Re: Transformation of object modells to relational structures

In <4p5bcp$bhj_at_wormer.fn.net> keystrk_at_feist.com (Jack L. Swayze Sr.) writes:
>
>answers_at_ix.netcom.com(Steve Long) wrote:
>
>>Or is it that you do not understand ERDs?
>
>Umm, well. I have been earning my income as a Data Moder (or Data
>Analyst, or Data Administrator) for almost 11 years now. I was ERD
>data modling before CASE tools were even heard of (which was before I
>was being paid for it). I have been exposed to at least 10 different
>ERD modeling representations (some of which, I must admit, I now
>forget). And (if the above isn't enough) I participated in the USAF
>contract that extedned the IDEF-1 modeling paradigm into IDEF-1X
>(which is what ERWin uses as its default diagrammatic technique).
>
Doesn't answer the question.

>>Can you substantiate your statement? Provide an object model that is
>>not equivalent to an ERD with an extention of adding methods as a
>>counter example and your statement will have more merit than mere
>>opinion.
>
>Sure, the one that made me realize the difference was the Fusion
>modeling paradigm (as created by the firm: Object Space). I even had
>one of the representitives of that firm tell me to my face that
>representing data requirements based on existence is diametrically
>opposed to the Object modeling approach which represents data
>requirements based on common (or different) 'responsibilities'
>(meaning functional responsibility).
>
I missed the counter example; is that another opinion in there?

>In other words: if you can imagine someone totally ignoring the data
>requirements, designing their system to be Highly modular, then
>determining that the best data structures is to have one and only one
>data structure for each and every process module, then 'encapsulating'
>that data structure so that the only way to get to the data structure
>is through that module you have designed - that is what the
>Smalltalk-like OO model is similar (but not identical) to.
>
What you have presented here is functional decomposition with extensions for data modeling. Another twist, but is it OO?

>>Steve
>>804-262-6332
>you didn't really want me to call you, did you?
>(if so, can I reverse the charges?)

Use your judgement.

>>--------------------
>
>------------------------------------------------------------
>
>'Keystroke'
>KeystrkTX_at_AOL.COM
>
Received on Thu Jun 06 1996 - 00:00:00 CEST