Thanks, Duane. Some comments below. (01)
Duane Nickull wrote: (02)
> Lee - Comments inline. Please post to the list for me.
>
> Leo Obrst wrote:
> > Unfortunately, an ontology will not only be a taxonomy or set of taxonomies.
> > So the question I think still stands: can an ebXML Registry handle a graph
> > structure, i.e., a network? On the one hand, an ontology can be considered a
> > collection of assertions (axioms) and be represented as a list. But
>implicitly
> > or explicitly, it is a graph with multiple inheritance as opposed to a
> > taxonomy with single inheritance, which is a tree.
> >>>>>>
> The ebXML registry is a mechanism that can offer publishing and
> discovery of artifacts. Therefore, by itself, it does not convey
> anything than user-defined assertions and declarations. It is the
> Registry Objects themselves that convey the assertions/declarations.
> The answer therefore, is yes. (03)
Ok, reasonable. (04)
> (05)
>
>
> We (the architects) decided that cleanly separating the mechanical
> functionality from the artifacts themselves was the best model to follow.
> (06)
Again, very reasonable. (07)
>
> A Regisstry Object could assert it was derived from multiple
> inheritance. This can also be reflected in the classification schemata
> and/or association mechanism. (08)
Ok. I guess this corresponds to ISO-IEC-11179's classification standard. Of
course,
what is a schema, what is a primary classification relation, etc. (we have
talked
about these issues with the 11179 folks)? (09)
>
> >
> > One issue of course is that if you define a property for a class, will an
> > ebXML registry enable inheritance of that property down the subclass graph?
> >>>>>
> If the information has been entered in the classification scheme
> instance or the associations, then yes. (010)
I wonder how this would work under ebXML, since it would really require some
knowledge and interpretation of the semantics of the taxonomic relation. And I
thought that the ebXML registry was agnostic as to that. Of course, the REGISTER
could expand/ "serialize"/enumerate or "extensionalize" ala a relational
database
all the possibilities him/herself (or equivalently, axiomatize in a list, as I
mentioned before), i.e., something like: (011)
class(a)
property(x)
property(y)
has_property(a, x)
has_property(a, y)
class(b)
property(z)
has_property(b, z)
subclass(b, a)
*has_property(b, x)
*has_property(b, y),
class(c)
property(w)
has_property(b, w)
subclass(c, b)
*has_property(c, x)
*has_property(c, y),
*has_property(c, z),
has_property(c, w) (012)
where the * properties are explicit enumerations of properties inherited from an
ancestor, but could be derived via an interpreter that knew the semantics of
transitivity defined on the subclass relation. This issue of course will come
up in
spades when "rules" or implications are entered. In fact, the above has an
equivalent axiomization using "rules". (013)
>
>
> > An ontology is both the vocabulary plus the meaning of that vocabulary in
> > machine-interpretable form. Really therefore, it is a logical/conceptual
>model
> > on steroids.
> >
> > Now, if the ontology or ontologies are created in another tool (which can
> > handle the above) and then entered into an ebXML registry, then what
>services
> > can we expect from the registry?
> >>>>>>>>>>>>>>>>.
> The registry does not care what the content is. Its' function is scoped
> to accepting artifacts for publishing via the RegistryObjectManager()
> interface and facilitating subsequent discovery of them via the
> RegistryObejctQueryManager() interface.
>
> How the ebXML Architecture handles this is via a set of specially
> designed artifacts. Core Components are artifacts that represent real
> world concepts. These data elements may "bless" refinements based on
> certain contextual circumstances, in which case they become Business
> Information Entities(BIE's). BIE's may be semantically equivalent to
> other data elements defined in other taxonomies. This equivalency
> assertion can be declared within the BIE data model itself (also in a
> rudimentry way via the Registry). (014)
> (015)
>
>
> The registry can also facilitate a discovery from the highest level of a
> company (the trading partner profile), down to more granular levels (
> business processes supported, documents required for each process,
> elements used on each document etc.)
>
> Additionally, registries may contain processing artifacts to allow
> application processing of other artifacts.
>
> For more information, I would recomment reading the ebXML v1.04
> Technical Architecture (http://www.ebxml.org).
>
> I hope this helps. (016)
Yes, I will look at this for further info. (017)
Thanks,
Leo (018)
>
>
> Duane Nickull
>
> > Thanks,
> > Leo
> >
> > marion.royal@xxxxxxx wrote:
> >
> >
> >>I forwarded this thread to Duane Nickull, of XML Global because I consider
> >>him to be an expert on ebXML Reg/Rep. Here is his response and I am
> >>including him as a courtesy copy should anyone wish to reply/comment.
