Re: Clean Object Class Design -- Circle/Ellipse
Date: Tue, 21 Aug 2001 03:23:54 GMT
Message-ID: <3b81d266.3858753634_at_news.grpvine1.tx.home.com>
On Mon, 20 Aug 2001 03:19:30 GMT, "Richard MacDonald"
<macdonaldrj_at_att.net> wrote:
>> I think I mentioned it in some earlier post.
>> Magnitude provides axiomatization of the ordering relationship ("<").
>> Number (derived from Magnitude) overrides "< " appropriately for its
>> domain.
>>
>Got it. You moved that up a little higher again. (I am a Smalltalker.)
>I'm still lost on the issue of the operations that "jump outside" the
>argument's domain. Are you saying that any operation that does so
>must move up the hierarchy until it encompasses both the argument
>and the result? So you disallow the "jump"?
I'm not sure about the meaning you assign to "jump". The way I prefer to put it is to say that one specialization/implementation should not redefine another. They all should be (parallel) implementations of a generic (ala CLOS) function.
Marc Gluch
Mindtap Inc.
Received on Tue Aug 21 2001 - 05:23:54 CEST