identical vs sameAs
ja...@openguid.net
I think it might help with adoption to use a property already
understood by tools, namely owl:sameAs.
It would restrict us to using OWL Full for the published links, which
allows the use of classes as individuals. I think the semantics might
be close enough to what I'm aiming for, namely linking real world
concepts.
In terms of using sameAs between classes, the spec states:
"In OWL Full, where a class can be treated as instances of
(meta)classes, we can use the owl:sameAs construct to define class
equality, thus indicating that two concepts have the same intensional
meaning."
It is less clear what happens when the property is used between
classes and individuals, but I'm guessing most tools interpret it as a
linked concept.
Thoughts?
Heiko Stoermer
Heiko Stoermer
University of Trento, Italy
Dept. of Information Science and Engineering (DISI)
http://disi.unitn.it/~stoermer
OKKAM id:
http://www.okkam.org/entity/ok5f23a5ce-a683-4c4d-ae73-b78cdc17aec1
Ryan Shaw
using owl:sameAs for individuals?
ja...@openguid.net
classes and instances. That is dependent on a target application /
ontology's view of the world.
As Heiko pointed out, there has been much discussion on the use of
sameAs in the LOD world. Some argue it's too formal for average
linking. Others say it's the best thing available. No conclusion was
ever reached, so I created a new property that didn't have formal
reasoning semantics. I was hoping to get the opinions of people on
this list...I know Bernard has written about possible predicates in
between owl:sameAs and rdfs:seeAlso.
I hope for adoption of oguid:identical. But at the same time I want
resources to be easily linked with existing tools. The best answer
might be to use identical by default, but provide a supplementary
downloadable file that uses sameAs if someone wants those semantics
for their purposes.
Bernard Vatant
I was off-line for a couple of days, otherwise I would have answered
right away to this.
> I am considering dropping the oguid:identical property all together.
> I think it might help with adoption to use a property already
> understood by tools, namely owl:sameAs.
>
Short answer : PLEASE, DON'T DO THAT ... or count me out :'( .
Long answer : openguid:identical is addressing exactly the issue set by
the lack of some efficient linking mechanism between owl:sameAs (too
strong) and rdfs:seeAlso (too vague).
Using owl:sameAs is interfering with the semantics of the URIs which are
linked, and it's exactly what you don't want to do with open GUID.
Precisely open GUID links are useful because they don't interfer
whatsoever with the semantics of the linked resources.
Paraphrasing Lao Tseu :
/thirty URIs pointing to one GUID/
/thnks to its vacuity/
/the GUID is useful/
So don't fill this useful void with semantics ...that is the main
conceptual point. Now for the technical aspects, read below if you like.
> It would restrict us to using OWL Full for the published links, which
> allows the use of classes as individuals. I think the semantics might
> be close enough to what I'm aiming for, namely linking real world
> concepts.
>
No, it is not. Using owl:sameAs, you declare equivalence of the
semantics of representations, whereas what you are aiming for is
asserting the equivalence of referents, which is beyond the semantics of
representations (maybe I project my dreams on yours, but so far I
understood they were the same. :-)
> In terms of using sameAs between classes, the spec states:
>
> "In OWL Full, where a class can be treated as instances of
> (meta)classes, we can use the owl:sameAs construct to define class
> equality, thus indicating that two concepts have the same intensional
> meaning."
>
> It is less clear what happens when the property is used between
> classes and individuals, but I'm guessing most tools interpret it as a
> linked concept.
>
Most tools will not interpret it at all, because there are no logical
tools dealing with OWL full properly ...
> Thoughts?
>
You got some
Bernard
--
*Bernard Vatant
*Senior Consultant
Vocabulary & Data Engineering
Tel: +33 (0) 971 488 459
Mail: bernard...@mondeca.com <mailto:bernard...@mondeca.com>
----------------------------------------------------
*Mondeca**
*3, cité Nollez 75018 Paris France
Web: www.mondeca.com <http://www.mondeca.com>
Blog: Leçons de Choses <http://mondeca.wordpress.com/>
----------------------------------------------------**
ja...@openguid.net
up :)
Rest assured, your dreams and mine are the same. I got a few hate
mails about reinventing properties and "gobbledygook" URIs that were
causing some mind circles.
Here is my corresponding slashed haiku (valid only if you pronounce it
"gwid"):
/Poor GUID, not pretty/
/You were called gobbledygook/
/Computers don't mind/
Heiko Stoermer
Bernard Vatant
Please forgive if I answer a question addressed to Jason, and to put it
abruptly that so far you've missed the point of Open GUID altogether.
Don't take it personal, it's a point difficult to grasp by most people,
and the more so that they are experts in semantics.
> a question from my side: why create new identifiers for classes anyway?
OpenGUID does not create new identifiers for classes or properties or
instances, it creates identifiers for referents before or beyond the
fact that they have been represented as classes or instances. Let me
take again an example. Let's consider "Organic food". In natural
language where it has emerged, this is neither a class nor an instance,
nor any kind of formal concept. It's a linguistic proxy pointing to some
ill-defined aspect of the world, of our experience, which at some point
the community has found to be salient. IOW, people coined this term to
make it clear organic food neede to be made ditinct from anything else
in the world. This is as void of semantics as possible : there is
something called organic food, which has now a linguistic identity.
Now when I need to build a formal representation of this concept, the
type of representation will depend upon what I want to do with this concept.
