Re: agile programming and agile db design

DA Morgan wrote:
> heylow wrote:
> > I am in a place where people peddle agile approach. Four months have
> > passed by, with two teams, each comprising 10 developers, working from
> > two states. No refactoring has done at db level; nor is it going to
> > happen. The shop hired a j2ee caching specialist, and he advised
> > pagination and on-the-fly sorting at db-level, and he parrots
> > scalability. Another snake oil product!
> >
> > For a query, we have joined 10 tables to get the required for a
> > particular screen. Many other queries need at least 7 tables. When
> > people designed tables in previous iterations, they had no clue about
> > the requirements for the future requirements.
> >
> >
> > Is this what agile programming and db design is about?
>
> This has nothing to do with Agile. This is people using the good
> reputation of Agile to cover their ignorance. Agile is about project
> management and planning. Not about making bad decisions.
>
> What hyou describe is the technology equivalent of draping oneself in
> the flag and calling upon others to not disagree because to do so would
> be unpatriotic.
> --
> Daniel A. Morgan
> University of Washington
>

heylow,

How is scalability a snake oil product? Designing so that the application will scale is extrememly important. Is the j2ee specialist emphasing the use of prepared SQL with bind variables and not allowing building SQL via concatenation of strings to values? If so, he got that right. The proper use of bind variables are not only important for scalability but also to prevent SQL injection.

As far as the number of tables in a join goes the fact that your one expample needs 10 tables really does not mean much. One of the very first joins written in our shop when we migrated to Oralce was a 10 table join. However, the great majority of queries in our system use no more than three tables. A fairly high number of queries in fact reference only one table.

You might be correct in your opinion; however, there is no way for the forum readers to know if you are correct or just gripping because you are being forced to do the work differently that you want to do it. If you really want a second opinion pick specific points and post enough detail for someone to make a recommendation.

HTH -- Mark D Powell -- Received on Tue Jan 23 2007 - 17:01:21 CST