Integrate Analogous of "Quality" for "Processes"

[jmkeil]

According to the definition of quality and specifically dependent continuant, quality must be used to characterize independent continuants only (e.g. with a measurement datum that is about the quality). But I see nothing in BFO and IAO to describe a process in a similar way.

IOF defines a class ProcessCharacteristic (as subclass of occurrent) and the according properties hasProcessCharacteristic and processCharacteristicOf to fill this gap.

This is probably relevant not only for industry processes. Would it make sense to integrate this or something similar in IAO?

[alanruttenberg]

I've been wanting to add something like this to BFO directly, but have only discussed it internally. The idea is similar but divides process dependents (my label for what IOF calls process characteristic) into two classes, continuant process dependents, which are continuants, and occurrent process dependents which are occurrents. Continuant process dependents depend on a process but are also specifically dependent on some independent continuant. Examples are things like speeds and some other rates. Occurrent process occurrents depend on processes but are not specifically dependent, for example duration, or the rate of creation of some product. The examples of occurrent process dependents need to be better fleshed out. Could you point me at the IOF ontology that has classes under process characteristic?

Are there use cases for process characteristics specific to IAO that you have? I'd be interested in hearing them. Any thoughts on my proposal would be very welcome.

In the absence of process dependents, the pattern I've used for measurement data is to say they are about the process, which isn't as specific as it could be, but has worked well enough for me.

[jmkeil]

Thanks for your answer.

Could you point me at the IOF ontology that has classes under process characteristic?

Are there use cases for process characteristics specific to IAO that you have?

I need to represent measurements like the amount of some material per time used in a production process. My intention was to model it similar as in O3PO (according publication). The only other example I found that uses iof:ProcessCharacteristic is MESO, but without informative definitions it is (for me) not helpful to learn something about iof:ProcessCharacteristic.

The reason why I proposed it for IAO is, that it would enable the use of a measurement datum for a process in analogy to the already possible representation of a measurement datum for a continuant via a quality. But I currently do not have a use case to link a measurement datum to a information content entity. So IAO might not be right place.

Any thoughts on my proposal would be very welcome.

To be honest, I am not sure, if I really got your idea. But this are my thoughts about it:

To my understanding of IOFs process characteristic and what you call process dependent, process characteristic should be a subclass of process dependent. (My preferred name for process characteristic would be process quality in analogy to quality.)

The process dependent classes should probably be subclasses of some occurent dependent classes, to be more generic. That would result in the following hierarchy:

• entity
  • continuant
    • specifically dependent continuant
      • quality
        • relational quality
    • occurrent dependent continuant
      • process dependent continuant (Description: depend on a process but are also specifically dependent on some independent continuant)
        • continuant process quality
          • speed
  • occurrent
    • process
    • occurrent dependent occurrent
      • process dependent occurrent (Description: depend on processes but are not specifically dependent)
        • occurrent process quality
          • duration
          • creation rate

On which continuant does a speed depend on in addition to the process (the movement)? The moving object (e.g. a car, the moon - some participant of the process)? The object moving relatively to (e.g. the earth - some environment of the process)?

Does the dependency on a continuant make it an continuant?

In the absence of process dependents, the pattern I've used for measurement data is to say they are about the process, which isn't as specific as it could be, but has worked well enough for me.

How do you distinguish measurements of different quantity kinds in this approach?

[alanruttenberg]

Thanks for the feedback on the proposal. Helps me think it through and as you will see below there are parts that need to better.

I need to represent measurements like the amount of some material per time used in a production process.

I had a look at their use of process characteristic. My gut is that they are continuants, what I would call continuant process dependents. However they don't match the gloss I gave in that there's not something they are specifically dependent on, or at least not an obvious one. Maybe they are specifically dependent on some fiat portion of liquid. Or there's a different dependency relation. I've been chatting about what we nickname "oxygen dependence" which is a kind of generic dependence (but of a different sense than the relation for generically dependent continuants). The prototype oxygen dependence is that as organisms we all need oxygen but not the same oxygen molecules at all times. Similar to permanent generic parthood.

But the reason I would consider them continuants is that they (almost) "exist in full at any time they exist". That is, at any given moment there's a rate of flow and it can change moment to moment. Occurrents don't work that way. But I need to sharpen this.

The reason why I proposed it for IAO is, that it would enable the use of a measurement datum for a process in analogy to the already possible representation of a measurement datum for a continuant via a quality. But I currently do not have a use case to link a measurement datum to a information content entity. So IAO might not be right place.

As I said, I don't think it belongs here in IAO, but it belongs somewhere.

To my understanding of IOFs process characteristic and what you call process dependent, process characteristic should be a subclass of process dependent. (My preferred name for process characteristic would be process quality in analogy to quality.)

I don't think "quality" in the label will fly, based on discussions in the past with Barry. He's the final arbiter of labels. In any case it's only a label. I don't know what a process characteristic is - it's a primitive, so I won't venture to classify it.

On which continuant does a speed depend on in addition to the process (the movement)? The moving object (e.g. a car, the moon - some participant of the process)? The object moving relatively to (e.g. the earth - some environment of the process)?

The speed depends on the movement at least. Speed, generally, is tricky because of relativity. You can always choose a frame of reference in which something doesn't move. But for everyday use I'd make it depend on both the thing moving and the thing it is moving relative to. In my own work I consider positions, distances and similar to be relational qualities.

Does the dependency on a continuant make it an continuant?

No. What I think makes something a continuant(IMO) is that it (the rate) exists at each moment and has no temporal parts. There are a couple of exceptions to the temporal part criteria under occurrent. Instants don't have temporal parts, nor do process boundaries. But those things don't have identity over time. The rate of flow is like the weight of a baby. There's one particular that changes over time.

However, if it specifically depends on a continuant then it's a specifically dependent continuant and I expect that what I call continuant process dependents belong below specifically dependent continuant.

In the absence of process dependents, the pattern I've used for measurement data is to say they are about the process, which isn't as specific as it could be, but has worked well enough for me.

How do you distinguish measurements of different quantity kinds in this approach?

With different types of measurement datum. Again, not ideal but practical and it's common to do this sort of thing with continuants as well. For example in OBI 3D structural organization datumis about a cell, not about some quality of the cell. In fact I would say it is about a quality but what I'm pointing to here is the pattern, which is analogous to what I'm suggesting you do, at least in the interim.

So, you would have, in the case of flow, flow rate measurement datum, subclass of scalar measurement datum, and is about some liquid flowing process or something else related to the flow (a well for oil).scalar measurement datum doesn't have a requirement that it's about a quality. It's only the defined subclasses of it that are about qualities.

Or you could use ValueExpression and the relationship isValueExpressionOfAtSomeTime (also in o3po) which ultimately is a subclass of isAbout. The only thing I would say about that is that I don't like that it is a property of describes as I think there are cases where they are prescribed, like when a value denotes a setting.

The only change that would happen if we did have process dependents would be that you would say the flow rate measurement datum is about (or perhaps some more specific relation) the dependent and have the dependency relationships to get at the process.

[jmkeil]

Thanks for the explanation. Given a process characteristic is a specifically dependent continuant, that raises the question for me whether a process characteristic is not a realizable entity that depends on process participants (e.g. the car, the processing facility) and process environments (e.g. the earth) and is realized by the process (e.g. the movement, the production process).

What do you think about?

If modeled that way, properties for process characteristic would have obvious points to hook into the property hierarchy.