Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!microsoft!jimad
From: jimad@microsoft.UUCP (Jim ADCOCK)
Newsgroups: comp.lang.c++
Subject: Re: asking an object for its type
Message-ID: <70951@microsoft.UUCP>
Date: 28 Feb 91 17:37:21 GMT
References: <23984@netcom.COM> <1190@sheol.UUCP> <1991Feb19.000449.22255@gpu.utcs.utoronto.ca> <607@taumet.com> <65451@brunix.UUCP> <11284@pasteur.Berkeley.EDU> <70870@microsoft.UUCP> <8608@exodus.Eng.Sun.COM>
Reply-To: jimad@microsoft.UUCP (Jim ADCOCK)
Organization: Microsoft Corp., Redmond WA
Lines: 23

In article <8608@exodus.Eng.Sun.COM> chased@rbbb.Eng.Sun.COM (David Chase) writes:
|In article <70870@microsoft.UUCP> jimad@microsoft.UUCP (Jim ADCOCK) writes:
|>A "trivial" solution to the run-time type problem ought to be closer
|>to O(1) than O(N), where N is the number of classes that need to have
|>some notion of run-time type.  ...
|
|Do note that (using algorithms known to me) run-time type-checks and
|subtype queries can be implemented much more efficiently in the
|single-inheritance case.  This doesn't directly help C++, unless you
|choose to limit the problem that you are trying to solve.

Sorry, I guess I made myself pretty unclear.  I meant to imply O(N)
programmer effort -- not O(N) time.  Given N classes, the programmer
is going to have to specify type information in O(N) places.  Clearly
this is a violation of encapsulation.  If C++ "supported" run-time
typing, the programmer ought to be able to specify the needed run-time
information in O(1) places -- in a single superclass, for example.

This could be done if C++ provided a way to specify some "templated" information
in a superclass, with that template expanded "automatically" without
programmer intervention in all subclasses.  But, the present template
definitions don't seem to allow for such a possibility.  Correct me if
I've missed something.