Re: [xml-dev] How you interpret your business rules has a profoundimpact

[Michael Kay <mike@saxonica.com>]

On 24/03/2011 22:25, Costello, Roger L. wrote:
> The question is important.
No it isn't, it's meaningless.

If date-of-death is earlier than date-of-birth, one of them is wrong but 
you don't know which, and it's foolish to pretend that you do.

The only situation in which I might consider declaring one of them 
invalid in preference to the other is if the data collection for one 
attribute is much more reliable than the data collection for the other.

(A common mistake is to reject newly entered data because it's 
inconsistent with existing stored data. In practice the new data is 
probably more reliable than the old.)

Michael Kay
Saxonica