Cameron,
In the EKB open source project (http://portal.modeldriven.org/project/EKB)
we use both of the interfaces you are describing together. You can check
models of various kinds in/out via subversion and these are then published in
RDF using the linked open data conventions. You can check in OWL, RDF, UML,
Any “EMF” model or schema and it comes out in RDF. Other adapters
can be added. You can then use the online interfaces such as SPARQL to access
the information. At this point it is read-only at the RDF interface but that
is just SMOP (simple mater of programming) – other variants of the EKB
have been bi-directional but that is not packaged up yet. These are “structural”
transforms to RDF, semantic integration is another layer.
There are enormous file-based resources
and these can’t be ignored, so we want a way for all of the enterprise
knowledge to be available and connectable with other file, repository and DBMS
based data.
Regards,
Cory Casanave
From: sio-dev-bounces@xxxxxxxxxxxxxxxx
[mailto:sio-dev-bounces@xxxxxxxxxxxxxxxx] On
Behalf Of Cameron Ross
Sent: Wednesday, April 07, 2010
1:00 PM
To: [sio-dev]
discussion
Subject: Re: [sio-dev] Fwd: [ontolog-forum] Sharing and IntegratingOntologies
Hi Ali,
Thanks for the detailed response. I agree with Ron that you
should consider putting this up on the wiki!
The Ontology Design Tool use case takes an ontology repository as
input, along with input from a user, to generate an ontology as an output.
Both Semantic Mapping Tool uses cases take an ontology repository as
input, along with mapping axioms input from a user, and generate an
interpretation of the "Similarity" and "Differences"
between ontologies. Although these are certainly useful use cases, they
don't address my original questions. To clarify, I will restate my
questions:
1) Are there use cases that describe uses of an
ontology repository that go beyond the development and sharing of
ontologies (i.e. something closer to the eventual end user use cases of the
ontologies I guess)?
2) Are there use cases that clearly justify the need for an
"Online Repository"? By "Online Repository", I
mean an information system deployed on a centralized server somewhere (or
cluster) that provides users the ability to CRUD and process repository
information in near-real-time. This is in contrast to what I'll refer to
here as a "Static Repository". Take the Eclipse open source
software project as an example. We may define the Subversion repositories
used to manage the Eclipse software artifacts (e.g. source code) as
a "Static Repository". Users checkout artifacts from the
repository to their own local copy of the repository. Users
then CRUD and process their copy of the repository, synchronizing their
copy with the maser repository on occasion. Both Online and Static
Repositories have merit with the appropriate solution dependent on the intended
uses of the repository.
I'm not advocating either approach, but it would be great to see the
use cases that clearly justify one approach over the other. I get the
sense that we have tacitly decided that an ontology repository must be
an Online Repository. However, if there are no specific use cases
for an Online Repository then I would recommend that a Static Repository be
considered as the effort to deliver a production quality Online Repository is
much, much higher. There are all kinds of tools, systems, portals and
experience that we could leverage for a Static Repository. There are
examples of initiatives that manage some very large collections of engineered
software artifacts. The Eclipse Foundation, Apache Foundation and Linux
come to mind.
3) Does it make sense to consider ontologies, or
their modules, to be engineered software artifacts?
4) Are there significant differences between a collection of CLIF
or OWL files and a collection of Java files that force us to consider an
entirely different approach to constructing a large collection of such
artifacts within a collaborative environment?
On Wed, Apr 7, 2010 at 11:01 AM, Ali Hashemi <ali.hashemi+ontolog@xxxxxxxxxxx>
wrote:
On Tue, Apr 6, 2010 at 6:27 PM, Cameron Ross <cross@xxxxxxxxxxxx>
wrote:
I guess I'm struggling to understand the value of ontology
repositories beyond their ability to facilitate the development and
sharing of ontologies. If there isn't any value beyond this, then
what is the need for an "online" repository. Wouldn't a system
that
allows for the sharing and collaborative development of static
ontology artifacs (e.g. files) suffice? SourceForge comes to mind.
