @)
     (println (* @ @)) ) )

abc # Not a number 7 # Number 49 # -> print square xyz # Not a number 3 # Number 9 # -> print square</lang>

Python

Pythons REPL has _. <lang python>>>> 3 3 >>> _*_, _**0.5 (9, 1.7320508075688772) >>> </lang>

Racket

Racket doesn't have a "built-in" concept of a topic variable, but one is easy to add to the language with some macros. In fact, the subject of adding such a facility to a language using a hygienic macro facility is a very popular topic in some macro circles, and Racket can do it very well using syntax parameters. In the following there is a demonstration of two implementation approaches, the first uses a "parameter" -- a runtime value that is bound to a value in some dynamic extent, and the second uses a "syntax parameter" which is something that refers indirectly to a plain binding, and this binding can be adjusted by macros to point at an existing "real" binding. See the end of the code for usage samples. (Note that there is no point to talk about how these things behave wrt scope: since Racket is flexible enough to implement these with very different scopes...)

<lang racket>

1. lang racket

(module topic1 racket

 ;; define $ as a "parameter", but make it look like a plain identifier
 (provide $ (rename-out [$if if] [$#%app #%app]))
 (define current-value (make-parameter #f))
 (define-syntax $
   (syntax-id-rules (set!)
     [(_ x ...) ((current-value) x ...)]
     [(set! _ val) (current-value val)]
     [_ (current-value)]))
 ;; make an `if' form that binds it to the condition result
 (define-syntax-rule ($if C T E)
   (parameterize ([current-value C])
     (if $ T E)))
 ;; function application with []s uses the topic variable for the first arg
 (define-syntax ($#%app stx)
   (syntax-case stx ()
     [(_ f x y ...) (equal? #\[ (syntax-property stx 'paren-shape))
      #'(parameterize ([current-value x]) (f y ...))]
     [(_ f x ...) #'(f x ...)])))

(module topic2 racket

 ;; better variant: define `$' as a syntax parameter, which is adjusted to an
 ;; actual local binding; make it work in `if', and have a function definition
 ;; form that binds it to the actual arguments
 (provide $ (rename-out [$if if]) defun)
 (require racket/stxparam)
 (define-syntax-parameter $ (λ(stx) (raise-syntax-error '$ "not in scope")))
 (define-syntax-rule ($if C T E)
   (let ([c C]) (syntax-parameterize ([$ (make-rename-transformer #'c)])
                  (if c T E))))
 (define-syntax-rule (defun name body ...)
   (define (name arg)
     (syntax-parameterize ([$ (make-rename-transformer #'arg)])
       body ...)))
 )

(module sample1 racket

 (require (submod ".." topic1))
 (if (memq 2 '(1 2 3)) (cadr $) 'missing)
 ;; => 3
 (define (foo) (list (sqrt $) (* $ $)))
 [foo 9]
 ;; => '(3 81)
 )

(require 'sample1)

(module sample2 racket

 (require (submod ".." topic2))
 (if (memq 2 '(1 2 3)) (cadr $) 'missing)
 ;; => 3
 (defun foo (list (sqrt $) (* $ $)))
 (foo 9)
 ;; => '(3 81)
 )

(require 'sample2) </lang>

REXX

With this new definition of topic variables, the closest thing REXX has to a topic variable is probably function/subroutine arguments being "passed" to the target function/subroutine/routine/procedure. <lang rexx>/*REXX pgm shows something close to a "topic variable" (for funcs/subs).*/ parse arg N /*get an arg from CL, maybe a 3? */ say mysub(N) ' ◄───' /*invoke a function to square it.*/ exit /*stick a fork in it, we're done.*/ /*──────────────────────────────────MYSUB subroutine (function)─────────*/ mysub: return arg(1)**2 /*return the square of passed arg*/</lang> output   when the following input is used:   3

9   ◄───

Ruby

In Ruby the topic variable is $_ (same as Perl).

<lang ruby>while DATA.gets # assigns to $_ (local scope)

 print             # If no arguments are given, prints $_

end __END__ This is line one This is line two This is line three</lang>

Output:

This is line one
This is line two
This is line three

example: <lang ruby>DATA.gets p [$_.to_i ** 2, Math.sqrt($_.to_i)] #=> [9, 1.7320508075688772] __END__ 3</lang> The style of programming using $_ as an implicit parameter is gradually losing favor in the Ruby community.

Sidef

The underscore (_) topic variable is defined at compile-time in every block of a program. To call a method on it, we can just use the prefix dot (.) operator, followed by a method name, which is equivalent with _.method_name <lang ruby>say [9,16,25].map {.sqrt}; # prints: [3, 4, 5]</lang>

Standard ML

The SML language itself does not define a topic variable, but interactive implementations may define their own. For example the SML/NJ REPL defines a topic variable named it which is bound any time the user types an expression in the REPL (as opposed to a declaration).

<lang sml>- 3.0; val it = 3.0 : real - it * it; val it = 9.0 : real - Math.sqrt it; val it = 3.0 : real -</lang>

UNIX Shell

The shell $? is a kind of limited topic variable that holds the return value of the last function called. However, using it in a function will change its value, so following the echo below, the dollarhook will now contain the return value of zero indicating a successful echo:

<lang sh>multiply 3 4 # We assume this user defined function has been previously defined echo $? # This will output 12, but $? will now be zero indicating a successful echo</lang>

zkl

No topic variable pre se (a variable name can be a single character however), but underscore has a special meaning in some cases. Its use is scoped to the expression it is used in. <lang zkl>a,_,c:=List(1,2,3,4,5,6) //-->a=1, c=3, here _ is used as "ignore" 3.0 : _.sqrt() : println(_) //-->"1.73205", _ (and :) is used to "explode" a computation

                           // as syntactic sugar

1.0 + 2 : _.sqrt() : _.pow(4) // no variables used, the compiler "implodes" the computation

   // --> 9

</lang>