Issue 511: Measurements and Dimensions

ID: 
511
Starting Date: 
2020-09-09
Working Group: 
3
Status: 
Done
Background: 

Posted by Robert Sanderson on 9/9/2020

Dear all,

I believe that there is an inconsistency in the model for measurements and dimensions.

E54 Dimensions are associated directly with E70 Things using P43 has dimension.  So not every class can have dimensions, only those that are descendents of E70.

However E16 Measurement's property P39 measured has a range of E1 CRM Entity, meaning that while (for example) an E53 Place cannot have a dimension, it can be measured to have a dimension. This seems inconsistent that an entity that cannot have dimensions can still be measured.

I propose that the range of P39 measured be changed to E70 Thing to resolve this inconsistency.

I would also be okay with the other direction by changing the domain of P43 has dimension to be E1 CRM Entity, however that seems like a much more significant change, and would result in quite strange side effects such as Dimensions having Dimensions.

Posted by Thanasis on 9/9/2020

Good point, but it seems to me that being able to measure a Place is pretty important. Otherwise we have to measure through the physical object/site reference or the declarative space as part of a conceptual thing.
 

Current Proposal: 

Posted by Robert on 9/9/2020

Thanks Thanasis.  Yes, there's various dimensions that are associated with non-Things, and I agree that Place is particularly easy to justify.

Place:  Area. The county of Los Angeles has a dimension of 4751 square miles. If the place is approximate, then the radius of a centroid would be an obvious dimension to record, or height/width for bounding box defined Places.

Time-Span:  Duration is already a property of a Time-Span that refers to a dimension (P191). This could then be a subproperty of P43, or deprecated in favor of a classification on the Dimension.

Temporal Entity and Spacetime Volume are a bit strange in relation to Time-Span. Does the Period have the duration or the Time-Span, or both? What if they're different

Conversely Dimensions seem like they should not have Dimensions.

Posted by Athina on 9/9/2020

As I remember, this problem was discussed in issues 229 and 307, which are declared closed. However, I am wondering, if it is related to the issue 293?

Posted by Robert on 9/9/2020

Oh, I'm sorry for duplicating your original issue 229 Athina! Great minds think alike, perhaps?

It seems like 229 was closed in favor of discussing it as part of 307, but that discussion didn't happen. 293 does seem like a good venue, but ultimately there is an inconsistency right now (as you recognized long ago!) that should not persist in version 7.0 regardless of whether ObservableEntity / Observation are moved into CRM Core or not.

Posted by Martin on 09/09/2020

On 9/9/2020 7:47 PM, Robert Sanderson wrote:

 

Oh, I'm sorry for duplicating your original issue 229 Athina! Great minds think alike, perhaps?

 

It seems like 229 was closed in favor of discussing it as part of 307, but that discussion didn't happen. 293 does seem like a good venue, but ultimately there is an inconsistency right now (as you recognized long ago!) that should not persist in version 7.0 regardless of whether ObservableEntity / Observation are moved into CRM Core or not.

I agree.

Posted by Martin on 09/09/2020

Indeed durations can be measured as distances between two events. Einstein regarded that length measurements are two simultaneous events matching both endpoints of a yardstick with something. This is the basis of the Theory of Relativity.

Basically we have to do with signals.

To be discussed!

Posted by Martin on 2/3/2021

Dear All,

Let me take up this issue, after new considerations:

Background:

Posted by Robert Sanderson on 9/9/2020

Dear all,

I believe that there is an inconsistency in the model for measurements and dimensions.

E54 Dimensions are associated directly with E70 Things using P43 has dimension.  So not every class can have dimensions, only those that are descendents of E70.

However E16 Measurement's property P39 measured has a range of E1 CRM Entity, meaning that while (for example) an E53 Place cannot have a dimension, it can be measured to have a dimension. This seems inconsistent that an entity that cannot have dimensions can still be measured.

I propose that the range of P39 measured be changed to E70 Thing to resolve this inconsistency.

I would also be okay with the other direction by changing the domain of P43 has dimension to be E1 CRM Entity, however that seems like a much more significant change, and would result in quite strange side effects such as Dimensions having Dimensions.

……

Posted by Robert on 9/9/2020

Thanks Thanasis.  Yes, there's various dimensions that are associated with non-Things, and I agree that Place is particularly easy to justify.

Place:  Area. The county of Los Angeles has a dimension of 4751 square miles. If the place is approximate, then the radius of a centroid would be an obvious dimension to record, or height/width for bounding box defined Places.

Time-Span:  Duration is already a property of a Time-Span that refers to a dimension (P191). This could then be a subproperty of P43, or deprecated in favor of a classification on the Dimension.

Temporal Entity and Spacetime Volume are a bit strange in relation to Time-Span. Does the Period have the duration or the Time-Span, or both? What if they're different

Conversely Dimensions seem like they should not have Dimensions.

 .......................

