home *** CD-ROM | disk | FTP | other *** search
- Path: access4.digex.net!not-for-mail
- From: ell@access4.digex.net (Ell)
- Newsgroups: comp.lang.c++
- Subject: Re: Pure Virtual Destructor Question
- Date: 18 Feb 1996 23:23:40 GMT
- Organization: The Universe
- Message-ID: <4g8cds$7do@news4.digex.net>
- References: <4fas7a$7ns@comet2.magicnet.net> <4fecq0$k4e@news4.digex.net> <rcauvin-0902960901140001@quadostimpy.natinst.com> <4fp5r2$dm2@news4.digex.net> <rcauvin-0602960913040001@quadostimpy.natinst.com> <4fqmp5$pbi@stc06.ctd.ornl.gov>
- NNTP-Posting-Host: access4.digex.net
- X-Newsreader: TIN [UNIX 1.3 950824BETA PL0]
-
- Matt Kennel (kennel@msr.epm.ornl.gov) wrote:
- : Roger L. Cauvin (rcauvin@natinst.com) wrote:
- : > 2. The abstractness of a class is conceptual and does not depend on it
- : > declaring a function with no implementation.
-
- The ARM and DWP say an abstract class is where at least one function is
- declared with the pure specifier '=0'.
-
- : > All that abstractness
- : > implies is that it is not appropriate to instantiate an object of the
- : > class directly, since the class is incompletely defined.
-
- I'll stick with the ARM and DWP on this one.
-
- : > Providing a
- : > default implementation for a function, but declaring it pure virtual,
- : > indicates that the implementation is incomplete and must be completed by
- : > derived classes.
-
- However, the defintion should be "complete" for the abstract class where
- the pure virtual was declared.
-
- : Perhaps then, we should call these "incomplete" or "partial" classes,
- : instead of "abstract classes". This better expresses the notion of
- : "not instantiable" rather than "provides specification".
- :
- : I would prefer to reserve the term "abstract class" for classes which
- : only specify and do not implement anything.
-
- Again, I'll stick with the ARM and DWP on what is or isn't an "abstract
- class".
-
- Elliott
-