Re: Attribute name prefixes, domains, joins, ISO 11179

mAsterdam wrote:

> Maybe you could reconsider not being interested in naming domains.
> Suppose you'ld have a 'PaintCode' domain instead of the ambiguous 'Code'
> (and have a paintcode authority - you /do/ need one).
>
> Q: What kind of PaintCode is it?
> A: The PaintCode for the paint used on the Widget_TopHalf.

Yes, this certainly solves the problem, because we now have two places to "define" what class of thing the attribute belongs to: The column name, and the domain name. But I'm trying to reconcile all of this with 11179, the objective of which seems to be to capture the semantics of a data element in it's name without having to refer to a second source of information (in this case, name plus type). Received on Wed Nov 30 2005 - 03:31:06 CET