W3C home > Mailing lists > Public > www-rdf-logic@w3.org > August 2007

RE: Beginner question: How to describe a subgraphs structure?

From: Stéphane Monteil <stephane@engineering-studio.com>
Date: Sat, 25 Aug 2007 13:20:28 +0200
To: "'Heiko Stoermer'" <hstoermer@gmail.com>
Cc: <www-rdf-logic@w3.org>
Message-ID: <000501c7e709$f1f02e80$6402a8c0@SMONTEIL>
Hi Heiko,


Thanks for these further references. It always helps actual development to
learn some general principles and design ideas from the community.

I will get back to you soon to provide a url and account to our Beta
platform - A web semantic application for everyone ;)






De : Heiko Stoermer [mailto:hstoermer@gmail.com] 
Envoyé : vendredi 24 août 2007 09:53
À : Stéphane Monteil
Cc : www-rdf-logic@w3.org
Objet : Re: Beginner question: How to describe a subgraphs structure?


Hi Stéphane,

you will probably get disappointed when you start looking into RDF and
contexts, especially if you don't want to do research, but actual
development.. In fact, I did some work on this topic over the last few
years, and you might want to browse some of my articles, which also have
references to related work done by other people: 


On 8/23/07, Stéphane Monteil < stephane@engineering-studio.com
<mailto:stephane@engineering-studio.com> > wrote:

Hi Heiko,

Thanks for your answer. Your illustration is pretty clear and 
straightforward and it looks well adapted to model my problem in a simple

I will have to learn more about RDF containers to understand if there could
be a benefit for me. At this stage, I don't really see why they exist as it 
looks possible to describe some groups of things directly with straight RDF

The mention of reification led me to the notion of contexts that is actually
a requirement for me.
I work on a web application that allows to express some stories or memories 
through multimedia content. The stories are made of resources that can
change dynamically, depending on the user's selections (like the time frame
or the geographical region).



De: Heiko Stoermer [mailto:hstoermer@gmail.com]
Envoyé: jeudi 23 août 2007 16:02
À: Stéphane Monteil
Cc: www-rdf-logic@w3.org
Objet: Re: Beginner question: How to describe a subgraphs structure?


What you might want to experiment with is a combination of blank nodes,
reification and the RDF container constructs (alt, bag, seq). 

>From your example however I think that you don't really need that - it can
be modeled with straight RDF, just as you verbalized it below, with sth

event1 hasPicture pic1, pic1 takenBy person1 
event1 hasPicture pic2, pic2 takenBy person1
event1 hasPicture pic3, pic3 takenBy person2
event1 location loc1
event1 date '2004-12-12'

This will create a little graph that has event1 as the root element, with 
some properties like location and date, and some related pictures, each one
with its photographer.

Does this help? If not, you might want to submit a more detailed example ;)


On 8/23/07, Stéphane Monteil < stephane@engineering-studio.com> wrote:

Hi all,

I am a beginner with semantic web languages and my question is probably
basic. Yet, I have not found a clear answer through the W3C documents I read

so far.

I would like to describe a graph (a set of triples) as a subject (a
resource) by itself, and so on for a set of subgraphs.
To express it differently, in my graphs some of the nodes could contain
(reference) another graph.

1) Is it a consistent approach from a formal representation point of view?
2) Is it possible to represent that directly with RDF or any other
standardized language (OWL, etc.)?

My objective is to be able to consider a composite content of subjects as a
subject by itself. The real world application could be for example to be
able to manage a collection of pictures related to a specific event. Each 
picture is a subject (with several predicate - objects) and the grouping
event is a higher level subject itself that can have semantic properties.

Remark: The RDF Graph Modeling Language (RGML) seems to propose a kind of 
formal solution, but it does not look standardized and I would prefer to
stay within something generic (ie: a graph can be a node).

Thanks in advance for your help!

Stephane Monteil

Web site: http://www.engineering-studio.com

Received on Saturday, 25 August 2007 11:20:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:48 UTC