> >>
> >>Duane Nickull wrote:
> >>
> >>Yes please:
> >>
> >>The ebXML Registry could work as a Terminology servver since it is
> >>largely agnostic to the Registry Objects themselves. What would have to
> >>be studied is the classification trees, artifact structure (+ syntax)
> >>and specialized associations needed to facilitate setting an ebXML
> >>registry up as a terminology server.
> >>
> >>Some of this work has been started based on providing a semantic
> >>equivalency function between elements of disparate taxonomies by
> >>relating them to each other within certain contexts. This work involves
> >>taking the UN/CEFACT Core Components methodologies and deriving an XML
> >>syntax representation of both CC's and BIE's. Those artifacts are then
> >>referenced via an ebXML and a uni or bilateral reference is made between
> >>classes of equivalent objects.
> >>
> >>I support an open source project. It may be nice to tie in some of TB
> >>Lee's work on semantic Web.
> >>
> >>Duane
> >>
> >>marion.royal@xxxxxxx wrote:
> >>
> >>>Duane
> >>>Would be happy to tie you into this thread if you would like to respond.
> >>>
> >>>--------------------------
> >>>Marion A. Royal
> >>>U.S. General Services Administration
> >>>202.208.4643 (Office)
> >>>202.302.4634 (Mobile)
> >>>
> >>>Sent from PDA - Please excuse fat thumbs.
> >>>
> >>>------------------------------------------------------------------------
> >>>
> >>> ----- Original Message -----
> >>> From: ontolog-forum-bounces
> >>> Sent: 03/04/2003 12:06 PM
> >>> To: ontolog-forum@xxxxxxxxxxxxxxxx
> >>> Subject: Re: [ontolog-forum] Proposal for UBL Ontology Project
> >>>
> >>>
> >>>In a message dated 3/3/2003 10:58:11 AM US Mountain Standard Time,
> >>>farrukh.najmi@xxxxxxx writes:
> >>>
> >>>
> >>>>A good way to do a virtual project is to do an open source project.
> >>>
> >>>
> >>>Fully agree this is the way to go.
> >>>
> >>>
> >>>>I would like to propose that the proposed UBL ontologies be managed
> >>>>using ebXML Registry as an Ontology Server. There are many interesting
> >>>>features that an ebXML Registry has to offer as an ontology server. A
> >>>>partial list includes:
> >>>
> >>>
> >>>
> >>>This is interesting as I have not thought of the ebXML registry as
> >>>an Ontology server. For example, I do not believe the RIM supports
> >>>the formal notion of 'subclassOf" which would be critical. While I
> >>>believe we could use a custom association with this label, that is
> >>>weaker than the notion of subclass being built into the RIM. For
> >>
> >>example,
> >>
> >>>a formal notion of subclass would allow the child information object to
> >>>automatically inherit the attributes of the parent. Please correct me
> >>>if I am misunderstanding the RIM or its implications.
> >>>
> >>>Additionally, I would recommend the Ontology classes be associated
> >>>with a terminology registry for each concept (in essence equating a
> >>>class with a concept). Following step 3, in the protege Ontology 101
> >>>document, we need to enumerate important terms in the Ontology.
> >>>I am proposing a step beyond enumeration to formal definition with
> >>>concept, terms and referents. Is the ebXML registry suitable for a
> >>>terminology
> >>>registry? Or do people know of others?
> >>>
> >>>- Mike
> >>>-------------------------------
> >>>Michael C. Daconta
> >>>Chief Scientist, Advanced Programs Group
> >>>McDonald Bradley, Inc.
> >>>www.daconta.net
> >>
> >>--
> >>VP Strategic Relations,
> >>Technologies Evangelist
> >>XML Global Technologies
> >>****************************
> >>ebXML software downloads - http://www.xmlglobal.com/prod/
> >>
> >>_________________________________________________________________
> >>To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
> >>Msg Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> >
> >
> > --
> > _____________________________________________
> > Dr. Leo Obrst The MITRE Corporation
> > mailto:lobrst@xxxxxxxxx Intelligent Information Management/Exploitation
> > Voice: 703-883-6770 7515 Colshire Drive, M/S H305
> > Fax: 703-883-1379 McLean, VA 22102-7508, USA
> >
> >
> >
> >
>
> --
> VP Strategic Relations,
> Technologies Evangelist
> XML Global Technologies
> ****************************
> ebXML software downloads - http://www.xmlglobal.com/prod/ (019)
--
_____________________________________________
Dr. Leo Obrst The MITRE Corporation
mailto:lobrst@xxxxxxxxx Intelligent Information Management/Exploitation
Voice: 703-883-6770 7515 Colshire Drive, M/S H305
Fax: 703-883-1379 McLean, VA 22102-7508, USA (020)
_________________________________________________________________
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
Msg Archives: http://ontolog.cim3.net/forum/ontolog-forum/ (021)
|