We have to distinguish measurement from dimension. In order to measure something in a narrower sense, I need an observation of something material. Dimensions can also be result of computation, evaluation and estimation (forms of Attribute Assignment).

If we look at measuring in the narrower sense, we can count the characters of a text on paper, but not the abstract text. The logical representation of a text can be evaluated for its dimensions.

We cannot measure a place, but features at a place. See also Issue 388. But clearly, we can measure duration and extent of processes, and comparing a clock, which provides a duration from the last sync event, with some other transient situation or microevent, in order to calculate absolute time.

So, we may assign the ability to be observed to E18 physical things and E4 Period, or more narrowly to E5 Event.  The ability to be observed appears to need some common ontological nature, a certain materiality interacting with measurable signals. Even the lightning creates a plasma hose lasting some milliseconds. That would need a new class “Observable Entity” as range.

Otherwise, we may regard measuring physical things and measuring processes as independent. Then, we would need another measurement class, such as “static measurement” versus “dynamic measurement”.

Dimensions of other things, such as places in the abstract geometric sense of the CRM, need not be based on a common property. The place can only have diameters and areas as dimentions, and may be some more exotic ones. The dimension in the phenomenal timespan is of course that of the respective period etc. So, my argument being that E53 Place, E52 Time-Span have their own properties with range Dimension, without being regarded as observable (rather results of observation).

I’d propose the following:

Reduce in CRMbase Mesaurement , P40 observed dimension, to E18 Physical Thing. Add 3 different properties “has dimension” in CRMBase to E70 Thing, E53 Place, E4 Period (or E2 Temp Entity).

Extent CRMSci by E18, E4 IsA Observable Entity, and extend Mesaurement P40 observed dimension,  from E18 to Observable Entity.

Alternatively, introduce “Dynamic Measurement”  in CRMSci.

Posted by Robert Sanderson on 2/3/2021

Martin wrote in particular:

  Reduce in CRMbase Mesaurement , P40 observed dimension, to E18 Physical Thing. Add 3 different properties “has dimension” in CRMBase to E70 Thing, E53 Place, E4 Period (or E2 Temp Entity).

I agree with your argumentation, and believe that the changes in CRM Base would be:

P39 measured:

  Range changes from E1 CRM Entity to E18 Physical Thing

PXX1_has_dimension 

  Domain: E53 Place

  Range: E54 Dimension

PXX2_has_dimension

  Domain: E4 Period

  Range: E54 Dimension 

to be cognate with P43 has dimension for E70s.

The question would remain about the measuring of Non-physical Things, such as the number of symbols in a E90 symbolic object... but I don't have that use case, so am happy to leave the discussion to someone that does :)

Posted by Martin on 03/03/2021

Dear Robert,

Yes, exactly.

My argument about measuring non-physical things is that it does not constitute an observation process, but an abstraction from observable things. We can always use Attribute Assignment for such evaluations.

So, we can assign the word count to a text, without using E16 Measurement.
 

Posted by Daria Hookk on 03/03/2021

It's always comparison with something, we fix difference.

Posted by Robert Sanderson on 03/03/2021

On Wed, Mar 3, 2021 at 5:54 AM Martin Doerr <martin@ics.forth.gr> wrote:

My argument about measuring non-physical things is that it does not constitute an observation process, but an abstraction from observable things. We can always use Attribute Assignment for such evaluations. So, we can assign the word count to a text, without using E16 Measurement.

Understood, and agreed. The scope note for E16 is clear that is for measuring "physical properties ... by ... direct observation of particular states". 

A word count would be an Attribute Assignment of the Dimension to the Linguistic Object, potentially using a particular specific object as a witness for the symbols. Of course, I can count symbols in my head, but then I am not observing the symbols physically, and therefore it is not a Measurement.

If I am not able to be at the SIG session where this is discussed, please count this as my vote in favor of the resolution of the issue.

 

In the 49th joint meeting of the CIDOC CRM SIG and SO/TC46/SC4/WG9; 42nd FRBR – CIDOC CRM Harmonization meeting, MD & RS argued in favor of reducing the range of P39 measured to E18 (from E1 that it used to be). The change was to be implemented in v7.1.1 (next official). 

Points raised:
* one cannot measure instances of E53 Place, or E2 Temporal Entity using P39 any more: It will be resolved by introducing properties like P43 has dimension but for places and periods (will not be implemented in v7.1.1). 
* there are also implications for CRMsci (i) S21 Measurement, (ii) S15 Observable Entity and (iii) the relation btw O24 measured & P39 measured, and O12 has dimension and P43 has dimension. 
* the change allows non-monotonic reasoning

Decision: 
- The proposal to reduce the range of P39 to E18 will be implemented in v7.1
- the scope note will be updated accordingly and put to an e-vote. 
- E16 Measurement needs to be checked as well (list of outgoing properties needs be updated f.i.). 
- The migration path from old to new version of P39 also needs to be updated. 
- The two new properties (assigning dimensions to places and temporal entities) are intended for CIDOC CRM v7.2. They might even not be made part of CRMbase and be introduced to CRMgeo (for places) or CRMsci instead. To be discussed in a new issue
- Implications for CRMsci to be discussed in a new issue

HW: RS & MD to edit the scope note of E16, P39, check FOL representations.

