Re: Nearest Common Ancestor Report (XDb1's $1000 Challenge)

On 1 Jun 2004 09:45:55 -0700, Neo wrote:

>> > XDb1 can accept two things whose attributes do not allow them to be
>> > distinguished (ie two persons named john). Why? Because such
>> > situations can come up in the real world.
>>
>> Yes. And in the real world, extra attributes get added right away to
>> create a way to distinguish them. My wife started at a new job in March.
>> She is called José. One of the women she supervises is also called José.
>> Everybody at my wife's office now calls my wife "José D"
>> and the other lady "José T".
>
>True, but define the time period "right away".

Hi Neo,

"right away" = "as soon as the need to distinguish arises"

> In the example given
>above, "right away" could be from a few hours to a week and in other
>cases it could be infinite. During that time an RM implementation may
>not accept the second person due to constraints (which can also be
>added to XDb1 by user code), while XDb1 could. During that time, XDb1
>could make decisions based two persons with same name

Decisions that have a 50% chance of being incorrect.

>, where as RM's
>would be based on only one person. Both methods have pluses and
>minuses. XDb1's approach maybe more applicable for AI-type situations.
>
>> Why? Because in the real world, there's a NEED to distinguish things from
>> one another, and if they happen to have the same name something else is
>> used to make that possible.
>
>True, sometimes in order to meet a need, an attribute, constraints and
>code have to be added. But having a NEED doesn't guarantee it is
>always satisfied.

In cases like this, it has to be satisfied before facts about the things can be stored and retrieved in a meaningful way.

Best, Hugo

-- 

(Remove _NO_ and _SPAM_ to get my e-mail address)