From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-0.8 required=5.0 tests=BAYES_00,INVALID_DATE
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,93fa00d728cc528e
X-Google-Attributes: gid103376,public
X-Google-Thread: 1108a1,93fa00d728cc528e
X-Google-Attributes: gid1108a1,public
X-Google-ArrivalTime: 1994-11-03 10:40:37 PST
Path: 
 bga.com!news.sprintlink.net!howland.reston.ans.net!cs.utexas.edu!convex!news.duke.edu!eff!blanket.mitre.org!linus.mitre.org!linus!mbunix!eachus
From: eachus@spectre.mitre.org (Robert I. Eachus)
Newsgroups: comp.lang.ada,comp.object
Subject: Re: SOLVED! Decoupled Mutual Recursion Challenger
Date: 2 Nov 94 14:56:02
Organization: The Mitre Corp., Bedford, MA.
Message-ID: <EACHUS.94Nov2145602@spectre.mitre.org>
References: <1994Oct18.221751.15457@swlvx2.msd.ray.com>
 <CxwAuy.1CL@inmet.camb.inmet.com>
	<38289r$79m@oahu.cs.ucla.edu> <1994Oct19.143843.372@wdl.loral.com>
	<38fi4r$l81@oahu.cs.ucla.edu>
	<1994Oct24.174231.1897@swlvx2.msd.ray.com>
	<38hcv3$j85@baleen.cs.ucla.edu>
	<1994Oct25.155420.27353@swlvx2.msd.ray.com>
	<38pulp$ovg@oahu.cs.ucla.edu> <EACHUS.94Oct28105216@spectre.mitre.org>
	<38rguj$9cd@oahu.cs.ucla.edu>
NNTP-Posting-Host: spectre.mitre.org
In-reply-to: jmartin@oahu.cs.ucla.edu's message of 28 Oct 1994 11:46:43 -0700
Xref: bga.com comp.lang.ada:7480 comp.object:8157
Date: 1994-11-02T14:56:02+00:00
List-Id: <comp.lang.ada>

In article <38rguj$9cd@oahu.cs.ucla.edu> jmartin@oahu.cs.ucla.edu (Jay Martin) writes:

  > Could you give an example of how you could shoot yourself in the foot
  > using the package forward idea.  How do you implement mutual recursion
  > using straight Ada83 derived types?  The abstract type solution and
  > downcasting are not deemed very exceptable by many of the lang.ada
  > participants.  Jay

    The representations CAN'T be mutually recursive without some
variant records and other tricks.  The package forward idea would
allow such declarations. (There is no easy place to decide that you
have such a loop.)  On the other hand, even in Ada 83 it is possible
to defer the actual declaration of the record object corresponding to
the type to the body of the package.  If you do this, you get pointer
semantics for objects of the type, which may have to be adjusted in
the interface, but--especially if you make the type limited--that is
not a major issue.

    Yes, yes.  If you want to have a primitive operation on Office
which returns a Person and vice-versa, you can do it in one package
spec, or three, or four, but not two.  I still fail to understand the
degree of mouth frothing directed at this "weakness" of Ada.  As we
have seen in the past several weeks, Ada provides lots of different
ways of modeling this system.  That there are some models in other
languages which can be duplicated more easily than others should not
come as a surprise.  Let's focus on the "right" way to do it in Ada
instead.


--

					Robert I. Eachus

with Standard_Disclaimer;
use  Standard_Disclaimer;
function Message (Text: in Clever_Ideas) return Better_Ideas is...