March 2021

[Crm-sig] 511 e-vote

Athanasios Velios via Crm-sig Fri, 19 Mar 2021 03:43:12 -0700

Dear all,

 

At the last session of the last CRM SIG meeting we discussed issue 511 and voted to accept the reduction of the range of property P39 measured from E1 CRM Entity to E18 Physical Thing. Homework was assigned to check how scope notes and related properties are affected, recommend changes and call an e-vote for those. I am listing the required changes below. With regards to those changes, the possible votes are:

 

 * Yes = accept/agree

 * No = do not accept/agree

 * Other = With other you can either introduce a caveat (e.g.: 'Yes,

  but there is a typo on word x, fix it.') or you can write VETO, if

  you wish to stop the proposed change from happening, in which case

  you should also write a justification and reformulate the issue

  (e.g.: 'VETO, this change is unacceptable because it violates the

  following principle...')

 

1. E16 Measurement

 

 

Changed to clarify that E16 Measurement requires observation, including an update to an example and the removal of two examples.

 

 

From

 

Subclass of: E13 Attribute Assignment

 

Scope note:

 

This class comprises actions measuring quantitative physical properties and other values that can be determined by a systematic, objective procedure of direct observation of particular states of physical reality. Properties of instances of E90 Symbolic Object may be measured by observing some of their representative carriers which may or may not be named explicitly. In the case that the carrier can be named, the property P16 used specific object (was used for): should be used to indicate the instance(s) of E18 Physical Thing that was used as the empirical basis for the measurement activity.

 

Examples include measuring the nominal monetary value of a collection of coins or the running time of a movie on a specific video cassette.

 

The E16 Measurement may use simple counting or tools, such as yardsticks or radiation detection devices. The interest is in the method and care applied, so that the reliability of the result may be judged at a later stage, or research continued on the associated documents. The date of the event is important for dimensions, which may change value over time, such as the length of an object subject to shrinkage. Methods and devices employed should be associated with instances of E16 Measurement by properties such as P33 used specific technique: E29 Design or Procedure, P125 used object of type: E55 Type, P16 used specific object (was used for): E70 Thing, whereas basic techniques such as "carbon 14 dating" should be encoded using P2 has type (is type of): E55 Type. Details of methods and devices reused or reusable in other instances of E16 Measurement should be documented for these entities rather than the measurements themselves, whereas details of particular execution may be documented by free text or by instantiating adequate sub-activities, if the detail may be of interest for an overarching query.

 

Regardless whether a measurement is made by an instrument or by human senses, it represents the initial transition from physical reality to information without any other documented information object in between within the reasoning chain that would represent the result of the interaction of the observer or device with reality. Therefore, inferring properties of depicted items using image material, such as satellite images, is not regarded as an instance of E16 Measurement, but as a subsequent instance of E13 Attribute Assignment. Rather, only the production of the images, understood as arrays of radiation intensities, is regarded as an instance of E16 Measurement. The same reasoning holds for other sensor data.

 

Examples:

 

  • measurement of height of silver cup 232 on the 31st August 1997 (fictitious)
  • the carbon 14 dating of the “Schoeninger Speer II” in 1996 [an about 400.000 year old complete Old Palaeolithic wooden spear found in Schoeningen, Niedersachsen, Germany in 1995] (Kouwenhoven, 1997)
  • The pixel size of the jpeg version of Titian’s painting Bacchus and Ariadne from 1520–3, as freely downloadable from the National Gallery in London’s web page <https://www.nationalgallery.org.uk/paintings/titian-bacchus-and-ariadne> is 581600 pixels.
  • The scope note of E21 Person in the Definition of the CIDOC Conceptual Reference Model Version 5.0.4 as downloaded from <http://www.cidoc-crm.org/sites/default/files/cidoc_crm_version_5.0.4.pdf> consists of 77 words.

 

 

In First Order Logic:

 

E16(x) ⇒ E13(x)

 

Properties:

 

P39 measured (was measured by): E1 CRM Entity

P40 observed dimension (was observed in): E54 Dimension

 

 

To:

 

 

Subclass of: E13 Attribute Assignment

 

Scope note:

 

This class comprises actions measuring quantitative physical properties that can be determined by a systematic, objective procedure of direct observation of particular states of physical reality.

