Re: Clean Object Class Design -- Circle/Ellipse

From: Universe <universe_at_saltmine.radix.net>
Date: 21 Jul 2001 01:25:25 GMT
Message-ID: <9jalm5$76s$2@news1.Radix.Net>

In comp.object S Perryman <q_at_q.com> wrote:
> <D_at_B.A> wrote in message news:QLG57.872$Y47.13304_at_www.newsranger.com...
>> In article <9TF57.167$%16.43276332_at_radon.golden.net>, Bob Badour says...  

>> 1. Is the case Square/Rectangle really similar to Circle/Ellipse?
>> [...]

>
> It is not similar. It is the *same* .
> The problem that occurs when taking the spec into a solution, and therefore
> coming to
> grief with the Liskov Substitution Principle (LSP) , was called by Peter
> Wegner
> "subset subtyping" .

Anyone who comes to LSP grief by applying domain objects and relationships to the solution is an OO modelling lame. It should be clear from the Sqaure/Rectangle discussion that we can physical design, and implement via OO code, any given analysis/specification/problem space object interactions. There is no "wall" or barrier to this, except the arteriosclerotic, insipidly limited and technocratic, Cheney, craftite way of seeing objects, software and the world. :-}

Elliott

>
> Go to Google etc to search comp.object for the full name of the paper by
> Wegner
> that discusses the problem. Follow the thread to see what potential impl
> solutions can
> be used to alleviate the problem
>
> BTW, not sure why comp.theory is on the list.
> But be warned, once you get those guys involved, what you'll get as "theory"
> will
> make the comp.object debates look like Janet and John book 2 ... :-)
>
>
> Regards,
> Steven Perryman
>
>
 

-- 
http://www.radix.net/~universe
    "Don't Need No Right wing, Craftite, Reaganite Uncles"
Enjoy! * Hail OO Modelling * Hail the Wireless Web * @Elliott 2001

Received on Fri Jul 20 2001 - 20:25:25 CDT