W3C home > Mailing lists > Public > public-owl-dev@w3.org > October to December 2006

RE: representing non-state relationships in OWL

From: Pat Hayes <phayes@ihmc.us>
Date: Sat, 30 Dec 2006 12:19:26 -0800
Message-Id: <p0623091fc1bc76e31058@[]>
To: "Hans Teijgeler" <hans.teijgeler@quicknet.nl>
Cc: "'West, Matthew'" <matthew.west@shell.com>, "Sullivan, Jan" <jan.sullivan@informationlogic.co.uk>, <public-owl-dev@w3.org>

>Hi Pat,
>Thanks! Why did you not copy Chris and the forum?

Pure carelessness, sorry.

>ISO 15926 is for lifecycle (10-100 years) information about process plants.
>This includes process streams, so comparable with your ocean currents.

Indeed. You may be better sticking with the 'full 
4-d' view, then, as it is often called. But you 
should be aware that there are powerful forces 
who feel strongly that this is not the right way 
to do things, and will resist you. There are 
already some claimed 'standards' , such as DOLCE, 
which are founded and based on a systematic 
rejection of the idea of continuants having 
temporal parts. (THe obvious rejoinder, BTW, is 
that your ontology will be entirely about 
occurrents, which is a fancy way of saying that 
everything is a process. Which I think you will 
be likely to agree with.)

>Please see below for some responses.
>-----Original Message-----
>From: Pat Hayes [mailto:phayes@ihmc.us]
>Sent: Saturday, December 30, 2006 9:27
>To: Hans Teijgeler
>Subject: RE: representing non-state relationships in OWL
>>What about defining an OWL Class "Activity", then a subClassOf that
>>called "Meeting", and then the meeting you had in mind as an instance
>>of it. Peter then isHost of that instance of Meeting.
>>What still is missing is the temporal aspect.
>>Our solution for that so far found a cold shoulder in the OWL scene.
>>That solution is to define "temporal parts"[1]. The "temporal whole"
>>of Peter is Peter between his birth and his
>>(future) death. A temporal part is a part of that whole-life time span
>>in which a certain fact is true. In this case there is a temporal part
>>of Peter that "isHost" of that particular meeting.
>That is one way to do it. It is probably the most expressive and general way
>to encode temporal information, but the generality comes at a price.
>If you only wish to encode relatively simple facts about things that retain
>their identity through time but may change their properties
><HT> What to do with the properties that are no longer valid?

That raises another issue, how to keep track of 
'actual' time. THat is really an orthogonal issue 
to how to reason about time in general. My answer 
would be to abandon the idea that the KB is 
supposed to represent 'now', and instead treat 
everything in it as having an implicit or 
explicit date attached, and then reckon what is 
true 'now' by an external process of selecting on 
the basis of the actual date, what is sometimes 
called timeslicing. SO that the KB itself is 
timeless, and does not need to be continually 
updated simply because the clock has shifted. 
(Temporal DBs use a similar distinction between 
transaction and valid times.)

>, which includes most things in normal everyday circumstances (contrast,
>say, things like ocean currents or atmospheric weather events) you might
>consider a simpler scheme which distinguishes between continuants and
>occurrents (roughly, things and events.) Then there is often no particular
>need to distinguish temporal parts of continuants, since they are 'the same
>thing'throughout their lifetime.
><HT> There is hardly ANYthing 'the same thing' throughout one's lifetime.
>DNA comes to mind, but that is it.

I tend to agree, but many do not. Being the 'same 
thing' in this sense does not mean that it must 
be unchanging, only that is is identified in the 
same way. For example, you and I are arguably the 
same 'person' throughout our lives. There is a 
huge literature on these issues, and it sounds as 
though your view is what is sometimes called 
'process-oriented' - which is also my view, by 
the way, as well as that espoused by no less an 
authority than A.N.Whitehead - but this way of 
thinking, in which you and I today are not the 
same as you and I yesterday, is often claimed to 
be incoherent, wrong, philosophically in error, 
incomprehensible, etc. etc. So don't expect a 
smooth ride, is all Im saying. See for just one 
example the debate summarized in

