Re: Conceptual, Logical, and Physical views of data

From: mAsterdam <mAsterdam_at_vrijdag.org>
Date: Fri, 02 Sep 2005 01:09:08 +0200
Message-ID: <43178978$0$11071$e4fe514c@news.xs4all.nl>

dawn wrote:
> mAsterdam wrote:

>>David Cressey wrote:
>>>mAsterdam wrote:
[snip agreement]

>>>>Does it matter if the technical
>>>>architecture is given or not (say: we will use DBMS xyz)?
>>>Here's the way it works in practice for me:
>>>The conceptual model is implementation independent.
>>
>>Agreed.

>
> I would agree that is the idea, but it seems rare that it works out
> this way.

When the implementation discussions slow down, the differences at the conceptual level become clear. Q: What's the difference between a terrorist and a methodologist? A: You can reason with a terrorist.
(C) Martin Fowler

>>But what does it mean?
>>
>>1. If there is no explicit conceptual model
>>in an actual project different people will
>>assume different models (not just homonym
>>synonym stuff - ever tried modelling after
>>a series of take-overs? Assumptions go deep).

>
> so true

:-)

>>2. How do we make the conceptual model explicit?

>
> This is probably a naive response, but where I used to choose an early
> prototyping tool, possibly paper & pencil, even in the analysis phase
> of a project, I now use web pages. I can mock these up even in place
> of showing any end-user an erd (or uml class diagram) with any level of
> detail in it. If you model the conceptual data with web pages with
> foreign keys turned into links and sample data values, you can get a
> lot of bang for the buck.
>
> I haven't done that from an existing implementation, but I can imagine
> doing so and I would think it could be similarly useful. Then if you
> want it all on one (giant) page, use a web diagramming tool to show the
> pages and their links. One page is one (typically strong) entity. You
> can use various xhtml features to include the definitions of terms on
> mouseover (I haven't done that).
>
> You might be looking for a more formal modeling approach, in which case
> others will know a lot more than I.

You are the mathematician. I'm not. I'm just an experienced DB guy. (With a good methodologigal (J.J. Klant) education, though. Main (not full) disclosure: economics with a math accent).

>>Is there an effective formalism which can serve >>as a modelling language before the logical model?

> I'm guessing that saying I can do what I need using xhtml makes me
> sound on the not-exactly-professional side, but so be it.
>

>>ORM? (Object Role Models - aside:
>>some thought I was talking about Object Relation
>>Mapping - a non-issue).
>>
>>>The logical model is data model dependent  (relational v. object oriented),
>>>but independent of product, volume, load, and resources.

>
> That seems close to what happens, although plenty of people seem to
> think that the logical model is db independent. --dawn

It is not. Just think of how to handle the decision of choosing a primary key between the alternatives (another thread). Not necessary in say a CODASYL context.

>>>The physical model is dependent on all of the above. Received on Thu Sep 01 2005 - 18:09:08 CDT