Jack,
“Type” and “Class” are pretty much used interchangeably in ontology engineering. And the difference between “intrinsic” and “extrinsic” largely maps respectively
to “substantive” and “accidental” properties but with some caveats. 1) eye color may be substantive (if it cannot be modified), but hair color is not; 2) “run a mile in X minutes” is accidental, and is probably a relation such as bestSpeed(Person, FootSpeed)
or an event (see below).
Attributes are typically either modelled simply as ranging over simple data types or more complexly as ranging over possibly complex properties. However, your
“behavior” notion is probably nearly always a relation between two objects. And more complex behavior is modelled as an event:
run(Event, Person, BestSpeed, FootRace, Distance). Which might have an instantiation such as:
run(e1, jim-ryan, 3:51.1[minutes], run-race23, mile).
I.e., there is a run event e1 that the person jim-ryan ran of kind run-race for a mile distance.
Or something similar.
Thanks,
Leo
From: ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:ontology-summit-bounces@xxxxxxxxxxxxxxxx]
On Behalf Of Jack Ring
Sent: Sunday, May 06, 2012 3:02 AM
To: Ontology Summit 2012 discussion
Subject: Re: [ontology-summit] First Model Bench Challenge
Matthew, We may have quite different labels for the same concepts so let me try some others
In systems work it is important to discern between intrinsic attributes and extrinsic attributes.
Examples for a person are intrinsic = blue eyes while extrinsic = can run a mile in 4 minutes.
Examples for a system are intrinsic = what it is (content and structure) while extrinsic = what it does (stimulus:response behavior).
In object technology we talk about classes (content and structure) as distinct from objects that are instantiated from the pattern provided by the class and, properly engergized, can do things like respond to qualified
messages.
In numbers we can talk about structure, e.g., Date = 5/5/2012 and about usage, e.g. can be added or subtracted from other Date values but not multiplied.
Software people use the word 'type' to designate usage. Seems to me this indicates one aspect of extrinsic attribute.
So I consider class to be semantically equivalent to content/structure and type to be semantically equivalent to behavior potential.
Any suggestions about how to clean up my act?
On May 3, 2012, at 2:18 PM, Matthew West wrote:
I do not know what distinction you make between class and type. I make none.
This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.
Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.
What if marriage is established by husband and wife and license or certificate is simply an associated artifact?
Accordingly, common law, becomes a class of marriage, not a type.
On May 3, 2012, at 7:56 AM, Matthew West wrote:
I didn’t say you did not need a suitably underspecified term, or at least that it would be useful. I said that this would
not be sufficient to do anything useful, since not everything that is recognised as a marriage by one authority is recognised by another. This is really a case where you need to catalogue the different types of marriage (recognised by different authorities)
and the cross recognition that exists. This is potentially a long list (by shopping list standards).
This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.
Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.
Surely for the purpose of federation, you need to be able to agree on a suitably under-specified common term, that has been abstracted enough that all the possible models you
want to federate can have their term for marriage and any related terms be represented as specializations of those. Which means you need to understand the breadth of the issues that Amanda describes in order, not to model them, but to ensure that any ontologies
which turn up with those concepts in will not break the model.
For instance if you start with a "most general" representation of marriage that assumes one male and one female, then some ontologies that may exist out there will not be able to be federated.
Mike
On 03/05/2012 08:18, Matthew West wrote:
Even if federation were a purpose, I think you would still want to know which authorities recognised what kind of marriage.
This email originates from Information Junction Ltd. Registered in England and Wales No. 6632177.
Registered office: 2 Brookside, Meadow Way, Letchworth Garden City, Hertfordshire, SG6 3JE.
It doesn't fly in the face of ontological principles -- or my views on methodology -- at all! As I mean "purpose," Federation *is* a purpose (purpose != system). A very good
one, I might add. :-)
Having worked on ontology for federation, and on ontologies to model the information content of some of the data being federated, I agree emphatically that when ontologizing
for federation, it is of great importance to avoid saying too much about the concept. The method is to find a level of definition that makes minimal commitments while being usefully distinctive and giving somewhere to plug the more specific concepts into,
just as you say.
