Re: OO Design induces an existential crisis

On 29 Jul 2005 04:08:57 -0700, hansewetz@xxxxxxxxxxx wrote:

>AndyW wrote:
>> On 27 Jul 2005 06:27:19 -0700, hansewetz@xxxxxxxxxxx wrote:
>>
>> >AndyW wrote:
>> >> On Sun, 10 Jul 2005 10:49:01 +0200, "Ilja Preuß" <it@xxxxxxxxxxxxx>
>> >> wrote:
>> >>
>> >> >AndyW wrote:
>> >
>> >> Identity: Coffee cup.
>> >
>> >> This is a mind boggler - because in purism the cup only has one state
>> >> and that is that it exists. This is the state apart from the one
>> >> implied by its intended usage. Currently the one on my desk merely
>> >> exists since its not being interacted with.
>> >
>> >This idea of 'purism' sounds like pure nonsense. When you model a
>> >problem, using OO or non OO methods, you base your model on your
>> >perception of the world. Depending on the context and your objective
>> >you will model your cup in different ways.
>> >
>> I suspect that you missed the very first post in the thread that sets
>> out the model being used in the rest of the thread.
>>
>> If you had read it, then it sets out that there are different
>> techniques used when modeling pure object, abstractions, services and
>> mixins. The main technique is to understand what each is and learn
>> to recognise them, then use the appropriate technique to model them.
>
>I have started to believe that this discussion is not so useless after
>all. However, it is not because of the content it is useful but because
>of its lack of content. It many ways it is a poster child for some
>other discussions on the newsgroup.
>
>You are stating some ideas about how to model problems/solutions using
>objects, abstractions etc. What is missing is a 'description' off
>what these things are. I admit you supply a half baked description
>based on 'tangibility' that does not hold up after a few seconds
>scrutiny. Your ideas also lack an 'explanation' for why they should
>work, i.e., help in building better software.

Really.

Actually, i'm using an extension of the Booch model (identity, state
and actions). Its been extended with tangibility to define the
difference between an object, service and mixin (the latter two are
not tangible).

services and mixins come from the taligent object model. They are
used to classify two different types of entity - something that is
data only and something that only has functions (and no data).

This concept is also used heavily in most Framework descriptions and I
suspect in many patterns books.

It is based on both the fundemental rules of classical OO and also the
fundemantal rules of human cognition.


.