W3C home > Mailing lists > Public > public-qt-comments@w3.org > January 2010

[Bug 8662] [XQuery 1.1] Need an easy way to define a recursive inline function

From: <bugzilla@wiggum.w3.org>
Date: Wed, 06 Jan 2010 18:26:08 +0000
To: public-qt-comments@w3.org
Message-Id: <E1NSaa8-00066d-Mf@wiggum.w3.org>
http://www.w3.org/Bugs/Public/show_bug.cgi?id=8662





--- Comment #1 from John Snelson <john.snelson@oracle.com>  2010-01-06 18:26:08 ---
1) Construct to Define Local Recursive Functions
   ---------------------------------------------

There are two related possibilities here:

a. Allow definition of named, scoped, local functions, ie:

declare function outer($a)
{
  declare function f1($b)
  {
    $a, f2($b)
  };

  declare function f2($b)
  {
    $b, f1($b)
  };

  f1#1;
};

b. Many functional languages have a recursive "let" whose variables are all in
scope for any of the bound expressions:

letrec $f1 := function($b) { $a, $f2($b) },
       $f2 := function($b) { $b, $f1($b) }
return $f1

I think I'm personally in favour of a solution like this.

2) Reference to self
   -----------------

I dislike the idea of a special named function which is only in scope
sometimes, and has a different type signature depending on where it is in
scope. If we went this route, I would prefer a special variable like $this or
$self of type function(*), which contains a reference to the current function.

Of course as you point out, this would not solve the mutually recursive
function problem.

3) Recursive Global Variables
   --------------------------

Another (possibly tangential but related) issue is the rules about not allowing
a global variable that references itself. It seems odd to allow this:

declare function f() { 1, f() };

but not this:

declare variable $f := 1, $f;

Both the function f() and the variable $f evaluate perfectly well on an
implementation using lazy evaluation.

Maybe we could make our prolog rules simpler by allowing variables to refer to
themselves.


-- 
Configure bugmail: http://www.w3.org/Bugs/Public/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
Received on Wednesday, 6 January 2010 18:26:10 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:45:41 UTC