RE: ESIS, Groves and XML

Subject: RE: ESIS, Groves and XML
From: "Frank A. Christoph" <christo@xxxxxxxxxxxxxxxxxx>
Date: Wed, 21 Jul 1999 15:38:49 +0900
Daniel Mahler wrote:
> "Frank A. Christoph" wrote:
> >
> > I think this is essentially correct. I don't know about the OO
> methodology
> > stuff.
> >
> > Daniel Mahler wrote:
> > > 1 A grove is a lableled directed graph.
> > >   There a specified subgraph, which is a tree.
> > >   This is the parse tree of the document.
> > >   Remaining arcs are put in by subsequent processing.
> >
> > Actually, it is a multigraph, not just a graph.
>
> The distinction being parallel arcs?

Yes, there is a set of arcs between nodes, and each arc is assigned a label.

> > I don't think it is accurate to relate the notion of parse tree
> with trees
> > in the grove.
>
> Can you explain why?

For one thing, the children property of an element node is its content,
which does not include the attributes. But you would expect a parser to also
return the attributes of an element. As it happens, the node class for
attribute-assignments has a children property which points to the value of
the attribute, so it turns out that attribute values are in the subgraph,
but not attribute names. Also, you can see by this example that the subgraph
in question is not connected; or, in other words, there are many trees in a
grove. (However, I don't think this is the origin of the term "grove." :)

Hm, I suppose you actually could think of this as sort of a parser result. I
guess that's up to you.

BTW, there is one important thing I forgot to mention which is wrong about
your model: a labeled arc's target can be a sequence of nodes, and not just
a single node. For example, the content of an element is  usually a list.

--FC


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist


Current Thread