W3C home > Mailing lists > Public > xml-dist-app@w3.org > September 2002

Re: Issue 302: Graph edges that do not terminate

From: Jacek Kopecky <jacek@systinet.com>
Date: 09 Sep 2002 21:07:42 +0200
To: Marc Hadley <marc.hadley@sun.com>
Cc: Martin Gudgin <mgudgin@microsoft.com>, XMLP Dist App <xml-dist-app@w3.org>
Message-Id: <1031598462.18324.52.camel@krava>

 Marc,
 let me go around it a bit.
 SOAP Encoding rules serialize a given graph starting at a given thing
(be it an edge or a node, this is left unspecified in the spec AFAIK).
This process is a recursive one - serializing a node also means
serializing all its outgoing edges, serializing an edge also means
serializing the target node etc. With this process you never get to
inbound-only edges unless you start from one.
 The problem here is that having edges that don't have a source or
target node can be very confusing, especially when combined with the
fact that it's not the whole graph that is serialized, only the parts
reachable by directional edge traversal. Although implied, I don't think
this fact is explicitly mentioned in the Encoding section.
 So, my preference (not pushed unless yall suddenly decide we should go
this way) would be to stop toying with edges that don't originate or
terminate at a node; and instead define a special node, nil, which is a
singleton and at which an edge can only terminate (representing the
usual programming language null); and a special node, root, which is a
singleton and at which an edge can only originate. Then the process of
serialization could be simply defined to start with an edge (because
that is what becomes an element II) and all would be nice and simple.
The root node would never be serialized, meaning no confusion.
 It's a pity we have so little time. 8-)
 Best regards,

                   Jacek Kopecky

                   Senior Architect, Systinet Corporation
                   http://www.systinet.com/



On Mon, 2002-09-09 at 18:26, Marc Hadley wrote:
> 
> On Monday, Sep 9, 2002, at 11:13 US/Eastern, Jacek Kopecky wrote:
> >
> >  as I see it, mandating xsi:nil would do, but I don't think it's
> > advisable; it's been discussed at lengths before.
> >  I think the Encoding should say how the "root" element (and its name)
> > is formed, but I don't think inbound-only edges are necessary and I
> > think they would be overkill (what if there are two inbound-only edges
> > in a graph?)
> Quite possibly a naive question but: what's wrong with a graph having 
> two inbound-only edges (i.e. two 'roots') ?
> 
> Marc.
Received on Monday, 9 September 2002 15:07:47 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:11 GMT