Re: Logical data design, Pick style

Oops -- this was for cdp. Sorry. --dawn

dawn wrote:
> I'll be back in the office next week Tuesday, but while I'm online I
> thought I'd toss out a question. I've done a little work on the
> following question, but have more to do and thought others might be
> interested in jumping in on this one:
>
> If you don't do logical data modeling relational model style, using
> 1NF, then how do you model the data?
>
> 1. Entity-relationship models or the analogous UML diagrams are good
> for identifying and communicating conceptual data models. The topic of
> conceptual modeling is bigger than this, but the focus of this question
> is the "logical data model" or "implementation data model."
>
> 2. Take the conceptual model and look for strong entities -- those
> that exist independent of other entities. Model those as files.
>
> 3. Identify unique keys for these files.
>
> 4. Identify some high level types for these files, including the old
> designations of master files, transaction/event files, and code files.
>
> 5. Decide whether you will pour all code files together or use one
> code file for typical abbreviation-description tables (see a previous
> discussion on this topic a few months ago).
>
> 6. Add in all properties for these entities as attributes/fields in
> these files.
>
> 7. Add in foreign keys where appropriate as well as the redundant
> "return links"
>
> 8. Look for functional dependencies and put it into Boyce-Codd normal
> form sans 1NF.
>
> I have a lot of other notes, but this should be good enough for a
> discussion starter, I hope. Your thoughts? --dawn
>
> P.S. Blog is at http://www.tincat-group.com/mewsings, latest entry is
> on MV NULLs.
Received on Thu Mar 30 2006 - 03:25:58 CEST