From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <lealanko@cs.helsinki.fi>
X-Original-To: caml-list@yquem.inria.fr
Delivered-To: caml-list@yquem.inria.fr
Received: from mail2-relais-roc.national.inria.fr (mail2-relais-roc.national.inria.fr [192.134.164.83])
	by yquem.inria.fr (Postfix) with ESMTP id 29830BC57
	for <caml-list@yquem.inria.fr>; Fri, 17 Dec 2010 10:19:52 +0100 (CET)
X-IronPort-Anti-Spam-Filtered: true
X-IronPort-Anti-Spam-Result: ApMBAJe5Ck2A1gkBe2dsb2JhbACWFo4/AQEWIgUfw2eFSgQ
X-IronPort-AV: E=Sophos;i="4.59,360,1288566000"; 
   d="scan'208";a="84283920"
Received: from courier.cs.helsinki.fi (HELO mail.cs.helsinki.fi) ([128.214.9.1])
  by mail2-smtp-roc.national.inria.fr with ESMTP/TLS/DHE-RSA-AES256-SHA; 17 Dec 2010 10:19:51 +0100
Received: from melkinpaasi.cs.helsinki.fi (melkinpaasi.cs.helsinki.fi [128.214.9.14])
  (AUTH: PLAIN cs-relay, TLS: TLSv1/SSLv3,256bits,AES256-SHA)
  by mail.cs.helsinki.fi with esmtp; Fri, 17 Dec 2010 11:19:50 +0200
  id 00093E78.4D0B2B36.000047F8
Received: by melkinpaasi.cs.helsinki.fi (Postfix, from userid 37211)
	id A3AA48DF5F; Fri, 17 Dec 2010 11:19:50 +0200 (EET)
Date: Fri, 17 Dec 2010 11:19:50 +0200
From: Lauri Alanko <la@iki.fi>
To: caml-list@inria.fr
Subject: Unimplemented modules as top-level signatures
Message-ID: <20101217091950.GA28557@melkinpaasi.cs.helsinki.fi>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
User-Agent: Mutt/1.5.20 (2009-06-14)
X-Spam: no; 0.00; lauri:01 alanko:01 ocaml:01 complicates:01 foo:01 mli:01 foo:01 pitfalls:01 functors:01 lauri:01 alanko:01 signatures:01 minor:01 modules:02 defining:02 

A minor annoyance in OCaml is that although you can (and must) put
each top-level module into its own file, there is no corresponding
mechanism for defining module types on their own: you always have to
put a "module type" definition inside another module, which
complicates scoping, and with standard tools adds an extra level of
indentation to the signature definition.

However, with 3.12 there is solution of sorts: write the signature
definition in foo.mli without an accompanying foo.ml, and then refer
to the signature as "module type of Foo".

Is this horrible style? Are there some pitfalls I should be aware of?
Or are there better solutions to my desire to avoid nested scoping?
Top-level functors would also be nice to have... :)


Lauri Alanko
la@iki.fi