Perhaps this a naive view, or it has been discussed previously. If
so, some links would be much appreciated.
Cameron.
Kojeware Corporation
Dear Cameron,
Thanks for the question. I can provide three general use cases, and one
concrete example.
General
Use Case
1) Ontology Design Tool (ODT)
Motivation: I am what an ontologist / knowledge engineer would call a
"domain specialist" or a "subject matter expert." I have
extensive experience in my field and want to formalize my knowledge. I don't
have the time nor desire to become an expert in logic, but I want to be able to
express my work in a machine readable format that is shareable with others in
my field and might possibly ease interface with those in fields peripherally
related to mine.
Goal: Provide a mechanism for a SME to formalize intuitions.
Actors: User + ODT
Triggers: Need for formal axioms
Pre-requisites: COLORE
Solution: Using the inbuilt advantages afforded with a formal language at least
as expressive as first order logic, we can communicate with the SME using
examples (tarski-models) only, and navigate the repository to find the best set
of axioms which correspond to their intuition.
Base Course:
1) The user logs into COLORE. (it only works on COLORE atm)
2) The user names the relation(s) she wishes to formalize
3) User provides at least one example of their relation in "action" -
essentially a Tarski style model. The model can be represented visually, or inputted
as a plain text (see referenced papers for more on this).
4) ODT searches the ontologies in the repository (COLORE) to find
"Core-Hierarchies" and bounds in each hierarchy that match the user's
intuition.
5) For each Core Hierarchy, ODT presents a Tarski style model in the same
representation that the user inputted.
6) The user decides if this example corresponds to their intuition.
7) repeat 5-6 until search space exhausted
8) Present user with axioms for their intuition.
For a more detailed explanation of how this works, please see Chapter 4 of my
Master's thesis, or for a very
brief version, this paper: http://stl.mie.utoronto.ca/publications/design-repository.pdf
, or for something in between, wait for an upcoming journal paper. Chapters 1-3
of the thesis discuss the repository and how the media that is logic on the web
allows all this to happen.
========================================
2) Semantic Mapping Tool (SMT) - Same Domain
Motivation: I am an organization who has developed an ontology and would like
to interface said ontology with one developed by others. I need to determine
what / how and where our ontologies overlap. This works for any number of ontologies,
not just two.
Goal: Given test mapping axioms, determine how two or more ontologies are
"similar" and "different."
Actors: User + SMT
Triggers: Interoperability
Pre-requisites: COLORE
Solution: Create an image for each target ontology in the repository. Exploit
repository structure to determine "similarity" and
"difference." Again, this only works on COLORE, and requires at least
first order expressivity.
Base Course:
1) User logs into COLORE
2) User provides candidate mapping axioms. I.e. the user guesses that A is
related to B, but is unsure how and to what extent, wants to see how A is in
fact related to B. (This can also be automated...)
3) SMT generates an image of the user's target ontologies in COLORE, given the
mapping axioms
4) SMT analyzes images to determined "Similarity" and
"Differences" in the target ontologies
5) SMT provides the user with partial interpretations of the ontologies into
one another and into COLORE (in some cases, SMT performs automatic abduction).
For more on this, see chapter 5 of my Master's thesis or alternatively, wait
for an upcoming journal paper. The definition of "Similarity" and
"Difference" in the thesis are a bit outdated, you'll have to wait
for the journal paper for the most up-to-date definitions.
=======================================
3) Semantic Mapping Tool - Interdisciplinary
Discovery / Sharing
Motivation: I want to know if knowledge developed in some other domain is
useful for me. Can I reuse work done by others in a seemingly disparate field?
Is there a way to get around the research silos and specialized jargons that
have popped up? i.e. cell diffusivity and permeability being similar to
electrical resistivity... Essentially, I want to discover 'conceptual or
structural metaphors' that connect disparate domains to one another.
