Re: object algebra

"Neo" <neo55592_at_hotmail.com> wrote in message news:4b45d3ad.0402260913.2e7125bc_at_posting.google.com...
> > Could you show us the parenthesized representation of a three node
> > directed graph, with nodes A, B, and C, and arcs from A->B, B->C,
> > and C->A? If you believe that relations are unnecessary, then you
> > should be able to do it without directly encoding a relation into
> > your expression.
>
> I am not sure I understand the problem correctly.
> I take a blank piece of paper.
> I draw 3 nodes on it.
> I label the nodes A, B and C.
> I draw a line with an arrow pointing from A to B.
> I draw a line with an arrow pointing from B to C.
> I draw a line with an arrow pointing from C to A.
> The result looks something like the figure below:
>
> -->A-->B-->C
> |----------|

That's right on so far.

> Which aspects of the above diagram do you want me to represent? Did
> you want it to form a loop? Are you asking me to represent the above
> diagram with a relational expression using only A, B, C and the
> operators intersection, union, negation and grouping with parentheses?

In an earlier post you claimed that a parenthesized notation was superior to RDM, and gave an example:

   ((((ab)cd)e)f)(gh)

I'm not 100% sure I understand what you mean by that notation but I think I do: it's a pair of trees, with leaf nodes a, b, c, d, etc. Is that right? (Or perhaps its a single tree with two children of the root.)

If this parenthesized notation is truly superior, it should be at least as easy to express the three node graph as it is in RDM. I pick this example because it's something that's pretty easy to represent with relations but that many other formats have a hard time with. For example, XML pretty much falls on its ass with this example, having to resort to adding sufficient facilities as to allow one to express arbitrary relations. This leaves one wondering what all the stuff that went in in the first place was for.

Marshall Received on Fri Feb 27 2004 - 05:10:53 CET