Having federation as a purpose gives us that guideline. If, say, we were providing semantics for demographic data, we might need one or more of the more specialized definitions.
Those can be integrated via the minimal concepts in the ontology that serves the federation primarily.
Note how different that is, though, than where the discussion was going. Federation was not a stated goal, and discussion was down into the details. So, we *could* assume federation
as the purpose, and continue, or we could assume some other purpose, and perhaps continue. But the exercise did not appear to have such a scope.
The actual purpose of the example was to demonstrate composite concepts – i.e. that the concept of “spouse” is meaningful within the
context of marriage. However we could assume that the further purpose is to define common concept to allow information federation regarding marriages, spouses and such.
Your premise is that it is “rarely feasible” to have such a model without s specific purpose (i.e. system?) in mind. Yet, one of the
fundamentally purposes of federation is to correlate information from independently conceived systems. If there are no concepts shared beyond those “designed in” then no communication between or involving such systems is possible.
The complexities of how various institutions decide to classify marriage is interesting but, to my mind, not that relevant. If we over
axiomatize concepts they are never correct. A federation model should seek to say as little as possible about a concept to define it (specializations of that concept may then add more details – such as common law marriage in Virginia). The details of why
certain individuals or institutions may or may not consider certain individuals to be in a state of marriage is less important than they do. For example, if I were asking for a list of marriages across a collection of jurisdictions, I will get a set. How
they derived each subset is their business – what we would like is for the general definition to be compatible with whatever policies each jurisdiction imposes. What jurisdictions we “trust” is another matter.
I know this flies in the face of some peoples ontological principles, in that we may not know the differentia for marriage even if
we have a well-defined and identifiable concept. However, in real practice the concept does exist and is in regular use in some form that allows us to communicate beyond the scope of a designed system.
Subject: Re: [ontology-summit] First Model Bench Challenge
I'm going to answer this, in part because it is interesting, and in part because it illustrates why I generally don't respond to this sort of group attempt to formulate a generally good
ontology of (x), when neither Subject Matter Experts nor readily researchable use cases are involved. I think the attempt is rarely feasible, and usually involves a wrong methodology for applied computational ontology development. Just about every subject
is much more complex than expected, and more contextual, when you actually get into it. In practice, working ontologists do (or should, as a rule) determine how detailed and complex the model needs to be, and with respect to what, with reference to intended
scope: what does this model need to support (including potential for reuse/extension)? What assumptions of the use context should be made explicit and then used to bound the model? If we have no such reference, the exercise is endless. If we have one but have
no experts, we are modeling what we, professional ontologists but amateurs in the area of concern think about it, not the reality or practice.
In this case, I'm not an expert, not a lawyer, etc... but may have above average knowledge for the group, since I have some relevant practical experience, requiring getting to know several
states' laws regarding common-law marriage and also its treatment by, for example, health insurance providers and the IRS.
Common-law marriage is governed by law, but different jurisdictions set the specifications differently (or do not recognize common-law marriages at all). The treatment is, in fact, often
unclear.
Let clause(a) = "P1 and P2 are competent adults, as defined in the state of jurisdiction"
Let clause(b) = "P1 and P2 are competent adults, as defined in the state of jurisdiction, and exactly one of P1 and P2 is male and exactly one of P1 and P2 is female, as defined in the
state of jurisdiction"
Note that legal competence and legal gender each may change over time, and it can therefore be the case that while the length of cohabitation of P1 and P2 is known, either the competence
or the legal gender of P1 or P2 may change during that cohabitation and by doing so cause P1 and P2 to satisfy clause(a) or clause(b) when they did not before, and the date of the change may be unclear under the law in the state of jurisdiction
Most states in the US that have common-law marriage currently have something based on clause (b), but not all do.
