From: Dawn M. Wolthuis <>
Date: Mon, 31 May 2004 07:23:44 -0500
Message-ID: <c9f853$n1n$>

"x" <> wrote in message news:40bb10bc$
> **** Post for FREE via your newsreader at ****
> "Alfredo Novoa" <> wrote in message
> > On Sat, 29 May 2004 08:42:21 -0400, "Laconic2" <>
> > wrote:
> > Pointers are about HOW. How to transverse access paths.
> No. Pointers are about WHAT. What are the access paths. :-)

I think this is at the hub of a significant difference of opinion. I am unaware of "how" pointers - those are behind the scenes from my perspective. I recall reading programs with an occasional alternation of memory locations (considered bad programming even back then), but I don't see any of that these days.

I agree with x that the specification of a "foreign key pointer" in a graph structure is about what, not how. This is similar to a specified view that includes a join - the join specification is about what, not how. One difference is the difference in the language of set theory compared to graph theory. We talk about traversing a graph, which has the language of a process, but that is how one talks about the "what" of a graph, not how it is implemented.

If a specification of a path in a graph is a "How" then so is the specification of a join in a view. --dawn Received on Mon May 31 2004 - 14:23:44 CEST

Original text of this message