If I want to sort specific instances of food products found on my market
into organic and non-organic, I will certainly build a class "o:Food"
and two disjoint subclasses "o:OrganicFood" and "o:NonOrganicFood", with
their respective URIs.
If I want to index documents or publications about organic food, I will
certainly define a skos:Concept "c:OrganicFood" in some Concept Scheme.
A third representation b:OrganicFood would represent an instance of
business domain for people or companies
The semantics of those representations, conveyed by descriptions of
those two URIs, are certainly different and even inconsistent with each
other, for example.
foo:MySweetTomato rdf:type o:OrganicFood
ex:OrganicMonthlyReview rdf:type foaf:Document
ex:OrganicMonthlyReview dc:subject c:OrganicFood
bar:TheOrganicGarden rdf:type bar:Store
bar:TheOrganicGarden bar:hasActivity b:OrganicFood
What does Open GUID brings is a link between the o, c and b
representations of Organic Food.
openguid:identical says exactly "there is some common referent beyond
those representations, although they do not convey the same semantics,
and hence can't be merged by a owl:sameAs."
Why yet another URI, when I have already three different ones for three
aspects of the same referent? Exactly to say that : there is a common
referent, and this URI acts as a hub, or a binder, between its various
representations. It has no semantics whatsoever beyond that, it does not
inherit of any semantics of the resources it links, nor add any extra
semantics to them. Having an URI for the binder is a way to anchor this
binding in the Web infrastructure. The open GUID is not another URI for
the class, it's a URI for the referent *before it's been represented as
a class*.
> Should we not assume that a class in an OWL ontology (if it is
> published on the web) is uniquely identified by its URI?
Indeed it is, as a class. But no representation is exhausting its
referent, as the above example shows (hopefully).
Does that help?
Bernard
>
> Heiko
>
> On Fri, Oct 3, 2008 at 10:53 AM, <ja...@openguid.net
> <mailto:ja...@openguid.net>> wrote:
>
>
> Awesome response! I was hoping my idle threat would get someone fired
> up :)
>
> Rest assured, your dreams and mine are the same. I got a few hate
> mails about reinventing properties and "gobbledygook" URIs that were
> causing some mind circles.
>
> Here is my corresponding slashed haiku (valid only if you pronounce it
> "gwid"):
>
> /Poor GUID, not pretty/
> /You were called gobbledygook/
> /Computers don't mind/
>
> On Oct 3, 2:07 am, Bernard Vatant <bernard.vat...@mondeca.com
> <mailto:bernard.vat...@mondeca.com
> <mailto:bernard.vat...@mondeca.com>>
> > ----------------------------------------------------
> > *Mondeca**
> > *3, cité Nollez 75018 Paris France
> > Web: www.mondeca.com
> <http://www.mondeca.com><http://www.mondeca.com>
> > Blog: Leçons de Choses <http://mondeca.wordpress.com/>
> > ----------------------------------------------------**
>
>
>
>
> --
> Heiko Stoermer
> University of Trento, Italy
> Dept. of Information Science and Engineering (DISI)
> http://disi.unitn.it/~stoermer <http://disi.unitn.it/%7Estoermer>
> OKKAM id:
> http://www.okkam.org/entity/ok5f23a5ce-a683-4c4d-ae73-b78cdc17aec1
>
--
*Bernard Vatant
*Senior Consultant
Vocabulary & Data Engineering
Tel: +33 (0) 971 488 459
Mail: bernard...@mondeca.com <mailto:bernard...@mondeca.com>
ja...@openguid.net
It is becoming clear to me that what is clear in my head is clearly
not being made clear to others. I need to clear this up!
I will start with a blog post, and maybe pull some of it into the spec
or another reference document. Do you mind if I use your organic food
example?
Cheers,
Jason
> *Mondeca**
> *3, cité Nollez 75018 Paris France
> Web: www.mondeca.com<http://www.mondeca.com>
Bernard Vatant
> It is becoming clear to me that what is clear in my head is clearly
> not being made clear to others. I need to clear this up!
>
come to this with a mind "as an empty pot", and not everybody knows how
to do that.
> I will start with a blog post, and maybe pull some of it into the spec
> or another reference document. Do you mind if I use your organic food
> example?
>
feel free to use. There is no copyright on this :-) .
Bernard
--
*Bernard Vatant
*Senior Consultant
Vocabulary & Data Engineering
Tel: +33 (0) 971 488 459
Mail: bernard...@mondeca.com <mailto:bernard...@mondeca.com>
Simon Reinhardt
> > It would restrict us to using OWL Full for the published links, which
> > allows the use of classes as individuals. I think the semantics might
> > be close enough to what I'm aiming for, namely linking real world
> > concepts.
>
> No, it is not. Using owl:sameAs, you declare equivalence of the
> semantics of representations, whereas what you are aiming for is
> asserting the equivalence of referents, which is beyond the semantics of
> representations (maybe I project my dreams on yours, but so far I
> understood they were the same. :-)
representation of a document sent back by a Web server? Certainly not
since not all resources are documents and so not all resources have
such representations. So what exactly is the representation of the
resource denoted by the URI <http://www.w3.org/People/Berners-Lee/
card#i> and how exactly is it different from the "representation" of
the resource denoted by the URI <http://www4.wiwiss.fu-berlin.de/
bookmashup/persons/Tim+Berners-Lee>? What are the problems of using
owl:sameAs here, using the pure RDF model?
Simon