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: 103376,18f6de557e6897b2
X-Google-Attributes: gid103376,public
From: "John G. Volan" <johnvolan@sprintmail.com>
Subject: Re: circular unit dependency
Date: 1997/06/03
Message-ID: <33944BC7.969@sprintmail.com>#1/1
X-Deja-AN: 245867469
References: <3386d96f.171920@noticias.ibernet.es>
 <9A7E8196B8D7EE83.E6C868B798076E45.6F1AD9E8B3E01F66@library-proxy.airnews.net>
 <mheaney-ya023680003105972352110001@news.ni.net>
 <33932F31.4399@sprintmail.com>
 <mheaney-ya023680000206972122410001@news.ni.net>
 <33942317.2BCD@pseserv6.fw.hac.com>
Reply-To: johnvolan@sprintmail.com
Newsgroups: comp.lang.ada
Date: 1997-06-03T00:00:00+00:00
List-Id: <comp.lang.ada>


W. Wesley Groleau (Wes) wrote:
> 
> John,
> Your example is reasonable, and a solution would be a good thing,
> but the problem is not a disaster.  I can only remember once in
> over ten years that I wished I could separately package two types
> but couldn't.

Of course it's not a disaster. There are, after all, a variety of
avoidance techniques, some of which I enumerate in section [3] of my
FAQ.  I think what happens is that most Ada programmers naturally fall
into the technique suggested by [3.2]: Arbitrarily choose one class to
be client-only and the other to be supplier-only, force the association
to be one-way.  I suspect many Ada programmers are not even conscious
that they are doing that. I exploit it whenever I can.  It can let you
go quite a long time without ever facing the Day of Mutual Reckoning.
:-)  However, this kind of one-way biasing is somewhat artificial and
stilted when, conceptually, the object classes are really peers.  The
alternative, mutually dependent packages, only _seems_ awkward because
of this quirk (or rather, this hole) in the language. If Ada already had
one of the enhancements that have been suggested (see section [5]), I
wager we wouldn't be having any of these religious debates about how
packages were "meant" to be used.

I just think, as we begin entering an era of increasingly large and
complex domains (e.g., distributed client/server object systems), these
kinds of situations will increase in frequency.  It would be a good idea
to have all our ducks in a row. I believe that separately encapsulated
yet mutually dependent object classes are a legitimate design pattern,
one that ought to be available as an option within every engineer's
repertoire.

------------------------------------------------------------------------
Internet.Usenet.Put_Signature 
  (Name       => "John G. Volan",
   Employer   => "Texas Instruments Advanced C3I Systems, San Jose, CA",
   Work_Email => "johnv@ti.com",
   Home_Email => "johnvolan@sprintmail.com",
   Slogan     => "Ada95: World's *FIRST* International-Standard OOPL",
   Disclaimer => "My employer never defined these opinions, so using " & 
                 "them would be totally erroneous...or is that just "  &
                 "nondeterministic behavior now? :-) ");
------------------------------------------------------------------------