In digging around for a tool to convert the .pds into a Topic
Map, I discovered Adobe’s RoboHelp advertises it can export a TopicMap. Does
anyone have experience with this tool and/or file structure? Does anyone have an
Adobe CS4 /license available for taking advantage of RoboHelp’s free
trial period (which I’m hoping we can try)?
Re: Relatioships in Ontologies
I’d be happy if we could just get the first relational operator
working, and it would be the word “change”, as broadly specified in
the Federal Enterprise Architecture’s Glossary, in the sense of a
pipeline model, as in X “changes to “ Y and Y “changes from”
X.
Thanks and have a good day,
Jim Disbrow
Program Manager, ET.gov
From: Bob Smith
[mailto:bobsmithttl@xxxxxxxxx]
Sent: Thursday, February 04, 2010 1:31 PM
To: Jack Park
Cc: ravi sharma; Disbrow, Jim; peter.yim@xxxxxxxx; Owen Ambur; John
Young; Wood, Elaine (NREL)
Subject: Re: Energy Water Nexus vis-a-vis the Climate Case Study for the
Ontology Summit 2010
You make a good case, Jack.
In addition to the ontologist's technical domain there exist
billions of dollars in policy and program work that could possibly benefit from
a cleaner and useable map of how energy, water, climate options relate to each
other in practice.
For example, a large battle is taking place over the best
ways of obtaining enough water for California's residents, businesses,
fisheries, and agriculture.
This is an urgent problem that will not go away soon.
So working along the lines you suggest, with a vision of the
scale of the problem, is useful.
PS - Attached is a presentation from Marco Gonzoles that I
presented at a Board meeting last night. It simply shows the battle over a
Water Desalination proposal.
Bob,
In answer to previous questions here, I know of no tool (doesn't mean there
isn't one) that would read Jim's diagram and transliterate it into a topic map
(or any other structure). I can only imagine wetware doing the work, perhaps
crowd-sourced ala wikipedia. That is, each node in his diagram is a topic, an
individual page in the wiki; each node plays roles in relationships with other
nodes. Those relationships should be topics as well (not traditional labeled
arcs or even HREFs). That way, they can be targets of debates, embellished with
tags, annotations, scopes, etc.
I am not talking in terms of "garden variety" XML topic maps; I am
talking about those which are created according to the dictates of the Topic
Maps Reference Model (TMRM) which makes no ontological commitments to
"Topic" "Association" (which is not a topic) and
"Occurrence". There are really only two objects in the universe
of the TMRM, make that three: A proxy for each subject (SubjectProxy) roughly
the equivalent to each of the TAO, and property (key/value) objects
(SubjectProperty) with which to define and represent each subject. A proxy is a
container of properties. The third object is a "legend" which
publically defines the property types used to identify the subjects, and the
merging rules (the way those properties are interpreted in your map for
purposes of merging proxies--remember, the prime discipline in topic mapping is
adhering to the dictates of a good map: one and only one location for any given
subject).
Nothing in what I am saying here takes anything away from either XML topic maps
or from ontologies of any kind. They each suit their purposes well. My views
are based more on realities I sense (interpret) when observing the sensemaking
practices of really intelligent people trying to figure out how to merge
ontologies, how to model universes of discourse, etc. I am suggesting that
people collaboratively map their territory with less rigor in the specifics of
representation schemes; let those fall out later. Just answer these questions:
what are the subjects? and what are the attributes (properties, roles, etc) of
those subjects?
Energy and Climate, two enormous fields of discourse, will take a while.
Jack
On Thu, Feb 4, 2010 at 9:18 AM, Bob Smith <bobsmithttl@xxxxxxxxx>
wrote:
That sounds like a very good suggestion: start with a
general topic map of "the territory" as an exploration platform -
when ready enough.
Some questions then emerge from your suggestion - one of
which is the map between Jim's PDF file of the Cycles and "the
territory" of interest (Scope).
On Thu, Feb 4, 2010 at 8:49 AM, Jack Park <jackpark@xxxxxxxxx>
wrote:
Let me just offer a short comment before I blow out of here
for a meeting. I mention topic maps because I eat and breathe topic maps. That,
in my view, is not different from ontologies; we can do ontologies in topic
maps, but we can do a few more things than a concise ontology anticipates.
That's not particularly important to ontology engineers who are tasked to
control laser surgical instruments, but to those who are assigned tasks of
understanding complex systems like biological, ecological, climate, etc, topic
maps add something. Let me explain.