 

An instance of E16 Measurement may use simple counting or tools, such as yardsticks or radiation detection devices. The interest is in the method and care applied, so that the reliability of the result may be judged at a later stage, or research continued from the associated documents. The date of the event is important for dimensions, which may change value over time, such as the length of an object subject to shrinkage. Methods and devices employed should be associated with instances of E16 Measurement by properties such as P33 used specific technique: E29 Design or Procedure, P125 used object of type: E55 Type, P16 used specific object (was used for): E70 Thing, whereas basic techniques such as "carbon 14 dating" should be encoded using P2 has type (is type of): E55 Type. Details of methods and devices reused or reusable in other instances of E16 Measurement should be documented for these entities rather than the measurements themselves, whereas details of particular execution may be documented by free text or by instantiating adequate sub-activities, if the detail may be of interest for an overarching query.

 

Regardless whether a measurement is made by an instrument or by human senses, it represents the initial transition from physical reality to information without any other documented information object in the reasoning chain that would represent the result of the interaction of the observer or device with reality. Therefore, determining properties of an instance of E90 Symbolic Object is regarded as an instance of E13 Attribute Assignment, which may be inferred from observing and measuring representative carriers. In the case that the carrier can be named, the property P16 used specific object (was used for): should be used to indicate the instance(s) of E18 Physical Thing that was used as the empirical basis for the attribute assignment. For instance, inferring properties of depicted items using image material, such as satellite images, is not regarded as an instance of E16 Measurement, but as a subsequent instance of E13 Attribute Assignment. Rather, only the production of the images, understood as arrays of radiation intensities, is regarded as an instance of E16 Measurement. The same reasoning holds for other sensor data.

 

Examples:

 

  • measurement of height of silver cup 232 on the 31st August 1997 (fictitious)
  • the carbon 14 dating of the “Schoeninger Speer II” in 1996 [The carbon 14 dating of an approximately 400.000 year old complete Old Palaeolithic wooden spear found in Schoeningen, Niedersachsen, Germany in 1995] (Kouwenhoven, 1997)

 

 

In First Order Logic:

 

E16(x) ⇒ E13(x)

 

Properties:

 

P39 measured (was measured by): E18 Physical Thing

P40 observed dimension (was observed in): E54 Dimension

 

2. P39 measured (was measured by)

 

 

Changed to reduce the range to E18, update the scope note and the second example.

 

 

From

Domain: E16 Measurement

 

Range: E1 CRM Entity

 

Subproperty of: E13 Attribute Assignment. P140 assigned attribute to (was attributed by): E1 CRM Entity

 

Quantification:

 

many to one, necessary (1,1:0,n)

 

Scope note:

 

This property associates an instance of E16 Measurement with the instance of E1 CRM Entity to which it applied. An instance of E1 CRM Entity may be measured more than once. Material and immaterial things and processes may be measured, e.g., the number of words in a text, or the duration of an event.

 

Examples:

 

  • 31 August 1997 measurement of height of silver cup 232 (E16) measured silver cup 232 (E22) (fictitious)

 

 

In First Order Logic:

 

P39(x,y) ⇒ E16(x)

 

P39(x,y) ⇒ E1(y)

 

P39(x,y) ⇒ P140(x,y)

 

 

To:

 

 

Domain: E16 Measurement

 

Range: E18 Physical Thing

 

Subproperty of: E13 Attribute Assignment. P140 assigned attribute to (was attributed by): E1 CRM Entity

 

Quantification:many to one, necessary (1,1:0,n)

 

Scope note:

 

This property associates an instance of E16 Measurement with the instance of E18 Physical Thing upon which it acted. The instance of E16 Measurement is specific to the measured object. An instance of E18 Physical Thing may be measured more than once with different results, constituting different instances of E16 Measurement.

 

Examples:

 

  • 31 August 1997 measurement of height of silver cup 232 (E16) measured silver cup 232 (E22) (fictitious)
  • the carbon 14 dating of the “Schoeninger Speer II” in 1996 (E16) measured the “Schoeninger Speer II” (E22) [The carbon 14 dating of an approximately 400.000 year old complete Old Palaeolithic wooden spear found in Schoeningen, Niedersachsen, Germany in 1995. See also
  • E16 Measurement] (Kouwenhoven, 1997)

 

 

 

In First Order Logic:

 

P39(x,y) ⇒ E16(x)

P39(x,y) ⇒ E18(y)

P39(x,y) ⇒ P140(x,y)

 

 

 

 3. P140 assigned attribute to (was attributed by)

 

 

Only updated the reference P39 measured as a sub-property of P140.

