Measured Qualities

[swartik]

I have some questions about qualities obtained through measurement and how to represent them.

Suppose I have a Material Entity e and I want to express that its Temperature t is 20 degrees Celsius. Here's my assumption on how to express this in CCO.

I define an Act of Measuring individual aom. Its input is e. Its output is an Interval Measurement m.

I assert the following relationships:

• aom has-input e.
• aom has-output m.
• m has-value 20^^xsd:integer.
• m is-an-interval-measurement-of t.
• e bearer-of t.

(As far as I know, the data property to express the integer value of an ICE isn't finalized yet. has-value is my approximation.)

According to Modeling Information with the Common Core Ontologies 1.3, you express the unit of measurement by:

• Defining an IBE ibe.
• Asserting:
  • ibe is-carrier-of m.
  • degree-celsius-measurement-unit is-measurement-unit-of ibe.

So, two questions:

1. Is the first set of elements and property assertions the correct way to express temperature – more generally, the correct way to express a continuant's quality that's known through measurement?

2. In CCO 2.0, will property is-measurement-unit-of continue to be a subproperty of is-carrier-of?

[CarterBeauBenson]

@swartik

1. Looks right to me, although temperature likely is not a quality of an object, but a process profile that is an occurrent part of some process.
2. I believe so. You can look at the Information Artifact Ontology (IAO) to see an alternative view where values and units dangle from ICEs.

Carter

[swartik]

@CarterBeauBenson

1. In CCO V2.0 (commit 0fae911), class Temperature is a subclass of class quality. Is there a move afoot to make Temperature a subclass of class process profile?
2. Thanks for the pointer to IAO. I will take a look.

[alanruttenberg]

1. Looks right to me, although temperature likely is not a quality of an object, but a process profile that is an occurrent part of some process.

Yikes. Absolutely not. Temperature exists in whole at any time at which it exists, and like any other continuant, can change in time. It is represented as a quality in every ontology I'm aware of. It's almost the prototypical quality.

@swartik I have proposed the value specification pattern to CCO some time ago. It is what is used in OBO, by Dave Lutz in his work, and in my own work, among others. Value specifications are generalizations measurements but have the same "shape" as measurement. But they can be settings our outputs of simulation, for example.

The value specification proposal looks like what you have represented for the most past (and uses has-value as the property), but also has-unit is also asserted on the ICE, where it belongs.

[CarterBeauBenson]

@alanruttenberg
To be clear, I am non-committal to the process profile view.

Seems like the particles wholly exist, but their movement, expressed as energy, is an occurrent of some kind. Process profile (I know you disapprove of the class) would be the rate that the particles are vibrating at the time of measurement.

My doubt of the quality view stems from some conversations with materials engineers last summer. I may not have portrayed the argument well.

[swartik]

@CarterBeauBenson,

1. Degrees Kelvin now being defined in terms of Boltzmann's constant, expressing temperature as a process profile will please physicists.
2. The OED defines temperature as follows (emphasis mine):

   The state of a substance or body with regard to sensible warmth or coldness, referred to some standard of comparison; spec. that quality or condition of a body which in degree varies directly with the amount of heat contained in the body, and inversely with its heat-capacity; commonly manifested by its imparting heat to, or receiving it from, contiguous bodies, and usually measured by means of a thermometer or similar instrument. (Now the ordinary sense.)

So as usual it gets down to: Who is this ontology serving?

[alanruttenberg]

A rate does not a process profile make. Many process-dependent quantities are continuants in the same way other qualities are in that they continuously vary, maintaining identity. There is the speed of the car as shown on the odometer - always "the" speed of the car, continuously changing.

In many ontologies that deal with energy (maybe all I've seen, which is several), energy is considered a quality. Temperature is effectively average kinetic energy of molecules.

expressing temperature as a process profile will please physicists

This sounds like nonsense to me. It's a rare physicist who will understand what process profiles are. I don't.

[CarterBeauBenson]

re:1

If I had to guess, probably not. It is just something to flag that temperature in CCO may be misclassed.