Definition of "plan specification" and subtype "study design"

[zhengj2007]

From MBrochhausen@gmail.com on May 22, 2011 11:12:05

I have some issues with plan specification (ps), namely the fact that the ps tells its bearer what to do. Actually I have no issues was that, but where do we put a kind of specification that tells people (which are not the bearer of the specification) how to act, namely a coding schema (an information artefact telling people how to code medical services, procedure and supplies). Regarding this problem, I was wondering whether "study design" really is a subtype to plan specification.

Any ideas?

Best, Mathias

Original issue: http://code.google.com/p/information-artifact-ontology/issues/detail?id=112

[zhengj2007]

From alanruttenberg@gmail.com on May 23, 2011 02:57:15

Should say tells the bearer of it's concretization what to do. That would be true of agents that bear concretizations though not for all. Needs work.

[zhengj2007]

From bjoern.p...@gmail.com on May 23, 2011 07:41:00

[I also send a 'reply all' email to the list, thinking that would end up in the tracker, but apparently it doesn't, so I am repeating that email here].

The current definition states is: "a directive information entity that when concretized it is realized in a process in which the bearer tries to achieve the objectives, in part by taking the actions specified. Plan specifications includes parts such as objective specification, action specifications and conditional specifications."

This already refers to 'bearer', and only concretizations of GDCs can have bearers. More specifically, it also says that the concretization can be realized, which implies that the specific concretization is a realizable entity (and not e.g. a document). This whole line of thought becomes a little clearer when looking at OBI rather than IAO. There, the intermediary class 'plan' is named, and the logical definitions of 'planned process' are present that are missing in IAO.

[zhengj2007]

From MBrochhausen@gmail.com on May 25, 2011 04:03:09

sorry, I still don't seem to get, although I agree with Bjoern that the way OBI gives it is slights more understandable and consistent.

But to me it seems that what is represented by this class is something like a computer program. The bearer of the program tries to achieve the the objectives. If you have a clinical study design the bearer most likely is a document (digital or not). The bearer does not try anything, but the agents managing the study (machine or human agents, both) do try to achieve the objectives specified.

Maybe what we need is to distinguish "directive information entity" which is always something like a program, service or tool from plans (directives) which have bearers, but are basically informing third parties about objectives. In that case directive information entity might be ill-named (but what are names in ontologies, right?). Furthermore, in that case study design cannot possibly be a subclass to directive information entity.

Best, Matt

[zhengj2007]

From bjoern.p...@gmail.com on May 25, 2011 07:38:33

Let's see if I can explain the thinking better (I would argue that this part of OBI / IAO could use a formal writeup as a small paper; it is hard to understand). I may also miss parts of your concern, but first I want to be sure we are on a common understanding what the current implementation is meant to mean:

There is NO bearer of a plan specification. A plan specification is an 'information content entity' (ICE). An ICE can be 'concretized' in several forms, for example the content of 'Moby Dick' is an instance of an ICE that is concretized in several instances of patterns of ink in books, electromagnetic patterns on hard disks and neuron patterns in human brains. All of these patterns are specifically dependent continuants that are borne by material entiteis. In contrast, the ICE is a generically dependent continuant (because destroying a single one of its concretizations does not change it).

A concretization of a plan specification can be written down, memorized and communicated like other ICEs. But there is a specific concretization, namely a 'plan' which is only in those bearers that not only know of the plan specification but have the intend to realize it. A plan is therefore a realizable entity (like role, disposition and function).

In the case of a complex study design the bearer of the corresponding plan is typically a group of people. These people can still rely on written concretizations of the study design as a common reference, or verbal communication between each other to e.g. instruct technicians, which is just ICE exchange as usual. But as people are carrying out planned processes that are part of the study design execution, they are bearer of (parts of) the concretization of the study design.

Does that make sense?

• Bjoern

[zhengj2007]

From MBrochhausen@gmail.com on May 25, 2011 08:27:56

Bjoern, thanks for your clarification which is helpful, because it pointed me to a mistake I made: I forgot that human minds can be bearers of ICEs.

Yet, some problems remain: Why can we say plan specifications have no bearers, when the definition states the opposite?

I think, by now I have the impression that your argumentation introduces a double bottom that looks somewhat superfluous from the point of view that BFO (and probably IAO, too) tend to be quite re-iterative with respect to ICE anyway (and rightly so).

You say (correct me if I am wrong) that of any plan specification exists in multiple concretizations in specific pattern (very different ones). These patterns (SDC opposed to our original plan specification which is a GDC) are now borne by different bearers, which are IC. These pattern look like something similar to a quality. Even though that might be, we need to be careful this quality like thing is not to be mistaken as a plan or plan specification. A specific pattern of ink on paper can be a study design in one language and a poem in another language.

Anyways, the definition as given in IAO (and maybe in OBI) two seems to be wrong, because given that we have , say, of my cancer study design. One concretizations is such that it is a pattern in my brain, whereas another concretization is such that it is a specific pattern of ink on a piece of paper. Both of these concretizations would, as far as I understand you, have bearers. One would be e, the other would be the sheet of paper in front of me. In case of the sheet of paper neither the plan specification, nor the concretized plan specification tell its bearer anything.

I like the idea of making a difference between the two concretizations above in saying the concretization my mind is a plan, but we must check which BFO whether this is a fourth subtype of realizable entity.

Given all that the definition in IAO rings wrong to me and since this is what people see we ought to work on optimizing it. I offer my assistance there.

Best, Matt

PS: Does any of this make sense?

