home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!uknet!edcastle!dcs.ed.ac.uk!rjp
- From: rjp@dcs.ed.ac.uk (Rob Pooley)
- Newsgroups: comp.object
- Subject: Re: Object-Oriented Methodologies - Class Specifications
- Message-ID: <43376@skye.dcs.ed.ac.uk>
- Date: 7 Sep 92 13:17:23 GMT
- References: <43008@skye.dcs.ed.ac.uk> <715276480.1.p00058@mail.psi.net> <1992Sep1.220559.10346@m.cs.uiuc.edu> <1992Sep2.140048.11829@bcrka451.bnr.ca>
- Sender: nnews@dcs.ed.ac.uk
- Organization: Edinburgh University
- Lines: 36
-
- In article <1992Sep2.140048.11829@bcrka451.bnr.ca>, sjm@bcrki65.bnr.ca (Stuart MacMartin) writes:
- >
- > It is possible that "the context of its use" is that it be generally
- > useful. It is not intended for one application only. It is a utilitarian
- > foundation class, application independent. Such classes are necessary,
- > obviously. It makes sense to separate the application-specific needs from
- > the generic building blocks.
- >
- > Your argument appears to preclude designing classes for re-use. Or have I
- > misunderstood your comments?
- >
-
- That's a bit more extreme than I intended. My experience seems to show that it
- is rarely useful to design at too abstract a level. This may in fact make it
- harder to re-use the class.
-
- What I was trying to say was that there are probably limits to re-use and that
- it is better to be clear about the context in which you intend such re-use to
- take place. Geometrical shapes are a particularly slippery concept to deal
- with, since they can be intended for several different purposes and so
- understood in different ways.
-
- Inheritance is not "Platonic", i.e. you don't start from a pure definition and
- add compromises in terms of the world you use the concept in, do you? If not
- then that means of re-use is inherently rooted in the context of its intended
- end use.
-
- Re-use in composition may be a better candidate for a pure view of classes,
- but then the use of the objects derived from a class are defined in the
- aggregating class and the object itself merely carries enough information to
- define itself. Thus a circle is a tuple of radius and (possibly but not
- necessarily) a centre.
-
- Rob Pooley,
-
- Fading into the shades of imperfection
-