Re: A Normalization Question

> And further (this is the point which you stil don't
> get), the string 'brown' and the string 'brown' are ALSO two
> distinct and different things which are not redundant with
> respect to each other if they represent two distinct and
> different logical propositions.

Person Color Street
brown brown brown

We are getting closer. In the above tuple, the person, color and street in the tuple are distinct things. The string 'brown is also a distinct thing. The name of the first three things is the fourth thing, string 'brown'. The string 'brown' is redundant, not the person, the color or the street. RM makes this difficult to so. It is more apparent in TM. See www.xdb2.com/Example/ThingsNamedBrown.asp where one string 'brown' names many things without redundancy.

> > Updating the string 'brown' should have absolutely no
> > impact on the string 'green'. Updating 'brown' to 'nworb' which does
> > not affect 'green' is not an update anomaly.
>
> It is when you update every 'brown' to 'nworb' and in the
> process, also change all the last names which happened to
> be 'brown' to 'nworb' when you only meant to change the
> instances of the paint color.

If one only means to change the name of the paint color, one should not changing the string 'brown' as it names many things. In XDb2, if one changes what appears to be the name of the color, the code unrelates the color from 'brown', searches for 'nworb', if not found, creates 'nworb' and then re-relates the color to 'nworb'. The end result is there is only one string 'brown' and one string 'nworb' each of which can name 0 to many things. Received on Sun Jul 18 2004 - 04:45:17 CEST