 

 

From:

 

 

Superproperty of:  E14  Condition Assessment. P34 concerned (was assessed by): E18 Physical Thing

E16 Measurement. P39 measured (was measured by): E1 CRM Entity

E17 Type Assignment. P41 classified (was classified by): E1 CRM Entity

 

 

To:

 

 

Superproperty of:

 

E14  Condition Assessment. P34 concerned (was assessed by): E18 Physical Thing

E16 Measurement. P39 measured (was measured by): E18 Physical Thing

E17 Type Assignment. P41 classified (was classified by): E1 CRM Entity

 

 

 

4. P40 observed dimension (was observed in)

 

Added a second example:

 

  • the carbon 14 dating of the “Schoeninger Speer II” in 1996 (E16) observed dimension the carbon 14 based temporal distance from 1996 to the growth of the wood of the “Schoeninger Speer II” [The carbon 14 dating of an approximately 400.000 year old complete Old Palaeolithic wooden spear found in Schoeningen, Niedersachsen, Germany in 1995. See also E16 Measurement] (Kouwenhoven, 1997)

 

 

5. P43 has dimension (is dimension of)

 

Modified scope note to indicate that the long path through P39i was measured by applies to E18 Physical Thing and not E70 Thing (note that this change does not affect the domain of P43).

 

 

From:

 

 

 

Domain: E70  Thing

Range: E54 Dimension

 

Quantification: one to many, dependent (0,n:1,1)

 

Scope note:

 

This property records a E54 Dimension of some E70 Thing.

 

This property is a shortcut of the more fully developed path from E70 Thing through P39i was measured by, E16 Measurement, P40 observed dimension, to E54 Dimension. It offers no information about how and when an E54 Dimension was established, nor by whom.

 

An instance of E54 Dimension is specific to an instance of E70 Thing.

 

Examples:

 

  • silver cup 232 (E22) has dimension height of silver cup 232 (E54) has unit (P91) mm (E58), has value (P90) 224 (E60) (fictitious)

 

 

In First Order Logic:

 

P43(x,y) ⇒ E70(x)

P43(x,y) ⇒ E54(y)

P43(x,y) ⇐ (∃z) [E16(z) ˄ P39i(x,z) ˄ P40(z,y)]

 

 

To

 

 

Domain: E70 Thing

 

Range: E54 Dimension

 

Quantification: one to many, dependent (0,n:1,1)

 

Scope note:

 

This property records a E54 Dimension of some E70 Thing.

 

It offers no information about how and when an instance of E54 Dimension was established, nor by whom. In the case that the recorded property is a result of a measurement of an instance of E18 Physical Thing, this property is a shortcut of the more fully developed path from E18 Physical Thing through P39i was measured by, E16 Measurement, P40 observed dimension, to E54 Dimension. It offers no information about how and when an E54 Dimension was established, nor by whom. Knowledge about an instance of E54 Dimension need not be result of a measurement; it may be the result of evaluating data or other information, which should be documented as an instance of E13 Attribute Assignment.

 

An instance of E54 Dimension is specific to an instance of E70 Thing.

 

Examples:

 

 * silver cup 232 (E22) has dimension height of silver cup 232 (E54)

  has unit (P91) mm (E58), has value (P90) 224 (E60) (fictitious)

 

 

In First Order Logic:

 

P43(x,y) ⇒ E70(x)

 

P43(x,y) ⇒ E54(y)

 

P43(x,y) ⇐ (∃z) [E16(z) ˄ P39i(x,z) ˄ P40(z,y)]

 

 

Please send your e-votes by the 26th of March.

 

 

All the best,

 

 

Thanasis

 

 

Bottom of Form

Øyvind Eide via Crm-sig Tue, 23 Mar 2021 03:55:30 -0700

YES

All the best,

Øyvind

Christian-Emil Smith Ore via Crm-sig Tue, 23 Mar 2021 04:36:31 -0700

​Yes

Chr-Emil Ore

Weiss Christian SNM via Crm-sig Tue, 23 Mar 2021 05:28:23 -0700

Yes!

Christian

Hiebel, Gerald via Crm-sig Tue, 23 Mar 2021 05:43:06 -0700

YES,
Best,
Gerald

Robert Sanderson via Crm-sig Tue, 23 Mar 2021 06:14:03 -0700

YES

Thanks everyone!

Francesco Beretta via Crm-sig Tue, 23 Mar 2021 06:29:58 -0700

Dear all,

as already stated in the SIG meeting, I'm concerned with monotonicity, and more largely with substantially changing the substance of a class without changing its identifier: E16 remains E16 but "measuring the nominal monetary value of a collection of coins" is now _excluded_.

So what about all project's using E16 for that ? Not to mention the surface of Places as geometries and so many projects using E53 Place for representing a geographical place ? The surface of a place cannot be measured ?

Issue 511 starts from a useful consistency check :  "E54 Dimensions are associated directly with E70 Things using P43 has dimension.  So not every class can have dimensions, only those that are descendents of E70. However E16 Measurement's property P39 measured has a range of E1 CRM Entity, meaning that while (for example) an E53 Place cannot have a dimension, it can be measured to have a dimension. This seems inconsistent that an entity that cannot have dimensions can still be measured. I propose that the range of P39 measured be changed to E70 Thing to resolve this inconsistency."

Because of this argument : "My argument about measuring non-physical things is that it does not constitute an observation process, but an abstraction from observable things. We can always use Attribute Assignment for such evaluations. So, we can assign the word count to a text, without using E16 Measurement."

after a quite short discussion (in proportion to the relevance of the issue) we vote about the restriction of this same class to a quite different substance than the long period one.

