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=0.2 required=5.0 tests=BAYES_00,INVALID_MSGID,
	REPLYTO_WITHOUT_TO_CC autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: fac41,9a0ff0bffdf63657
X-Google-Attributes: gidfac41,public
X-Google-Thread: 1108a1,9a0ff0bffdf63657
X-Google-Attributes: gid1108a1,public
X-Google-Thread: 103376,4b06f8f15f01a568
X-Google-Attributes: gid103376,public
X-Google-Thread: f43e6,9a0ff0bffdf63657
X-Google-Attributes: gidf43e6,public
From: Loryn Jenkins <loryn@s054.aone.net.au>
Subject: Re: Software landmines (loops)
Date: 1998/09/06
Message-ID: <35F1B862.C1FFA000@s054.aone.net.au>#1/1
X-Deja-AN: 388259402
Content-Transfer-Encoding: 7bit
References: <m3ogt3qgca.fsf@mheaney.ni.net> <35f23ce2.7649859@news.erols.com>
 <6snn1b$c90$1@hirame.wwa.com> <35ef7dff.24318728@news.erols.com>
 <EyrHEC.My1@ecf.toronto.edu> <35f79e53.98130474@news.erols.com>
 <35F0B75E.F259FCE6@s054.aone.net.au> <6srcsl$qeo$1@hirame.wwa.com>
X-Priority: 3 (Normal)
Content-Type: text/plain; charset=us-ascii
X-Trace: news.mel.aone.net.au 905033866 12807 203.102.238.34 (5 Sep 1998
 22:17:46 GMT)
Organization: TekRite Pty Ltd
Mime-Version: 1.0
Reply-To: loryn@acm.org
NNTP-Posting-Date: 5 Sep 1998 22:17:46 GMT
Newsgroups: comp.lang.eiffel,comp.object,comp.software-eng,comp.lang.ada
Date: 1998-09-05T22:17:46+00:00
List-Id: <comp.lang.ada>

> There are two aspects to dependency management.  One is to isolate those
> things which ought not depend upon each other.  The other is to associate
> those things that ought to be associated.  i.e. the old duo of cohesion and
> coupling.  Your nicely worded phrase "conceptually modelling coherent sets
> of behaviors" is a good way of expressing the cohesion side of the equation;
> and I, of course, have no quarrel with it.  Objects, after all, ought to
> consist of conceptually cohesive groupings of functions and data.  Object
> models, on the other hand, should consist of an interelated set of such
> objects which have been decoupled as much as possible.

Well ... decoupled appropriately.

Perhaps this is simply a difference in emphasis. Obviously, you run into
trouble when you have bidirectional dependencies everywhere. Obviously,
you run into trouble if you have cyclic dependencies between clusters.

This is an extremely important aspect of architecture.

It just doesn't seem to me to be one of the fundamental, driving forces
of OO. (OO does enable it nicely, though.)

Loryn Jenkins