NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/comp/lang/pascal/6685

Internet Message Format | 1992-11-18 | 2.0 KB

Xref: sparky comp.lang.pascal:6685 comp.object:4257
Newsgroups: comp.lang.pascal,comp.object
Path: sparky!uunet!clarkson!news
From: [e-mail] (Harry Kooijman,Peyt314a,3808)
Subject: Re: BP/TP OOP is missing something...
Message-ID: <[e-mail]>
Sender: [e-mail]
Nntp-Posting-Host: sun.soe.clarkson.edu
Organization: Clarkson University
References: <[e-mail]>
Date: Wed, 18 Nov 1992 16:13:17 GMT
Lines: 30

From article <[e-mail]>, by [e-mail] (Duncan Murdoch):
> In article <[e-mail]> [e-mail] (Markku Sakkinen) writes:
>
>>In any flavour of OO Pascal without such parameters, the constructor
>>of a general matrix class (i.e. without fixed dimensions) would most
>>naturally create only a "handle" for an actual matrix.
>
> This doesn't seem like a very clean solution.  Does it mean that I'd have
> an array without a size after the automatic initialization?  Then I'd have
> to send a size message to it before I could use it?  It sounds to me as
> though the size message is the real initialization; automatic initialization
> doesn't work here, because it doesn't produce a usable object (i.e.
> establish the invariants of the class, as you said).
>

The language does not force you to have an "empty" constructor and that
the actual initialization (here for the matrix size) is done later, but
allows you to do both. Now that's freedom of choice! Actually it allows
classes/objects that size themselves according to the (local) needs. So
you can say I want a vector, use it for an array of size x, then
increase its size to y, use it, then decrease its size to z and use it.
When the vector goes out of scope, it knows what it last size was and
frees the memory it was occupying at destruction.



--
Harry Kooijman: Rigorous dynamic distillation simulation.
Dept. of Chemical Engineering, Clarkson University, Potsdam NY 13699.
[e-mail]