Re: Relations as Repeating Groups & Namespaces

From: Dawn M. Wolthuis <dwolt_at_tincat-group.com>
Date: Thu, 3 Jun 2004 11:03:22 -0500
Message-ID: <c9ni4h$ial$1_at_news.netins.net>


"x" <x-false_at_yahoo.com> wrote in message news:40bef890_at_post.usenet.com...
> **** Post for FREE via your newsreader at post.usenet.com ****
>
>
> "Dawn M. Wolthuis" <dwolt_at_tincat-group.com> wrote in message
> news:c9m55r$k99$1_at_news.netins.net...
<snip>
> Hierarchical namespaces tend to be dependent on the applications.
> Hierarchical namespaces tend to classify *things*.
> We need a method general enough to deal with any kind of facts.
> But I agree that "labeling" is one of the human abilities.

The entire logical model is interwoven with the applications (such as mountain man) with namespaces being no different. We separate out root nodes for namespaces based on different companies or subsidiaries or divisions.

Why do we have different rules for namespaces depending on where they are in the namespace tree? Is there any logic to that? Associate with any relation name the same things you associate with a schema (root namespace) and you have a more elegant structure, it seems (by being consistent). XML permits relations nested to any level. There is a point where it becomes confusing semantically (that's where some products cut off the hierarchy) but it surely is not at the level of relations. Permitting relations within any namespace (including within a relation) seems reasonable.

--dawn Received on Thu Jun 03 2004 - 18:03:22 CEST

Original text of this message