Here, are some examples of different U.S. State treatments:
(1a, 1b) If [clause(a)|clause(b)], and P1 and P2 live together for >=n years, then P1 and P2 are entitled to be treated as married by the state, should they request such treatment.
(2a, 2b) If [clause(a)|clause(b)], and P1 and P2 live together for >n years, then P1 and P2 are married in the eyes of the state, and if there is a separation, division of property, or
subsequent marriage of either P1 or P2 to another party, laws governing divorce apply.
(3a, 3b) If [clause(a)|clause(b)], and P1 and P2 live together currently and represent themselves as married to the community, then P1 and P2 are married in the eyes of the state. [No
minimum time requirement]
(4) There is no common-law marriage, but a common-law marriage established in another jurisdiction will be recognized if it was recognized by that jurisdictions while the parties lived
in that jurisdiction. If it satisfied the law of that jurisdiction but was not formally recognized by that jurisdiction, it will not be recognized [due to the difficulties of one jurisdiction's authorities deciding on the satisfaction of another jurisdiction's
laws].
(5) There is no common-law marriage, and a common-law marriage established in another jurisdiction, even if recognized formally by that jurisdiction and other authorities such as the US
Federal Govt, will not be recognized if there is anything unclear or inconsistent with this jurisdictions laws.
(3b) is the case Texas, by the way. That one is perhaps the most fascinating, because under it standard conditions, saying you're married makes it true. A speech act, indeed!
The US Govt will recognize a common-law marriage based on (3b), e.g. for tax purposes, if the representation is performed via some witnessed, notarized document, but may date the marriage
as beginning at some prior time.
(5) is the case in Virginia, according to several Virginia family law experts/practitioners of my acquaintance, Virginia courts will look for a way to decline to recognize common-law marriages
whenever possible.
In the case which gave rise to my awareness of these issues, a common-law marriage was established in Texas (3b), recognized by the US Federal Govt and various other entities, redundantly
re-established via civil ceremony in Massachusetts in a later year, then dissolved in Virginia. The Virginia experts representing the parties to the case were not in broad agreement, but did agree that the case ought to proceed using the date of the later
Massachusetts ceremony, not the date of the Texas recognition. One reason was the risk of having to start over if the court simply refused to recognize the Texas marriage. The primary reason, however, was that none of these experts felt that they could establish
a challenge-resistant start date for the Texas marriage! So it was, in their view, entirely unclear, despite its being governed by law, perhaps partly because of its government by too many laws!
So, on the specific point, the start of the situation can indeed be unclear. And on the general point, even the most basic structure of the situation and the event vary, even if we are
considering only the legal concepts and not the religious or other cultural concepts, and to represent any of them adequately for any use, you probably need to have some expert resources.
On Wed, May 2, 2012 at 4:52 PM, Wartik, Steven P "Steve" <swartik@xxxxxxx> wrote:
I wouldn't say it's unclear. It's governed by law, right? If you live together for some number of years, you are entitled to have the state recognize you as married. The date and time may
not be as precise as if you have a wedding ceremony but can nevertheless be established. (Okay, I'm not a lawyer, and that's probably not the exact legal definition.)
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012
Community Portal: http://ontolog.cim3.net/wiki/
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012
Community Portal: http://ontolog.cim3.net/wiki/
--
Mike Bennett
Director
Hypercube Ltd.
89 Worship Street
London EC2A 2BF
Tel: +44 (0) 20 7917 9522
Mob: +44 (0) 7721 420 730
www.hypercube.co.uk
Registered in England and Wales No. 2461068
|
_________________________________________________________________
Msg Archives: http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe: mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files: http://ontolog.cim3.net/file/work/OntologySummit2012/
Community Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2012
Community Portal: http://ontolog.cim3.net/wiki/ (01)
|