- From: <uche.ogbuji@fourthought.com>
- Date: Thu, 18 Jan 2001 18:43:13 -0700
- To: Arnold deVos <adv@langdale.com.au>
- cc: Dan Brickley <Daniel.Brickley@bristol.ac.uk>, Dave Beckett <dave.beckett@bristol.ac.uk>, Dennis van der Laan <LaanD@vertis.nl>, www-rdf-interest@w3.org
> On Fri, 19 Jan 2001 06:18, uche.ogbuji@fourthought.com wrote: > > > > OK, here's a question I seem to have to ask very frequently, but to which I > > never seem to get a useful answer. What on Earth is wrong with multiple > > inheritance? > > > > There is a need to distinguish between code and (data) > models on this point. There is nothing at all wrong with MI in data > modelling IMHO. In code there are issues, for example, the interpretation of > repeated inheritance. Oh, there's a lot more to distinguish than just that. There is a need to distinguish between interitance, subclassing and subtyping (I used to think this was a quiddity until I read an excellent research paper the cite of which I can't dig up right now). I point this out because my possibly wrong guess is that you're talking about code inheritance but RDF subclassing. But I think that even in code Multiple Inheritance is not automatically bad. One form of inheritance is mixing-in, which does not have the "diamond of doom" problems that you're hinting at with the "repeated inheritance" comment. Basically, MI is a tool. It can be used well or poorly, but again I await anyone who can show that it is axiomatically bad. However, since all here seem to agree that multiple subclassing in RDF is not, after all, a bad thing, I'm happy to drop the matter. Thanks. -- Uche Ogbuji Principal Consultant uche.ogbuji@fourthought.com +1 303 583 9900 x 101 Fourthought, Inc. http://Fourthought.com 4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA Software-engineering, knowledge-management, XML, CORBA, Linux, Python
Received on Thursday, 18 January 2001 20:43:26 UTC