[zhengj2007]

From bjoern.p...@gmail.com on May 25, 2011 08:57:07

In the textual definition:

"a directive information entity that when concretized it is realized in a process in which the bearer tries to ... "

the "bearer" is supposed to refer to the bearer of the concretization. I agree that this is confusing - mostly because we were aiming for short and readable english definitions, which should be read side-by-side with the logical ones (which are missing from IAO). So I would suggest to change it to

"a directive information entity that when concretized it is realized in a process in which the bearer of the concretization tries to ... "

I believe the 'plan' class, which is present in OBI and is a sibling of other realizable entities does what you are asking for: distinguish concretizations of plan specifications that are e.g. qualities of a piece of paper of the bearer from those that are realizables.

One thing this discussion is pointing out is that the split between IAO and OBI is problematic. IAO was originally only going to deal with pure ICEs, and therefore was not going to take planned process, plan and such things, all of which were originally developed in OBI.

One option would be to make a slim extension of BFO that would be imported by both IAO and OBI that would have only a minimal set of classes and relations including

• information content entity plan specification objective specification
• plan
• planned process

Alan had in the past been reluctant to do this because he still wanted flexibility to change definitions for ICE, but I think at this point it is time to bite that bullet.

[zhengj2007]

From mcour...@gmail.com on May 25, 2011 09:17:43

I think your original question arose from a maybe not very clear definition of "plan specification": "a directive information entity that when concretized it is realized in a process in which the bearer tries to achieve the objectives, in part by taking the actions specified. Plan specifications includes parts such as objective specification, action specifications and conditional specifications."

Here bearer means "the bearer of the concretization of the plan specification". This is probably clearer if we look at the parent class, directive information entity: "An information content entity whose concretizations indicate to their bearer how to realize them in a process."

Any directive information entity is concretized as an SDC.

This SDC then has bearers, and those are the bearers referred to in the definition (which could be made clearer)

Let us consider your example of cancer study. You are thinking about that study, making "plans" of how this should be done. You will probably write a grant, recruit patients, have an experimental protocol to follow, collect results, analyze those results, and then draw a conclusion. You will then summarize this in a publication.

This is all part of the plan specification, and all the elements described above are subparts: e.g., your objective specification may be "determine effect of drug X on malignant cell proliferation". Your study design would be "recruit 50 patients, obtain their consent, administer them a drug etc." Specific protocols would be "take 50ml of drug X, dilute to 1/10, and administer sub cutaneously, repeat 3h later"

At some point, those will get written down, and your plan specification is then concretized into its SDC, the plan. Your team of scientists will read and aim at execute your protocols, you will submit your proposal to an IRB etc... The bearer of the concretization of your plan specification is for example the scientist, which aims at performing tasks as described by the concretization of the plan specification. When you will write up a publication summarizing your results you will be a bearer etc.

[zhengj2007]

From mcour...@gmail.com on May 25, 2011 09:23:50

As OBI imports IAO, would it make sense to consolidate everything into IAO (i.e., your set of minimal classes and relations)? I agree that having plan in IAO would really help make things clearer. I also think we should document this better on the wiki: I started a page at https://code.google.com/p/information-artifact-ontology/wiki/DirectiveInformationEntity . Feel free to update.

[zhengj2007]

From MBrochhausen@gmail.com on May 25, 2011 09:25:50

Melanie,

thanks for summarizing what i think. I only slightly disagree with the last paragraph which seems to be a bit sloppy (if I may say so): The written down (thus concretized) plan specification is already an SDC which is borne by piece of paper. This bearer does nothing, regardless of the fact that there really exists bearers (my brain, your brain) that invest or trigger some action towards fulfilling the plan.

By now I am absolutely sure that a definition that needs so much explaining is suspicious of being wrong. Why mention the bearers of a completely different entity (concretization) anyway?

[zhengj2007]

From MBrochhausen@gmail.com on May 26, 2011 02:14:49

here is an afterthought:

what is the realizable entity a plan specification is about? That seems to be something different from the plan which (as I think) showed in my last e-mail (since the plan can be pattern of ink on a paper; correct me if I am wrong).

Best, Matt

[zhengj2007]

From bjoern.p...@gmail.com on May 26, 2011 17:16:57

A 'plan' can not be patterns of ink on a paper (I think Melanie didn't write that intentionally). It must be borne by an entity that intends to realize it.

There are two views what a plan specification is about. In my mind, an instance of a plan specification is about the class of planned processes (and participants in those processes) that realize its concretizations. Alan suggests that an instance of a plan specification is about instances rather than classes.

[zhengj2007]

From MBrochhausen@gmail.com on May 28, 2011 01:17:46

OK. I think I am with Alan here. But maybe let me give you my version and see what your reaction to that turns out to be: (before I start I need to say I am feeling extremely uncomfortable with the verb "concretize" and the term "concretization" since I am not really sure what this would be in BFO and I have a feeling we are creating unnecessary extra-entities.

S, there are plans which as Bjoern are borne by entities that intend to realize them. Thus, plans are good realizable entities (subclass to SDC). All plan specifications are about plans (all ICE needs to be about somthing, right?). Every instance of a plan specification can x (fill in whatever verb you like "inhere in", "be borne by" maybe even "concretize", whatever your preferred name for the relation between DCs and ICs you like) in different material entity (my brain, this sheet of paper, this audio tape). What is realized is the plan, not the plan specification. But the plan specification can transport my plan of how to seize world power to you, resulting in you having a similar but not same (SDC!!!) plan of how to seize world power).

What do you think?

[zhengj2007]

From alanruttenberg@gmail.com on March 10, 2014 10:21:06

After a discussion on an IAO/OBI call I wanted to point out a parallel issue with Objective Specification and Plan Specification. Both can be concretized as either quality or realizable entity (though at least one) and perhaps this dual concretization is an important differentia for the terms. When editing the terms both should be constructed on concert so that they are parallel.