On Sep 24, 2009, at 11:36 PM, Gunther Schadow wrote: (01)
> Friends, after today's meeting and following this discussion,
> I think it's a critical decision to set a scope for this
> project. Do you want to define units or quantities? If you
> want units, then you do not need to define mass or discuss
> what it means to apply 180 cm to yourself. (02)
We do not (and indeed cannot) *define* mass. But we do need to have
enough in the ontology to be able to apply it, by connecting its
concepts to those in other ontologies concerned with describing
engineering artifacts, biological systems, etc. (03)
> If you want to define quantities, you should consider the
> VIM's is lacking much detail that can branch out from
> 1.1 (1.1) quantity "property of a phenomenon, body, or substance,
> You can not apply 1.3 kg to your lump of cheese without
> specifying the property. Same as 180 cm to yourself. (04)
You have to refer to the property, but you do not have to "specify"
it. (Actually Im not sure what you mean by 'specify', so maybe....) (05)
> In some standards, formalized by IUPAC (there is a book
> called the "Silver Book" and the following literature:
>  Dybkaer R. An Ontology on Property for physical, chemical, and
> biological systems. APMIS. 2004; 112(Suppl. no. 117).
>  C-NPU. Subcommittee on Nomenclature, Properties, and Units in
> Laboratory Medicine. Available from: URL:
>  LOINC®. Logical Observation Identifiers Names and Codes.
> Available from: URL: http://www.regenstrief.org/loinc
> While the above are about laboratory medicine, the principles
> apply for physics as well.
> To specify a property, they say, you need to specify the "system"
> and it's "component" and "kind of property". (06)
Well then they are wrong. While this might be good discipline, it is
not *necessary* in order to write an ontology. It is also, by the way,
highly idiosyncratic. I would strongly suggest that we avoid the word
"system", which along with "context" is notoriously lacking in
precision, and only serves to confuse when it is used in a formalism. (07)
> So: "my (system) body (component) length (kind of quantity) is 180 cm"
> or: "this lump of cheese's (system) total (component) mass (kind of
> quantity) is 1.3 kg"
> To split out "component" from "system" does not always make sense (08)
Quite. This idea seems (?) to be based on a rather simplified ontology
of parthood having crept into the descriptions. (09)
> IMO (like above the total mass is does not require a component).
> It appears even from the VIM that to recognize a property, you need
> to have an object which possesses that property. Defining exactly
> the property is IMO too wide for this ontology to include, because
> it could be just about anything, e.g., it could be the "frequency
> by which I notice that I have not eaten dinner" -- far too much
> to specify in your ontology. (010)
Wrong way to think. The ontology does not need to *specify* very
broad categories like this. It only has to *allow* them, by not
imposing needless constraints. We can have a category of
'spatiotemporal entity' or 'continuant' (or... choose your favorite
high-level physical category) and have quantities be properties
(actually, property values) of them. (011)
> So, you probably need to exclude defining properties from your
> scope and begin with kind of quantity. But I suggest you even take
> kind of quantity as a primitive and focus only on dimensions.
> Since you are making UML models from the VIM, I am attaching a Word
> document that has a draft which I have made about a year ago based
> on VIM quotations. It's one of those I showed in today's slides
> without much further explanation. You may find some ideas there
> to use.
> Finally, to the rest of the discussion:
>>> David Leal wrote:
>>>>>> There are at least two ideas of what the members of "1.3 kg" are,
>>>>>> a) the members are mass tropes of different individual quantities
>>>>>> of matter;
>>>>>> b) the members are different individual quantities of matter.
>>> I wrote:
>>>>> I don't know what David's (b) means. I would have said the
>>>>> of the equivalence class designated "1.3kg" are either:
>>>>> a) 'mass tropes' of different individual things, or
>>>>> b) measurements of the 'mass tropes' of individual things.
>>> Matthew West wrote:
>>>> MW: I would say neither of these, but the individual things
>>>> The measurements are of these.
>>>> E.g. My lump of cheese is a member of the 1.3Kg equivalence class.
> Clearly that cannot be. (012)
Well no, it clearly CAN be. (I don't like that way of putting it, but
it does make sense.) (013)
> It is (014)
What is, exactly? Matthew was referring to a lump of cheese, right? (015)
> (a) the particular properties (this mass of this lump of cheese) (016)
No. The mass of the lump is clearly not the same thing as the lump
> (b) the measurement results (the reading of the scale with this lump
> of cheese on it) (018)
Even less is a lump of cheese a measurement. This is a category error.
