Re: Normalize until neat, Automate until Complete

Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net> wrote in message news:<pmvf72-di5.ln1_at_pluto.downsfam.net>...
>
> I would suppose I mean a system of primitive operations that can be used to
> compose data of arbitrary complexity which will follow predictable
> behaviors. I would assume the picking those primitives would take some
> thought, as opposed to simply saying, "hey, we have a speed problem on the
> gonkulator report, better materialize the commonly derived columns!" This
> approach, while being pretty realistic, offers no guidance or advice on how
> to avoid the problem in the future. Worse, the suggestion that these hard
> working folk will "normalize till it hurts" tends to guarantee they will
> get into trouble again and again.
>

I once evaluated a product called Versata, which takes the so called Business Rules approach. In fact, it was actually based on derived, or calculated columns. You could decide whether to materialize them or not. Deriving and/or updating them was handled by the "Business Rules" engine but the applications were using these columns as if they actually existed in the database.

There is an IBM Red Book that describes the product. See e.g.

http://www.redbooks.ibm.com/redbooks/pdfs/sg246510.pdf

Best Regards,
Lauri Pietarinen Received on Fri Nov 26 2004 - 00:56:45 CET