W3C home > Mailing lists > Public > public-xg-eiif@w3.org > October 2008

Re: Fwd: triples/ toward RDFizing the schema

From: Guido Vetere <gvetere@it.ibm.com>
Date: Tue, 7 Oct 2008 16:19:46 +0200
To: paola.dimaio@gmail.com
Cc: public-xg-eiif <public-xg-eiif@w3.org>, public-xg-eiif-request@w3.org
Message-ID: <OFED3C5075.0A1FF2E0-ONC12574DB.0047D7D1-C12574DB.004EB75D@it.ibm.com>
Paola,

>>In the meanwhile, I would suggest getting the XMI 1.4 serialization of 
the current schema, 
        >how do we get that

        Depends on the tool, I'm familiar with eclipse UML2 tools and 
Rational SA, maybe I can help you?

>I use the ontology to support a model of the world that would support the 
required functinality

According to the mainstream  (Gruber), an ontology is the specification of 
a conceptualization. I would add that the conceptualization should aim at 
capturing ontological commitments, i.e. statements on what exists in a 
partition of the World called 'domain of interest'. However, most of 
ontologies that you can find around are just ad hoc models specified with 
highly formal languages like OWL. Formality does not guarantee the quality 
of the conceptualization, though ..

>I am not sure how our ontology should model/represent/express  the 
information flow

If the ontology is intended for information exchange -- this is my 
understanding -- then maybe we can drop the information flow part, unless 
systems are not supposed to talk about their talking, which is logically 
cumbersome however .. :-)

Cordiali Saluti, Best Regards,

Guido Vetere
Manager & Research Coordinator, IBM Center for Advanced Studies Rome
-----------------------
IBM Italia S.p.A.
via Sciangai 53, 00144 Rome, 
Italy
-----------------------
mail:     gvetere@it.ibm.com
phone: +39 06 59662137
mobile: +39 335 7454658





paola.dimaio@gmail.com 
07/10/2008 13.49

To
Guido Vetere/Italy/IBM@IBMIT
cc
public-xg-eiif <public-xg-eiif@w3.org>, public-xg-eiif-request@w3.org
Subject
Re: Fwd: triples/ toward RDFizing the schema






Thanks Guido

I presume what you suggest below is going to be pretty straighforward to 
do
, and in fact any ontology editor will probably do, considering I am 
supposed to be learning the stuff, I ll have a go at some point. ( 
procrastinating)

however, putting a schema into rdf/owl is indeed the most trivial part of 
creating  maintaining using an ontology


In the meanwhile, I would suggest getting the XMI 1.4 serialization of the 
current schema, 

how do we get that
 
import it into Protégé (not necessarily 3.3) and see what happens. 

sounds fun
 
If the system doesn't crash, at the end you'll get a legal OWL file. Then 
mabye you will have to fix something by hand. 

expected

and what happens then? assuming we generate at some point a valid owl file 
of our 
schema, and we make it accessible,  how is it going to be used? how can we 
experiment with it?

I am personally still keen on html/xml because that's what our 
content/data exchange systems (standard cms) can work with at the moment, 
so for me the ontology is a way of making sure that the underlying 
representation schema is sound m basically I use the ontology to support a 
 model of the world that would support the required functinality

But I guess as we are at it, we may as well go all the way and put 
something out 
that we can play with til we know better 
 


As for the message exchange protocol, I'm not sure that 'ontologizing' it 
is really needed -- woudn't be part of the middleware? 

I am not sure how our ontology should model/represent/express  the 
information flow, lets think about it a bit, but whatever needs to be 'not 
casual'  I think, as at the moment is currently this information flow that 
fails (systems do not support the right information flow) and possibly one 
of the things that external representation can fix. I dont think we need 
to ontologise it, but I think we need to make sure that our schema makes 
the data flow as efficient and as transparent as possible, and the 
ontology can help us verify that
(I think)

thanks a lot G

pdm

 


Cordiali Saluti, Best Regards,

Guido Vetere
Manager & Research Coordinator, IBM Center for Advanced Studies Rome
-----------------------
IBM Italia S.p.A.
via Sciangai 53, 00144 Rome, 
Italy
-----------------------
mail:     gvetere@it.ibm.com
phone: +39 06 59662137
mobile: +39 335 7454658




paola.dimaio@gmail.com 
Sent by: public-xg-eiif-request@w3.org 
07/10/2008 04.04 


To
Guido Vetere/Italy/IBM@IBMIT 
cc
public-xg-eiif <public-xg-eiif@w3.org> 
Subject
Re: Fwd: triples/ toward RDFizing the schema








Guido and all
i am having  two additional thoughts that I would like to know what you 
think of

1) if our UML is not optimized, and the relationships are not streamlined, 
as in the case of our UML
I doubt that an ontology that would be derived from it with an automated 
process would be correct

so - correct me if I am wrong  - in order to feed the uml diagram to 
protege and obtain the desired output
(that is a model that can work and not break) we need to do a bit more 
work on those relations

however, if we work out the triples, we can use them to help us 
improve/rationalize our UML

