Re: Just one more anecdote

"Kenneth Downs" <knode.wants.this_at_see.sigblock> wrote in message news:bbq2s2-3ah.ln1_at_pluto.downsfam.net...

> Methinks that the RM is close enough to perfection (I did not say it was
> perfect, nor even close to perfect, just close _enough_) so that analysis
> and design are really the same thing. The process of analysis is the
> process of attempting to cast record-keeping needs in terms of normalized
> tables.

I disagree. Analysis describes the problem. Design describes the solution.

When design is used in place of analysis, it results in "thinking inside of the box", in situations where the box itself wasn't there, until somebody decided that the box was a suitable metaphor for the problem as stated.

In the original puzzle about thinking outside of the box, there are nine dots to be connected by four contiguous straight line segments. There is no box in the puzzle. The eye of the beholder sees a box, where there is no box, due to the layout of the nine dots. You can't solve the puzzle, unless you think outside of the box that isn't really there.

>
> ERD has never appealed to me because it seemed to be trying to make
> something easy that was in fact already easy. Because it was trying to
> make something more easy, it had to introduce elements that masked
reality,
> such as a M:M relationship that masks a cross-reference table. What's the
> point? The x-ref itself is sometimes useful for direct querying, so why
> not show it?
>
I'll admit that I'm overly taken with ERD. My initial admiration for it came from a case where someone had
taken an existing CODASYL database that captured a large part of the enterprise data and abstracted out and ERD from it.

This ERD, or a subset of it, allowed us to design a VAX Rdb/VMS database in about three days that would have required about three weeks under other circumstances. We were well on our way to prototyping the solution when management, in its wisdom, changed the target database from Rdb/VMS to Oracle. We survived that without a scratch, but I don't attribute that to good analysis. That was due to good design. That's another discussion.

Converting an ERD to a schema of relations is so straight forward that it may seem trivial. I don't think it is trivial, even though I think it's very easy. But that's another discussion. Received on Wed Aug 03 2005 - 15:19:26 CEST