Excluding, e.g. the monetary value of an entity, which is purely abstract.

My argument was rebutted in the SIG saying the replacement is Attribute Assignment and algorithms can do the job in the data. I partly agree but it seems to me that, given the radical change of substance, the consistency of the information produced before version 7.??? will be lost.

So why then not create a new class, with a new ID and a new substance, restricted in the mentioned sense, and deprecate E16 if wished but leaving it as is for the sake of consistency of legacy information and monotonicity ?

Given these arguments, I vote:

VETO.

All the best

Francesco

Christian-Emil Smith Ore via Crm-sig Tue, 23 Mar 2021 08:10:56 -0700

Hm, Your veto  has the consequence that the issue has to be postponed to the 
next physical(that is electronical) meeting. This leaves us in an interesting 
situation.

​"At the last session of the last CRM SIG meeting we discussed issue 511 and 
voted to accept the reduction of the range of property P39 measured from E1 CRM 
Entity to E18 Physical Thing. "

So this is not covered by the veto.  What is postponed is not the decision to 
reduce the  range  of P39 measured but the evaluation of the

" Homework [...] to check how scope notes and related properties are affected, 
recommend changes and call an e-vote for those. "

Best,

Christian-Emil

Franco Niccolucci via Crm-sig Tue, 23 Mar 2021 08:15:54 -0700

I strongly agree with Francesco. 

Some time ago I wrote a paper about reliability assessment, suggesting that it 
could be considered a sort of measurement, perhaps a subclass of E16. It was 
not a proposal for the CRM SIG, just considerations about the fact that at 
present there is no satisfactory attribute to qualify reliability. This issue 
is very important for example to re-use data according to the R part of the 
FAIR principles. Such “attribute” needs to be machine-actionable and available 
for composition if data undergo several re-use passages: in sum, it has to be 
quantitative and preferably numeric, in the broad scope of E60 Number. Thus if 
one re-processes data that are 80% reliable also the results will be 80% 
reliable, in the best case, or less. Another processing with an error of 70% 
will produce new data reliable at 56%. I remember a famous project where 
archive documents were digitized, OCR-ed and then mined with NLP: nobody ever 
calculated the reliability of the final results. 

So the topic is not irrelevant, nor just a matter of gut feeling.

This concept was, indeed, an Attribute Assignment - quite obviously as it is 
its superclass so every measurement is an attribute assignment. But what should 
be pursued, in my opinion, is the right balance between (i) the proliferation 
of classes and properties and (ii) an excessive generalization. I am a bit 
scared by (i) 200 properties as well as by (ii) assigning too wide roles to 
very general entities. 

Occam stated "Entia non sunt multiplicanda praeter necessitatem” and this in my 
opinion might possibly apply to the 200 properties. I would add that also 
unnecessary (praeter necessitatem) simplification should be better avoided, 
like in this case putting everything into a generic attribute assignment except 
some privileged activities that have a name (and a life) of their own.

A more philosophical argument would consider that there is no such thing as 
“(pure) observation” as opposed to "abstraction": any observation is influenced 
by the observer. Besides Heisenberg's indetermination principle, it is a much 
debated issue. The Galilean method, also known as the scientific one, and the 
historical method, both converge and are to be used in cultural heritage. So 
the justification that measurement is observation does not simplify, rather it 
complicates the decision.

Regards

Franco

By the way, for those who don’t know Latin, Occam’s razor means: "Entities 
should not be multiplied beyond necessity”.

Martin Doerr via Crm-sig Tue, 23 Mar 2021 11:11:37 -0700

Dear Franco,

Now I am fairly confused about with what you actually agree:

You write:

"This concept was, indeed, an Attribute Assignment - quite obviously as it is its superclass so every measurement is an attribute assignment. But what should be pursued, in my opinion, is the right balance between (i) the proliferation of classes and properties and (ii) an excessive generalization. I am a bit scared by (i) 200 properties as well as by (ii) assigning too wide roles to very general entities. "

Whereas Francesco argues that for *reasons of monotonicity* the range of P39 should have been remained at E1 CRM Entity, and that even the scope note of E16 Measurement always defined measurement as a result of observation without any doubt, the range of P39 should not conform with the scope note:

The old scope note begins:
"This class comprises actions measuring quantitative physical properties and other values that can be determined by a systematic, objective procedure of direct observation of particular states of physical reality."

You argue that the range of P39 to be E1 CRM Entity is NOT an excsessive generalization, whereas using E13 for quantitative properties of abstract items is an excessive generalization?