I can import the entire cancer.owl (26k classes) document into a topic map
platform: it comes in as an owl file and remains an owl ontology. I then
"harvest" that ontology to create topic maps "subjects" and
their relationships, nomenclature, etc. In that sense, I am wrapping the
owl entity with a topic map, not replacing it. I can still run the inference
engines on the owl document, but also, over in the collaboration platform, I
can embellish the cancer subjects with tags, annotations, links to web pages,
structured dialogues about this or that, etc. Later, if things suggest
improvements or changes to the owl document, they can be mapped back. I can do
the same thing with database artifacts. (No, I cannot walk on water ;-)
That is why I suggested starting with a general topic map of the territory as
an exploration platform.
The open source platform I am developing to do this is based on Semantic
MediaWiki coupled to an IBIS (structured) conversation extension and my topic
map (RDF) platform. It's not ready for prime time now, but will be within the
next very few months.
It's just a suggestion. Nothing more than that.
Cheers
Jack
On Thu, Feb 4, 2010 at 8:02 AM, ravi sharma <drravisharma@xxxxxxxxx>
wrote:
Have you tried AllegroGraph and Gruff? They are free up to a
million triples? (I think - please check, not 100% sure)
I have yet to study the topic map or the graphic for E-W
cycle.
On Thu, Feb 4, 2010 at 10:49 AM, Disbrow, Jim <Jim.Disbrow@xxxxxxxxxxx>
wrote:
Good morning, Jack / Peter / Bob / Owen
/Ravi / John
The structure of our ontology will
probably wind up being constrained by technology, so picking the technology
(e.g, the ontology building tool) will define the structure (a less than
optimal position, but we are constrained in many directions). Let’s try
to make the wisest choice!
Do you have a preferred tool for
creating our ontology? Do you know if (and how) this preferred tool handles
“associations” (a Topic Map term) or relationships (an OWL/RDF
term) robustly? Can the tool start with a vector-oriented .pds file, such as
the .pds behind the Physical Fuel Cycle .pdf (posted in Google Docs “EnergyWaterNexus@xxxxxxxxx”)?
I’d like to start with this graphic and rebuild it into a usable ontology
with explicit relationships between the various topics – and am prepared
to do it myself, manually, if necessary.
Or do you know of (or have) a free topic
map builder that can be used to create our ontology? Or a tool that would take
an output from one of the free ontology builders (e.g., TopBraid, Protégé) and work
for us? Do you know if (and how) these tools handle “associations”
(a Topic Map term) or relationships (an OWL/RDF term) robustly? If we can
follow Jack’s suggestion (below), it would be beneficial to the Case
Study’s outcome (my opinion).
On a related subject: DOE Island in
Second Life has opened its doors (in Beta), and Energy Ant is a virtual bot
there, using an AI that is driven by an AIML file. Per existing agreements, I
will have a large degree of input and control in the process for building out
the AIML file. I was thinking to use this in collecting the questions that
people (SL visitors) would like answered – and make the ontology
responsive to these question. If we could use these questions to create an
ontology directly, it would be great, but I don’t know if this is doable
– what do you think? Do you know of a tool that would create an AIML file
by transliterating from the .pds behind the Physical Fuel Cycle graphic posted
in Google Docs?
On another related subject: We are
currently getting the open-source code used by the US Air Force for testing
ontologies for robustness, so stay tuned for progress on this. We could use the
Ontology we build as the basis for testing whether we can answer the questions
that are posed to Energy Ant correctly – or not. The creators of this
tool have generously offered to help (with constraints based on their current
overloaded work schedule), and have used the free TopBraid tool in their
projects.
Who else should we be directing these
questions toward?
Thanks and have a good day,
Jim
The problems associated with EnergyWaterNexus@xxxxxxxxx
seem to have been fixed, and it should be working for all of us now.
From: Jack
Park [mailto:jackpark@xxxxxxxxx]
Sent: Thursday, January 28, 2010 6:58 PM
To: Disbrow, Jim
Subject: Re: Energy Water Analysis 2010
Jim,
These seem to be coming through fine. Replying to this email directly doesn't
work.
The pdf diagram seems ripe for transliteration into a topic map as a first step
towards an ontology. By putting it into a topic map, we can play with it in a
variety of ways before distilling it into any official ontology.
Jack
--
Thanks.
Ravi
(Dr. Ravi Sharma)
313 204 1740 Mobile
|
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/energy-water-nexus/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/socop-forum/
Unsubscribe: mailto:energy-water-nexus-leave@xxxxxxxxxxxxxxxx
Community Shared Files: http://ontolog.cim3.net/file/work/energy-water-nexus/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?EnergyWaterNexus (01)
|