Re: Modeling Data for XML instead of SQL-DBMS

David Cressey wrote:
> "dawn" <dawnwolthuis_at_gmail.com> wrote in message
> news:1161893884.428842.156110_at_m7g2000cwm.googlegroups.com...
> > OK, I'll review all the feedback and come back with a revised question
> > (once I have proper definitions for the logical data model and know
> > precisely for which model we need to know whether persistence will be
> > handled with UniData or DB2, for example).
>
> I'm going to suggest that there are at least two conflicting definitions of
> "logical data model" in use in the newsgroup, regardless of what the
> glossary says.

I think it is mum on this matter.

> One of the definitions might be called "preliminary design".
> The other might be called "programmer visible implementation."

Yes, yes, very good summary. I was using the former and after reading, I switched to the latter. However, that is being contested here.

> Note that,
> in the case of design choices that are invisible to the programmer, it
> doesn't matter why they are invisible.
>
>
> > Only because I need to rephrase the question and am apparently using
> > the term Logical Data Model incorrectly, yet I'm not certain whether if
> > you and I are both given the same conceptual data model and you are
> > implementing it in Oracle and I in UniData, whether we might have the
> > same logical data model, although different implementation data models,
> > or whether our logical data models would differ. Mine would include
> > multi-valued attributes, for example. Thanks for any clarification.
> > --dawn
> >
>
> Here's the way I learned the terms, back in 1984 (no pun intended).
>
> conceptual data model (CDM): Driven by requirements. results from data
> anlysis. |data model independent. Example: er modeling.

agreed.

> logical data model (LDM): design model, cannot be changed without (in some
> situations) requiring (some) application code rewrite.
> result of preliminary design, driven by conceptual model, plus some
> indication of how the data is to be used. Data model dependent, DBMS
> independent, with a class of DBMSes that are all based on the same data
> model. Exapmple: relational data model.

Yes, that is how I was using the term, as a model that is data model dependent. So, if I am going to implement in SQL Server or DB2, my logical data model will be different than if I am going to implement in UniData or OpenQM. Perhaps the confusion in the terminology is that if one has blinders on and see only one class of DBMS tools, then they think of the logical data model as actually being DBMS-independent, rather than DBMS-independent for a particular class of DBMS's (or for a particular "data model").

> physical data model (PDM): late design model, driven by logical data
> model, but also by data volume, resources, load, and speed priorities.
> DBMS product (and perhaps version) dependent. Example: p.d.m. for
> implementation on Oracle.

I have used physical data model for the model used by a DBMS tool. So the physical data model was "hidden" from me as a developer, until addressing a performance problem, for example.

> create script: this term makes sense only in the context of SQL DDL, but
> the same idea could be applied to other environments. DDL ready to be run
> by a DBMS to create an empty database.
>
> Note that this terminology survives just about intact, when used with Data
> Architect, pert of Power Designer in 1999. Except that DA doesn't discuss
> an LDM separately. You can think of the table desing in the PDM as being
> the LDM.
>
> I'm NOtT claiming that this terminology is standard, or even the best. I am
> saying that I've been using is to communicate with others for over 20 years.
> And I recommend it for that reason.

My terms, learned in the field and not the classroom, align with yours in the first two cases, not with the PDM. I have a ton of books and web sites here and I'm just not sure what folks think would be the most acceptable versions of these definitions. Thanks for passing along your defs. --dawn Received on Thu Oct 26 2006 - 23:17:33 CEST