Your following statement finds my complete agreement:

"A more philosophical argument would consider that there is no such thing as “(pure) observation” as opposed to "abstraction": any observation is influenced by the observer. Besides Heisenberg's indetermination principle, it is a much debated issue. The Galilean method, also known as the scientific one, and the historical method, both converge and are to be used in cultural heritage. "

If I am not completely mistaken, you make a coimplication here. Whereas the scope note says that observation is *necessary*, you argue that pure observation is *not sufficient* for a Measurement.

Of course any "systematic objective procedure" is a product of hypothesis and its application. No current scientist would be so naive any more. See also "scientific realism" by James Ladyman and others.

I am a bit confused when you make arguments here about missing reliability assessment for questioning observation:

Never in my studies and research in physics I have done, seen or heard of a measurement without reliability assessment. The scientific method of observation that any scientist learns at the university , the "systematic objective procedure", can only be objective in terms of a reliability assessment and error margins. Nothing in the scope note of E16 implies that the results are 100% precise and true. Reliability can not be seperated from result.

By the way, we should always be careful citing quantum mechanics. Heisenberg's indetermination principle and the following quantum mechanical laws very precisely, numerically, define *how* the observer influences the observed system. This is not questioning observation as a method. It clarifies what you observe.

You write:

"So the justification that measurement is observation does not simplify, rather it complicates the decision."

Neither did the scope note of E16 ever state that "measurement *is* observation", nor was that the subject of the requested e-vote.

Let me add however, that

  (i) We model bottom up, and the CRMbase is not the place for a complete theory of measurement.

  (ii) Therefore the whole issue 511 was very conservative to be consistent with the intended scope of E16, as long as we have no better, consistent formulation.

  (iii) As long as we have no better consistent specific formulation, we always resort to the next superclass, which is E13 so far. This is not overloading E13. It is "epistemic humility" about E16.

  (iv) We have connected Issue 511 in the last SIG to producing a much more comprehensive representation of measurement.

We will be pleased to come back to the reliability assessment. Currently, error margins are foreseen for P90 has value.

As usual, I have the feeling that in reality, objectively, we agree???

All the best,

Martin

Franco Niccolucci via Crm-sig Tue, 23 Mar 2021 12:43:51 -0700

and you may have noticed that I did not vote - I just wanted to stimulate 
reflection, and I will not bore you anymore. I will reply Martin directly.

Franco
 

Martin Doerr via Crm-sig Tue, 23 Mar 2021 11:22:01 -0700

Dear Francesco,

Your concerns well respected, please let me explain a few things:

Firstly, this e-vote is not about the reduction of the range of P39 from E1 CRM Entity to E18 Physical Thing.

The reduction was decided in the last CRM-SIG with good majority after considering all pros and cons.

Following our rules, a decision once made by the CRM-SIG can only be undone by raising a new issue, providing new additional arguments.

Therefore, the use of the VETO right should not be used to undo an orderly decision taken by the SIG.

Francesco Beretta via Crm-sig Tue, 23 Mar 2021 11:57:52 -0700

Dear Martin, Christian-Emil, all,

In order not to block a development that seems to be largely consensual, and considering that my veto apparently violates the SIG rules, I withdraw it and simply vote NO.

If the majority thinks that the problem I have pointed to is not a problem and that the inconsistency between previous versions of the CRM and the new one in relation to the substance of this class is not an issue, especially with respect to monotonicity, I personnally have nothing to add.

With all my best wishes

Francesco

Christian-Emil Smith Ore via Crm-sig Tue, 23 Mar 2021 13:18:57 -0700

Dear Francesco, 
You did not violate any rule. My point was simply that the reduction has 
already been accepted. The question was about an adjustment of the scope notes 
so that they were in harmony with the reduced range. 
I had some concern about the reduction but ended with the conclusion that it is 
only physical matter that can be measured in the way the CRM measurement is 
defined. 

We also had a discussion in the chat in the meeting about measuring phenomena 
in a text, the example was the frequency of the  definite article 'the' in a 
work of Shakespeare. I will claim that the frequency is measured by counting 
the number of the occurrences of the three letter sequence with  white space or 
delimiters on each side in a given manuscript or printed text. One may also 
count the number of  'the' when listening to a oral performance of a work, 
still it is a measurement of a physical thing, the sound. From such measurement 
on may deduce and attach properties to the abstract object.  The question is if 
one can directly measure dimensions of non-physical things like temporal 
entities and abstracts - may be.  May be such measurement will be instances of 
a class of human subjective decisions and thus instances of attribute 
assignment?

Best,
Christian-Emil

Martin Doerr via Crm-sig Tue, 23 Mar 2021 14:00:00 -0700

Der Christian-Emil, all,

I agree in all you say below.

I'd like to mention that we have the issue about position measurements, and data evaluation as inference making in CRMinf. I'd suggest to see these things together for a wider theory.

Best,

Martin

Martin Doerr via Crm-sig Tue, 23 Mar 2021 13:54:39 -0700

Dear Francesco,

Thank you very much for reconsidering and withdrawing your veto. By mistake, I send my message before finishing it.

Please let me be more analytical about the arguments.