Pat Hayes

>(On some views, it is in fact incoherent to even speak of temporal parts of
>continuants, but such philosophical rigor isnt very useful in practice.)
>Then you can treat continuants as individuals and their properties as
>'roles' whose values include a temporal aspect.
>For a survey of several alternative approaches to temporal descriptions, you
>might look at an old paper of mine:
>http://www.ihmc.us/users/phayes/TimeCatalog.pdf     especially section 2
>Pat Hayes
>>_schema.html , then select temporal_whole_part
>>PS I quote from that:
>>A <temporal_whole_part> is a
>><composition_of_individual> that indicates that one
>><possible_individual> is a temporal part of another
>><possible_individual>. The spatial extent of the temporal part is that
>>of the temporal whole for the period of the existence of the temporal
>>Relationships that apply to the whole
>><possible_individual> also apply to the temporal parts of the
>><possible_individual>, except when the relationships relate to the
>>temporal nature of the whole. So if a <possible_individual> is
>>connected so are all its temporal parts, but being a
>><whole_life_individual> is not inherited by its temporal parts.
>>NOTE Since <temporal_whole_part> is transitive (inherited from its
>>supertype) a hierarchy of temporal parts is possible, with a
>><whole_life_individual> at the top.
>>EXAMPLE 1 The relation that indicates that an operating period of a
>>pump is a temporal part of the pump can be represented by an instance
>>of <temporal_whole_part>.
>>EXAMPLE 2 The relationship that indicates that the time period known as
>>March 1999 is part of the period known as 1st Quarter 1999 can be
>>represented by an instance of <temporal_whole_part>.
>>Hans Teijgeler
>>ISO 15926 specialist
>>+31-72-509 2005
>>From: C Haley [mailto:cands589@yahoo.co.uk]
>>Sent: Monday, September 11, 2006 20:23
>>To: public-owl-dev@w3.org
>>I©–ve been reading through the various OWL documents and from what I©–ve
>>seen it appears that OWL is very good at representing state
>>information, either of classes or individuals, but does not seem to
>>allow for representing non-state relationships.
>>For example I can use OWL to represent the concept ‘man©– as a class,
>>represent Peter as an instance of that class, and I can define a
>>property stateOfHealth, and the concept ill, and create a triple to say
>>stateOfHealth(Peter, ill). This is representing a fact which defines
>>the state of an instance.
>>But suppose I want to represent the fact that Peter hosted a meeting in
>>the office yesterday.
>>Even if I created an artificial property ‘toHost©– and a blank node as
>>an instance of the concept ‘meeting©–, there is no way to attach the
>>time and location to the property.
>>Also I would want this property to derive from a URI representing the
>>concept of ‘hosting a meeting©–, but the OWL syntax seems to require
>>properties to derive from other properties, not from a generic URI. So
>>clearly this is not the correct way to represent an action.
>>Can anyone tell me if there are any
>>recommendations or documents describing the preferred solution to this
>>Alternatively is this an area where the existing OWL syntax/vocabulary
>>is likely to be extended - is anyone actively working on this issue at
>>present? Are there any draft recommendations in circulation?
>>Many thanks for any comments anyone can give.
>>Win a BlackBerry device from O2 with Yahoo!.
>>No virus found in this incoming message.
>>Checked by AVG Free Edition.
>>Version: 7.0.405 / Virus Database: 268.12.3/447 - Release Date:
>  >13-Sep-06
>>No virus found in this outgoing message.
>>Checked by AVG Free Edition.
>>Version: 7.5.432 / Virus Database: 268.15.29/607
>>- Release Date: 28-Dec-06 12:31
>IHMC		(850)434 8903 or (650)494 3973   home
>40 South Alcaniz St.	(850)202 4416   office
>Pensacola			(850)202 4440   fax
>FL 32502			(850)291 0667    cell
>phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
>No virus found in this incoming message.
>Checked by AVG Free Edition.
>Version: 7.5.432 / Virus Database: 268.16.0/609 - Release Date: 29-Dec-06
>No virus found in this outgoing message.
>Checked by AVG Free Edition.
>Version: 7.5.432 / Virus Database: 268.16.0/609 - Release Date: 29-Dec-06

IHMC		(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32502			(850)291 0667    cell
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes
Received on Saturday, 30 December 2006 20:49:48 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:58:14 UTC