Goal: Support interdisciplinary knowledge sharing.
Actors: User + SMT
Triggers: Interoperability
Pre-requisites: COLORE
Solution: The exact same as Use Case 2, except the target ontologies are from
different domains. Again this exploits a basic advantage that the medium of
logic affords us :P
Base Course:
see above use case 2.
======================================
4) Concrete Use Case - Reseed
Reseed is a non-profit organization seeking to transform how people relate to
space, land and food. There are three vectors to this organization:
(1) Community Intervention by Example (i.e. actual urban farming)
(2) Education and Knowledge Sharing (relevant to ontolog - the technology side
of collecting, collating and sharing gained wisdom)
(3) Effect Policy Change in Government and Business
Very brief overview -- food security is becoming an increasingly important
issue; we often fight against nature instead of learning how to best work with
it; people in urban environments often don't have a connection to the land they
live on, nor are they connected to the provenance of the food they eat.
To address this, reseed wants to incorporate permaculture principles ( http://en.wikipedia.org/wiki/Permaculture
), to bring farming into urban environments. Take a moment and ask why every
house has green front lawn? It is a relic from Edwardian England (or before, I
forget). A green lawn demonstrated that one was wealthy enough to set aside
land for no purpose but to be green and short.... In this day and age, our
widespread adoption of lawns seems a bit absurd, especially in light of the
resources they divert, consume and larger looming environmental considerations...
In effect, in Ontario
at least, a green lawn is a forest in its infancy. It is why weeds love it, as
they are nature's way of trying to increase the biomass density en route to
re-establishing a forest, and ultimately, it is why they present an
ecological and economic drain... But I'm beginning to digress.
Basically, the relevant (to ontolog) technology integration strategy of reseed
includes developing an online knowledge resource for people to share their
experiences of urban farming, and for new comers to learn the basic principles
and apply them to their local context. For example, wisdom from the Amerindians
that almost disappeared with their genocide a few hundred years ago, suggested
that Corn, Beans, Squash and a few other plants should be grown together. The
corn grows quickly and provides a base for the beans to sprout, while the
squash, broccoli etc. provide protection to the roots of these plants from hard
rain and larger creatures. All three together help mitigate the need for
leaving land fallow... Anyway, capturing and sharing this knowledge is
valuable.
Not only that, but being able to take as input certain contextual (i.e. local)
variables regarding climate: i.e rainfall, temperature, humidity etc. allows
our system to suggest a number of plants and farming strategies for someone
looking to apply the collectively learned wisdom to their local contexts.
So to make this more concrete for OOR, Reseed is looking to explore all plant,
farming, seed ontologies from the perspective of supporting local farming
initiatives. This means:
1) Being able to do a search on related ontologies
2) Be able to browse them to determine if they address our needs.
3) Be able to extract, reuse and/or extend axioms if relevant
4) Create pointers to and from said ontologies
5) Develop a new ontology to support the rest of our technology integration
So it is a combination of the above use cases, plus some located on the ontolog
wiki. There's a bit more to the reseed use of ontologies, but this email is
already long enough and I think it gives a pretty good impression of a use
case.
Best,
Ali
--
Founding Director, www.reseed.ca
Social Technologies Adviser, www.pinkarmy.org
(•`'·.¸(`'·.¸(•)¸.·'´)¸.·'´•) .,.,
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/sio-dev/
Join Community: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/sio-dev/
Unsubscribe: mailto:sio-dev-leave@xxxxxxxxxxxxxxxx
Community Shared Files: http://ontolog.cim3.net/file/work/SIO/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?SharingIntegratingOntologies
--
Kojeware Corporation
|
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/sio-dev/
Join Community: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/sio-dev/
Unsubscribe: mailto:sio-dev-leave@xxxxxxxxxxxxxxxx
Community Shared Files: http://ontolog.cim3.net/file/work/SIO/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?SharingIntegratingOntologies (01)
|