[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: APIs, messaging
- From: Jeff Lowery <firstname.lastname@example.org>
- To: 'Thomas Passin' <email@example.com>, firstname.lastname@example.org
- Date: Wed, 23 May 2001 11:57:11 -0700
> de-normalization of a
> data model to get performance can be plenty subjective!
Oh, I would agree with that one, but I still assert that data design is more
cut-and-dried than class design (and easier to get right). I say this
because I've spent more time and emotional effort in arguing about the
latter than the former. Usually class design gets resolved in favor of the
guy who writes them.
The main reason class design is more complicated is that classes have more
scope, which is the whole point about separating the data model from the
containment hierarchy of objects. First, write (or better yet, generate)
data-manipulation classes against the data model (incorporating the type
system of the data model into the class model), then add all the processing
logic on top of that. It's a layered approach. Gee! How radical!