Re: A Logical Model for Lists as Relations

mAsterdam wrote:
> Let's have a name for this way of describing a list as a relation.
> Is "numbered items" ok?
>
> > The other possibility is to represent list with refereences like this:
> >
> > table EmpList (
> > nodeId integer,
> > nextNodeId integer,
> > ... -- list element content
> > )
>
> Would "successive items" be a good one for this way?
>
> So, we have several ways to describing a list as a relation, at
> least "numbered items" and "successive items".

The difference between the two essentially reduces to the way how we represent the order. The first method is intensional. It encodes the elements of a list with integer numbers which are (totaly) ordered. The second method is extensional. The order is represented explicitly by binary relation.

> To me it isn't clear what discrete piece of information content
> is carried by an individual "number" or "successor" attribute.

This question is meaningless without understanding what "information content" is.

To reply to another thread where Bob raised the utility question, it is undoubtful that sequences appear all over the math. Arguably, sequence is even more frequently occuring concept than set. There is even a dedicated website dedicated to integer sequences: http://www.research.att.com/~njas/sequences/ Received on Thu May 11 2006 - 22:12:38 CEST