BTW, a mass is not a measurement, either. (019)
> (c) the expression of the measurement result, i.e., my statement
> that the mass of this lump of cheese is "1.3 kg". (020)
Even more less is a lump of cheese a statement or expression. This is
a *logical* category error. (021)
> I feel that the extension of the equivalence class (022)
?? Equivalence classes don't have extensions. (023)
> "1.3 kg" is
> (c) or even less, and that only by a relation from the elements of
> that equivalence class to the properties and ultimately objects
> possessing those properties is it that we think that these properties
> or objects are the elements of the equivalence class. But that is
> not really useful. Of course you can always define the equivalence
> class by transitively including the expression, measurement,
> property and object, but that would violate Occam's razor.
>>> This may be true of 'mass', but it doesn't generalize. The person
>>> Barkmeyer is not a member of the 180cm equivalence class.
>> Thats because 180 cm isn't a meaningful quantity applied to you,
>> without further specification. It would like saying that you don't
>> have a voltage, which is true but not very interesting. But as soon
>> you specify a meaningful length class (height, waist size) then the
>> analogy with mass works just fine. And you are a member of an
>> equivalence class of volumes, for example.
>>> The height of
>>> EdBarkmeyer is a member of the 180cm class. But the waist size of
>>> EdBarkmeyer is a member of a different class (predictions of one
>>> of my
>>> former students notwithstanding). In a similar way, the existence
>>> of a
>>> thing may be a member of one duration class, while some other
>>> is a member of a different duration class.
> I think that's too much already. The member of the 180 cm class is
> 180 cm and 0.18 m. (024)
The class being talked about here, though, is the one of all spatial
extents (or some such) which are equal in length and include a
measuring tape exactly 180 cm long. The things being measured, rather
than the measurements. (025)
> There may be a relation between the property
> instances of Ed Barkmeyer's height and/or waist circumference to that
> equivalence class, but that is more than a useful theory of units
> of measure requires.
>>>> MW: <snip> I think it is important to
>>>> distinguish two concepts:
>>>> 1. The mass something has,
>>>> 2. A measurement of the mass something has.
> "Measurement" is a dangerously ambiguous concept. Nothing "has
> a measurement". It is we who "make measurements". Thing have
> properties which we measure. A measurement is the "process of
> experimentally obtaining one or more quantity values that can
> reasonably be attributed to a quantity". (026)
The only viable philosophical stance for practical ontology
engineering is a robust naive realism. Things really do have a mass.
We can measure the mass that things (actually) have. I do not see
where there is any ambiguity in this statement. (027)
Pat Hayes (028)
>>> I think this is exactly the same pair as David's (a) and (b) above,
>>> the substitution of "something" for
>>> "individual quantity of matter", and my (a) and (b) above is the
>>> with a slightly different substitution. It seems to me that we are
>>> in violent agreement, except that we can't agree on the spelling of
>>> "some thing".
>>>> We should not be choosing to be interested in one or the other, but
> Actually, I would say you should choose to be interested in
> neither for the sake of the simplicity of the theory of units
> of measure. (029)
> Gunther Schadow, M.D., Ph.D. gschadow@xxxxxxxxxxxxxxx
> Associate Professor Indiana University School of Informatics
> Regenstrief Institute, Inc. Indiana University School of Medicine
> tel:1(317)423-5521 http://aurora.regenstrief.org
> <Quantity and Units Model.doc>
> Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/
> Subscribe: mailto:uom-ontology-std-join@xxxxxxxxxxxxxxxx
> Config/Unsubscribe: http://ontolog.cim3.net/mailman/listinfo/uom-ontology-std/
> Shared Files: http://ontolog.cim3.net/file/work/UoM/
> Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard (030)
IHMC (850)434 8903 or (650)494 3973
40 South Alcaniz St. (850)202 4416 office
Pensacola (850)202 4440 fax
FL 32502 (850)291 0667 mobile
phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes (031)
Message Archives: http://ontolog.cim3.net/forum/uom-ontology-std/
Shared Files: http://ontolog.cim3.net/file/work/UoM/
Wiki: http://ontolog.cim3.net/cgi-bin/wiki.pl?UoM_Ontology_Standard (032)