wdduncan / capabilities-ontology

Repository for coordinating work for potential capabilities onotology
0 stars 0 forks source link

Use case for capabilities #1

Open wdduncan opened 6 years ago

wdduncan commented 6 years ago

Use cases: Are there domains we want to focus on representing? Of these domains are there standard questions/queries that data represented using the ontology should be able to answer?

In the graphic, capability is a kind of realizable entity. That seems like a safe starting place. But what about a skill? Can entity lose a skill w/o being physically changed? This seems unlikely to me, which would mean a skill is a type of disposition.

The question above holds for capabilities in general: can an entity lose a capability and not be physically changed? If the answer to this is "yes", then we need to decide if capabilities are a kind of role, or a sibling of role.

Also, and this is pure speculation on my part, is a capability a kind of realizable entity (i.e., only exhibited/manifested in processes)? If not, then perhaps a kind of relational quality?

Having some use cases will help greatly in addressing these questions.

wdduncan commented 6 years ago

So the way we describe Capabilities in the military is (roughly):

Military Units (Organizations) are the bearers of Capabilities...for example: Maneuver Capability Lethal Engagement Capability Communications Capability Reconnaissance Capability etc.
These Unit Capabilities are made up by persons with individual skills:
Land Navigation Skills Marksmanship Skills Gunnery Skills First Aid Skills Radio Operation Skills etc.
This is measured by "Operational Readiness" ratings so that If one 'Tank Gunner' breaks his leg and cannot perform his duties then... The 'Tank Crew' is down and... The 'Tank Company's' Operational Readiness rating goes down. We can just put some new person in his Role as 'Tank Gunner'...they have to get qualified through hours of testing (what we call Tank Crew Qualification or Gunnery). Pieces of Equipment also have Capabilities such as: Weapon Rate of Fire Weapon Range Radar Capability Night Vision Capability etc.
The military has all of this worked out to a very fine level of detail (and documents everything) so that it makes for a good use-case to establish more general principles/theories pertaining to 'Capability' and 'Skill'.

I think we will find that the Systems Engineering domain is similar in its level of detail and documentation.

So whatever we propose for definitions and relationships between class terms, will reflect (enhance) those two domains.

Your question, "... are there standard questions/queries that data represented using the ontology should be able to answer?" will help determine:

the required classes and subclasses the relationships between them, and the level of granularity required (I guess that is a restatement of 1) We should distinguish between Reference Ontology (for extending from)
Application Ontology (quite specific queries)

wdduncan commented 6 years ago

Skills (in this context) sound like BFO functions: a person goes through training and as a result of the training is physically changed (acquires new muscle memory etc.) so that he or she can now (adequately) perform new tasks. For example, after basic training, a soldier now has a skill (at a certain level of proficiency) to fire a weapon. If the solider loses this skill, some physical change has taken place (injury, degradation of muscle memory etc.).

A soldier's particular job within a section/squad seems to be a role. For example, the tank gunner is the person who has been assigned that role on the tank crew. Hopefully, the person assigned a role also has requisite skill (i.e., function), but, as we all know, mistakes are sometimes made in doing this. You can be assigned a role you do not have the skill to carry out. For example, an infantryman most likely does not have the skill to be a helicopter pilot. Moreover, the solider's role can change without necessitating a physical change. A tank gunner can be promoted to tank commander without necessary physically changing.

Measures of operational readiness (and the like) seem like types of information content entities.

If you find these account lacking, can you say a little more about what is not satisfactory? Are the labels "function" or "role" worrisome? If so, an alternative label can be used. For example, you can use the label "capability" for "function".

wdduncan commented 6 years ago

I do not think it is clear as to whether a new category of capabilities is needed. For instance: Why do you have organizational capability instead of organizational function? What are you trying represent with the label "capability" that isn't already represented by the semantics of BFO:function? Is it aggregate aspect of an organization? Aggregates can bear multiple functions. An offensive has the function of blocking members from the opposing team and whose members each have slightly different functions (e.g., the center has the function to hike the ball). The offensive team (as a whole) has a function to move the ball into the end zone. This function of the offensive team is is some sense dependent upon the function of offensive line, which in turn is dependent on function of the individual members of the line. But, now, it seems, we are discussing relations amongst functions, and not necessarily a new category of things.

Again, if it is the labels that are the problem, alternative labels can be used.

If I was to represent your example in OWL, it would roughly (i.e., using pseudo syntax with time index ignored) look like this:

define classes

tank gunnery role subclass of role tank gunnery skill subclass of function (alternative label "capability") military organization subclass of organization organizational capability subclass of function (alternative label "capability")

define instances

sgt jones instance of human being sgt jones gunner role instance of tank gunnery role sgt jones aiming skill instance of tank gunnery skill tank crew 1 instance of military organization tank crew 1 maneuver capability instance of organizational capability tank crew 1 engagement capability instance of organizational capability

relate instances

sgt jones bearer of tank gunnery role sgt jones bearer of sgt jones aiming skill sgt jones member of tank crew 1 tank crew 1 bear of tank crew 1 maneuver capability tank crew 1 bear of tank crew 1 engagement capability

Question: What you trying to represent with the individual tank crew 1 capability? Is this an individual/instance or a general class?

