Re: Can Oracle save a tree?

"Niall Litchfield" <niall.litchfield_at_dial.pipex.com> wrote in message news:40900c1b$0$20507
$cc9e4d1f_at_news-text.dial.pipex.com...

> There is of course an alternative if you don't consider traditional ER
> modelling and design appropriate in your case. Don't use an RDBMS at all.
> Use some sort of XML and get the data to describe itself.:(

How can he do that if he can't describe the structure of the data in the first place? XML describes structure where there is one. It doesn't replace or fabricate it.

The problem most of these OO folks have in the first place is they don't have the foggiest what their data model is and what it can be used for. All they bother with is the GUI and the object model and manipulation. When it comes time to make their programs sufficiently general that they can store their data states and re-start, the whole thing crumbles.

It's the very old problem of giving RAD tools to folks who understand squat about data management: the result looks good but is incapable of making the transition from prototype to product. Not suggesting it's the case with Ramon specifically. But a very common problem about 15 years ago and again nowadays.

I had an experienced architect ask me in my last project to remove an entire table from my database design and let another OO application manage its contents, while passing me back a subset of the data in a view.

The fact that this table had parent table child tables in my schema and I simply couldn't apply RI if I turned it into a view went completely unnoticed with him until I specifically asked how he'd get my surrogate FKs back.

"Implementation details", "don't confuse implementation with design", yadda yadda. I often wonder in what sort of lah-lah-land do these folks live in? Haven't they realised yet they MUST primarily handle data? By itself an OO program/design/model is worth exactly squat unless it can handle existing and future data?

-- 
Cheers
Nuno Souto
wizofoz2k_at_yahoo.com.au.nospam