From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <jon@jdh30.plus.com>
Delivered-To: caml-list@yquem.inria.fr
Received: from nez-perce.inria.fr (nez-perce.inria.fr [192.93.2.78])
	by yquem.inria.fr (Postfix) with ESMTP id 5F197BB91
	for <caml-list@yquem.inria.fr>; Wed, 26 Jan 2005 15:30:19 +0100 (CET)
Received: from ptb-relay03.plus.net (ptb-relay03.plus.net [212.159.14.214])
	by nez-perce.inria.fr (8.13.0/8.13.0) with ESMTP id j0QEUI4I012555
	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO)
	for <caml-list@yquem.inria.fr>; Wed, 26 Jan 2005 15:30:19 +0100
Received: from [80.229.56.224] (helo=chetara)
	 by ptb-relay03.plus.net with esmtp (Exim) id 1CtoTN-000IXk-7y
	for caml-list@yquem.inria.fr; Wed, 26 Jan 2005 14:48:45 +0000
From: Jon Harrop <jon@jdh30.plus.com>
Organization: University of Cambridge
To: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] 'a Set?
Date: Wed, 26 Jan 2005 09:13:08 +0000
User-Agent: KMail/1.7.1
References: <727068A7-6F2C-11D9-8411-0003939A19AA@fas.harvard.edu>
In-Reply-To: <727068A7-6F2C-11D9-8411-0003939A19AA@fas.harvard.edu>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200501260913.09018.jon@jdh30.plus.com>
X-Miltered: at nez-perce with ID 41F7A97A.000 by Joe's j-chkmail (http://j-chkmail.ensmp.fr)!
X-Spam: no; 0.00; caml-list:01 wrote:01 functor:01 elt:01 higher-order:01 bool:01 bool:01 rec:01 rec:01 closures:01 cheers:01 polymorphic:01 functions:01 functions:01 argument:01 
X-Spam-Checker-Version: SpamAssassin 3.0.0 (2004-09-13) on yquem.inria.fr
X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=disabled 
	version=3.0.0
X-Spam-Level: 

On Tuesday 25 January 2005 23:54, Mike Hamburg wrote:
> Is there a clean way to do this without removing the code from set.ml
> and modifying it?

I do not believe so. I have also had to do this.

Compared to a flat set of functions, the functor approach has the advantage of 
enforcing a consistently used compare function. The same effect can be 
achieved with "elt = 'a" by writing a higher-order function which returns a 
record containing the Set.* functions using the given function argument as 
the compare function. Something equivalent to this:

  type 'a t = 'a list
  type 'a set =
    { empty : 'a t;
      is_empty : 'a t -> bool;
      add : 'a -> 'a t -> 'a t;
      mem : 'a -> 'a t -> bool }

  let rec add compare e = function
    [] -> [e]
  | h :: t -> match compare h e with
      -1 -> e :: h :: t
    | 0 -> e :: t
    | _ -> h :: add compare e t
  let rec mem compare e = function
    [] -> false
  | h :: t -> match compare h e with
      -1 -> false
    | 0 -> true
    | _ -> mem compare e t

  let make ?(compare=compare) () =
    { empty = [];
      is_empty = (fun s -> s=[]);
      add = add compare;
      mem = mem compare }

Possible issues with this are that building closures (i.e. in "make") is 
expensive and that the resulting type is monomorphic ('_a). You can probably 
get a polymorphic type most easily by putting the definitions of "add" etc. 
in the record definition, rather than partially applying their arguments.

Cheers,
Jon.