Here is an alternative representation using it as a general class: tank crew 1 capability subclass of organizational capability tank crew 1 maneuver capability instance of tank crew 1 capability tank crew 1 engagement capability instance of tank crew 1 capability

What does this representation not address that your representation does?

wdduncan commented 6 years ago

Bill D:

I will go line by line...

However, I do not think it is clear as to whether a new category of capabilities is needed. For instance: Why do you have organizational capability instead of organizational function? What are you trying represent with the label "capability" that isn't already represented by the semantics of BFO:function?

Organization's are the bearers of both Functions and Capabilities:

  1. An Organization has a Function, but it may not be Capable of carrying it out.
  2. An Organization is 'bearer of' both Organizational Function(s) and Organizational Capability(ies)
  3. Put another way, my Trash Can has a Function, but it is also Capable of being a door stop, drinking vessel, or weapon.

Is it aggregate aspect of an organization? Aggregates can bear multiple functions. An offensive has the function of blocking members from the opposing team and whose members each have slightly different functions (e.g., the center has the function to hike the ball). The offensive team (as a whole) has a function to move the ball into the end zone. This function of the offensive team is is some sense dependent upon the function of offensive line, which in turn is dependent on function of the individual members of the line. But, now, it seems, we are discussing relations amongst functions, and not necessarily a new category of things.

Perhaps put another way...

  1. An Organization is the 'bearer of' some Capability (e.g. an Offensive Football Capability)
  2. An Offensive Lineman 'has function' Blocking Function because he is the 'bearer of' some Blocking Skill
  3. The Offensive Football Capability in 1 is realized by all of the individual Functions/Skills of each Player.

If I was to represent your example in OWL, it would roughly (i.e., using pseudo syntax with time index ignored) look like this:

define classes

tank gunnery role subclass of role

Tank Gunner Role subclass of Person Role

tank gunnery skill subclass of function (alternative label "capability")

Tank Gunnery Skill subclass of Skill (Skill subclass of Disposition)

I don't think that a Skill is a Function...I have a set of Judo Skills, but to say they are Functions seems odd.

military organization subclass of organization organizational capability subclass of function (alternative label "capability")

Organizational Capability subclass of Capability

Capability subclass of Disposition

define instances

sgt jones instance of human being sgt jones gunner role instance of tank gunnery role sgt jones aiming skill instance of tank gunnery skill tank crew 1 instance of military organization tank crew 1 maneuver capability instance of organizational capability tank crew 1 engagement capability instance of organizational capability

relate instances

sgt jones bearer of tank gunnery role sgt jones bearer of sgt jones aiming skill sgt jones member of tank crew 1 tank crew 1 bear of tank crew 1 maneuver capability tank crew 1 bear of tank crew 1 engagement capability

Agreed

Question: What you trying to represent with the individual tank crew 1 capability? Is this an individual/instance or a general class?

That is supposed to be an instance of Tank Crew Capability

wdduncan commented 6 years ago

What you wrote would be fine if capability was better defined.

Here is Barry's definition of capability from the slides:

Capability = a disposition whose realization in the normal case brings benefits to an organism or group of organisms

As already discussed, something being a benefit is situational and difficult to pin down (e.g., I thought smoking was a benefit when I was teenager, but I now realize it was a detriment).

Can we be more precise with the formal axioms? Roles may help to address this b/c they come into being due to the bearer being is some situation/context. Although BFO won't allow for the capability to be directly related to a role (as far as I know), you can specify a role a necessary condition. Something along the lines of:

capability inheres in some (organism that bears some "benefit conferring role")

As for Judo being function or skill/capability, it depends on (I think) on how you define the terms 'artifact' and 'design'.

The definition of function (on ontobee) is:

elucidation: A function is a disposition that exists in virtue of the bearer’s physical make-up and this physical make-up is something the bearer possesses because it came into being, either through evolution (in the case of natural biological entities) or through intentional design (in the case of artifacts), in order to realize processes of a certain sort.

True, your Judo skill (most likely) didn't come into being through natural selection, but it did come into being through an intentional design process (e.g., training). Does the parenthetical statement "(in the case of artifacts)" mean that only artifacts can acquire functions through intentional design or is this just an example of intentional design? Are biological warfare agents artifacts? How about intentionally genetically modified food? My intuition is to say "no", but others may want to weigh in on this.

If we limit capabilities to only inhering in biological organisms, then function cannot be a subtype of capability, since functions inhere in non-biological entities. This would make capability a sibling of function instead of a parent.

BTW: Much of this discussion reminds of Dipert's book on artifacts.

wdduncan commented 6 years ago

​Function --> reason for existing It is no part of the reason for existing of your car that it can withstand ​a crash at 10mph or less without structural damage

Ok. But, why is it a capacity? My car has dispositions and structural qualities (e.g., the shape of its fame) that allow it to withstand such a a crash. Now it seems capacities also have quality aspect to them. Plus, it could very well be part of the manufacturing process that the car is withstand crashes at certain speeds.

The offense has the function to do the things you say It has the capability to play at temperatures down to -10 degrees Etc. etc.​

Do you have formal axioms for defining capacity?

Isn't this the equivalent as saying that the function of the offense can be realized by processes in which the participants of the process are located in a geographical region where the temperature is -10 degrees? Is using the label "capacity" just a shorter way of saying this? Again, I ask if capacity is best treated as a defined class?

I want to be clear about the intent of my questions: If we don't try to get really clear about these issues, capacities won't be used by the rest of the OBO community.