2) an important thing that our ontology does not yet model/address is the 
message exchange protocol
I assume all these relationships correspond to equivalent data/information 
flows, is that correct?
if so, i wonder what protocols should suppor them (EDXl comes to mind, 
being discussed in parallel) and when would we have to think about it

best

PDM



On Mon, Oct 6, 2008 at 4:48 PM, <paola.dimaio@gmail.com> wrote: 
Dear Guido
thanks for input
I am familiar with Protege, in fact I have heard of Protege Light or 
something, and that the recent versions are easier to use

I have been looking forward (and dreading at the same time) the day when I 
would have to learn
how to use it (I am a bit averse to doing too practical things)

However, that day is coming near as I will be attending the summer school 
at ASWC precisely with the intent
of getting down to that, as it obviously something that I -we_ really need 
to work with.  I heard also Jena is good

in fact, i think Protege allows for collaborative ontology editing (can 
you confirm?)
and this is something that we should be working on together (we hope that 
would include you as you sem to have a sense of what we are trying to do 
here)

So, let me ask,

1) what is the best way to work collaboratively on an ontology using 
Protege (or other tool), do we set up and run it on a server that everyone 
can access, or do we each download an instance on our desktops and let it 
synchronize when we have updates?

2) considering this is a collaborative ontology building exercise 
(multiple stakeholders) , is there any other tool/environment that would 
best support our task

3) Mandana, are you up for working together on this, anyone else has 
skills or would like to acquire such skills 


thanks

pdm 





On Mon, Oct 6, 2008 at 9:41 AM, Guido Vetere <gvetere@it.ibm.com> wrote: 

To start, have a look to Protégé http://protege.stanford.edu/  It is not 
an industrial tool but it's quite stable, is easy to learn and supported 
by a vast community. 
There's a plenty of plugins to extend Protégé basic functionalities,e.g. 
to import UML 1.4 Diagrams through XMI. 
In fact, OWL shares a number of basic modelling principles with OO 
languages: classes, properties, inclusions, etc. Then, depending on the 
expressiveness you need, you have other formal notions such as 
restrictions, disjointness, and so on. A reference on this matter is the 
Description Logic Handbook , where you can go in depth with the theory 
behind OWL if needed. At the end you'll get RDF triples based on RDF 
Schema + OWL Schema, i.e. you'll be using standard (formal) properties 
with a clear semantics, all blessed by W3C! :-) 

Cordiali Saluti, Best Regards,

Guido Vetere
Manager & Research Coordinator, IBM Center for Advanced Studies Rome
-----------------------
IBM Italia S.p.A.
via Sciangai 53, 00144 Rome, 
Italy
-----------------------
mail:     gvetere@it.ibm.com
phone: +39 06 59662137
mobile: +39 335 7454658



paola.dimaio@gmail.com 
Sent by: public-xg-eiif-request@w3.org 
06/10/2008 17.35 


To
Guido Vetere/Italy/IBM@IBMIT 
cc
public-xg-eiif <public-xg-eiif@w3.org>, public-xg-eiif-request@w3.org 
Subject
Re: Fwd: triples/ toward RDFizing the schema










Yes, Guido


sure!
Wouldn't we have to work out the triples anyway? Please outline your 
suggested method
thanks!
cheers
PDM

On Mon, Oct 6, 2008 at 8:01 AM, Guido Vetere <gvetere@it.ibm.com> wrote: 

Hi Paola, 
maybe is a silly question, but since we are developing an ontology and we 
like RDF triples, why don't we simply use OWL? We would get DL formal 
semantics and a plenty of OS tools for editing (e.g. Protégé) and 
reasoning (e.g. Pellet). 

Cordiali Saluti, Best Regards,

Guido Vetere
Manager & Research Coordinator, IBM Center for Advanced Studies Rome
-----------------------
IBM Italia S.p.A.
via Sciangai 53, 00144 Rome, 
Italy
-----------------------
mail:     gvetere@it.ibm.com
phone: +39 06 59662137
mobile: +39 335 7454658


paola.dimaio@gmail.com 
Sent by: public-xg-eiif-request@w3.org 
05/10/2008 04.36 


To
public-xg-eiif <public-xg-eiif@w3.org> 
cc

Subject
Fwd: triples/ toward RDFizing the schema













Craig, thanks for reply
I find the comments below educational (learning something)
so I am forwarding them to the list to see if someone has something to add

yes, CAPS are ugly, only here used to distinguish S/O from p



cheers, PDM

and no, I dont have a cat !





---------- Forwarded message ----------
From: C H <craighubleyca@yahoo.com>
Date: Sat, Oct 4, 2008 at 9:59 AM
Subject: Re: triples/ toward RDFizing the schema
To: paola.dimaio@gmail.com


Feel free to forward this if a discussion ensues.  No need to bug the
list with it otherwise.

> I am startedt to think of the schema being worked out by
> Mandana as triples

Wise.  Astonishingly good tools exist for manipulating RDF triples.

