Re: Relational vs network vs hierarchic databases

"erk" <eric.kaun_at_pnc.com> wrote in message news:1100616618.494396.253060_at_f14g2000cwb.googlegroups.com...
> > What about the web? It is based on a graph model, with links from
> one node
> > to another, with each node being a function mapping a URL to the
> value(s)
> > associated with it. I doubt that a relational model would have had
> the type
> > of sweeping impact of the di-graph of functions that is the web.
>
> And completely unstructured text had a sweeping impact prior to the
> web; the web would have been impossible prior to some critical mass of
> both data and technology. The web has no single designer, no single
> business function, and little purpose in common with most databases,
> relational or no. I think this is apples-and-oranges.
>
> - erk

If you view the URL as a primary key, with the links as an array of foreign keys (embedded lsit) and the html or page as a type of document, then you have a structure. I'm not talking about semantic modeling or splitting out the words on the web page. Think of the page content as if it were an mp3 or some other object.

There is a reason that viewing documents as nodes where we can jump from one to another caught on (bit time!) where viewing them in relations did not. I don't know what it is, but I think it has to do with the way our brains more naturally perceive such data. But, of course, like many of my posts, that is not a p-->q argument -- just an open question.

Cheers! --dawn Received on Wed Nov 17 2004 - 00:08:03 CET