> can someone correct the assertion?
>
> SUBJECT predicate OBJECT assumption:
>
> (whereby SUBJECT and OBJECT correspond to the entities in
> the schema, and the predicates to the relationships)
> would this be right?

Yes.  Another word for predicate is "relation" as in
entity-relationship diagram.  Generally the word "relation" is
reserved for the very strict style of table used in relational DBs and
the word "relationship" for ERDs which are much much looser.
Predicates are somewhere in between in the scale of strictness - a
wide range in between from pure logical predicate to vague assertions
piled up in something like semantic mediawiki (a tag scheme that
embeds RDF data into mediawiki pages, extraordinarily useful)

> question (do we have to model all the triples for the schema to work?)

No, but any kind of automated processing will stop dead if you don't
reduce all the relations to three-folded SPO
(subject/predicate/object) before you ask the robot lawyers to take
over.  They may do very strange things like sue your cat if you have
failed to reduce all the constraints to something they understand.
Try not to give them their own expense account, either - robot lawyers
can run up quite a bar bill at the gas bar.

By robot lawyers I mean RDF reasoners and so on, of course.  What else?

> AFFECTEDPERSON needs RESOURCE

Suggests others like "affected_person needs refuge_instructions" -
this ALL-CAPS thing is bad news, it prevents us from writing readable
sentences.  When an [[affected_person needs refuge instructions]] it
would be best to just be able to write it like that because then
humans and machines can both read it with no translation (assuming _
equates to space when rendered).

> ORGANISATION has CONTACTPERSON
>
> ORGANISATiON has CAPACITY is RESOURCE (N TUPLE)
>
> RESOURCE has TIME/LOCATION/OTHER ATTRIBUTE

While you're using them right here, be careful with preposition 
predicates.
An "is" and "has" must be used very specifically, usually by "is" we
mean "is-a-kind-of" and by "has" we mean "has-characteristic" or
"has-component" or "has-resource" (different things, a characteristic
is an inseparable attribute, a component is required for it to work
properly and a resource is something it can share or give away without
failing).

Consider also the time relationships required to deal with a temporal
database.  Korzybski said "is" and the verb "to be" were questionable
at best and could mean too many things, crossing the actual
operational time bindings we use in practice.  In real reality, we are
*remembering* or *explaining* the past which is different from
*sensing* or *comparing* the present state to other things present,
both of which are different from *envisioning* or *predicting* the
future.  The use of "is" and "are" in that sentence is the most basic
and if you don't respect that distinction you get into trouble - for
instance, confusing historical data with some future projection in
order to get some entirely bogus present "trend line".

(where economics goes wrong...)

> does this make sense to anyone on this list, or am I
> enterering another planet? etc etc

Makes perfect sense to me.  But I may have to ask a robot lawyer.  I
hope you don't have a cat.

> Paola Di Maio
> School of IT
> www.mfu.ac.th
> *********************************************






-- 
Paola Di Maio
School of IT
www.mfu.ac.th
*********************************************




IBM Italia S.p.A.
Sede Legale: Circonvallazione Idroscalo - 20090 Segrate (MI) 
Cap. Soc. euro 361.550.000
C. F. e Reg. Imprese MI 01442240030 - Partita IVA 10914660153
Società con Azionista Unico
Società soggetta all'attività di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
above) 



-- 
Paola Di Maio 
School of IT
www.mfu.ac.th
********************************************* 


IBM Italia S.p.A.
Sede Legale: Circonvallazione Idroscalo - 20090 Segrate (MI) 
Cap. Soc. euro 361.550.000
C. F. e Reg. Imprese MI 01442240030 - Partita IVA 10914660153
Società con Azionista Unico
Società soggetta all'attività di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
above)

IBM Italia S.p.A.
Sede Legale: Circonvallazione Idroscalo - 20090 Segrate (MI) 
Cap. Soc. euro 361.550.000
C. F. e Reg. Imprese MI 01442240030 - Partita IVA 10914660153
Società con Azionista Unico
Società soggetta all'attività di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
above) 



-- 
Paola Di Maio 
School of IT
www.mfu.ac.th
********************************************* 



-- 
Paola Di Maio 
School of IT
www.mfu.ac.th
********************************************* 


IBM Italia S.p.A.
Sede Legale: Circonvallazione Idroscalo - 20090 Segrate (MI) 
Cap. Soc. euro 361.550.000
C. F. e Reg. Imprese MI 01442240030 - Partita IVA 10914660153
Società con Azionista Unico
Società soggetta all'attività di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
above)



-- 
Paola Di Maio 
School of IT
www.mfu.ac.th
*********************************************


IBM Italia S.p.A.
Sede Legale: Circonvallazione Idroscalo - 20090 Segrate (MI) 
Cap. Soc. euro 361.550.000
C. F. e Reg. Imprese MI 01442240030 - Partita IVA 10914660153
Società con Azionista Unico
Società soggetta all?attività di direzione e coordinamento di 
International Business Machines Corporation

(Salvo che sia diversamente indicato sopra / Unless stated otherwise 
above)
Received on Tuesday, 7 October 2008 14:21:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 7 October 2008 14:21:19 GMT