Your arguments in the SIG meeting have been well understood and well respected, and been evaluated against the alternatives and *by no means *ignored or *regarded as irrelevant*.

Summarizing the arguments pro:

1) Even though the range of P39 until 7.1 was E1 CRM Entity, it should never has been used in CRM applications for things other than instances of E70 Thing. Therefore, the range of P39 should be restricted at least to E70 Thing.

2) According to the scope note of E16, a Measurement is the result of observing a physical thing.

   The old and new scope note begins:

"This class comprises actions measuring quantitative physical properties and other values that can be determined by a systematic, objective procedure of direct observation of particular states of physical reality."

3) Determining or inferring instances of P43 has dimension can be the result of different kinds of processes.

For instances of E28 Conceptual Objects, these processes are not E16 Measurement, but evaluation of results, as *clearly stated in the old scope note*:

"Properties of instances of E90 Symbolic Object may be measured by observing some of their representative *carriers*...." and "Regardless whether a measurement is made by an instrument or by human senses, it *represents the initial transition from physical reality to information without any other documented information object in between* within the reasoning chain that would represent the result of the interaction of the observer or device with reality. Therefore, inferring properties of depicted items using image material, such as satellite images, *is not regarded as an instance of E16 Measurement*, but as a subsequent instance of *E13 Attribute Assignment.*"

4) Since the processes for inferring instances of P43 has dimension can be documented using E13 Attribute Assignment, it is not necessary that the range of P39 includes non-physical things in order to document how a dimension of an instance of E70 Thing was found. This is already described in the old scope note.

Of *paramount importance* is the interpretation that a measurement implies the *physical presence *of an object of evidence. Physical presence is one of the most fundamental reasoning processes in the CRM, which must not be abandoned or confused. In how far results are repeatable, precise etc. is *all *secondary to the fact that a present physical thing has empirically be evaluated.

Therefore, the *e-vote is  about the consistency *of the correction of the scope notes with the basic meaning of the old scope note of E16, and the correct propagation of all ramifications of the already decided reduction of the range of P39.

5) Since in version 7.1 we do, for good reasons, no more require all properties of compatible extensions to be subproperties of CRMbase properties, S21 Measurement in CRMsci needs no more be subclass of E16 Measurement for formal reasons. The *inadequate range of P39* however *prevented* developing adequate generalization of E16 Measurement in CRMsci. The decision to reduce P39 to E18 Physical Thing in the ISO standard to come *enables* CRMsci to be developed as it should. Not doing it, *would have blocked CRMsci* for a decade.

Please allow me to answer below your statements:

as already stated in the SIG meeting, I'm concerned with monotonicity, and more largely with substantially changing the substance of a class without changing its identifier: E16 remains E16 but "measuring the nominal monetary value of a collection of coins" is now _excluded_.

As stated above by citing the old scope note, I kindly ask you to consider that we have good reasons not to regard the decision as "substantially changing the substance" of E16, but as respecting the very substance, in contrast to border cases.

It is not true that "measuring the nominal monetary value of a collection of coins" is now _excluded_. It is true that it no more explicitly meantioned as an important application. It was deleted because it is amgiguous about the evaluation method, and therefore regarded as not particularly useful.

The paramount application of E16 is conservation technology in museum, monuments, and archaeometry, not counting pixels of images or money. Would you indeed disagree?

So what about all project's using E16 for that ? Not to mention the surface of Places as geometries and so many projects using E53 Place for representing a geographical place ? The surface of a place cannot be measured ?

If you have followed e-mail discussions last year, we discussed that the surface of a Physical Feature, including settlements etc., can quite well be measured with the new model and *falls under E18 Physical Thing*.

In 7.1, E53 Place does not have P43 has dimension anyway, because E53 Place is not subclass of E70 Thing.

You may have missed in the last meeting that the assignment of dimensions to E53 Place was *decided *as a *new issue* for edition 7.2, because it needs more thinking.

Issue 511 starts from a useful consistency check :  "E54 Dimensions are associated directly with E70 Things using P43 has dimension.  So not every class can have dimensions, only those that are descendents of E70.

However E16 Measurement's property P39 measured has a range of E1 CRM Entity, meaning that while (for example) an E53 Place cannot have a dimension, it can be measured to have a dimension. This seems inconsistent that an entity that cannot have dimensions can still be measured. I propose that the range of P39 measured be changed to E70 Thing to resolve this inconsistency."

 

Because of this argument : "My argument about measuring non-physical things is that it does not constitute an observation process, but an abstraction from observable things. We can always use Attribute Assignment for such evaluations. So, we can assign the word count to a text, without using E16 Measurement."

 

after a quite short discussion (in proportion to the relevance of the issue) we vote about the restriction of this same class to a quite different substance than the long period one.

Excluding, e.g. the monetary value of an entity, which is purely abstract.

My argument was rebutted in the SIG saying the replacement is Attribute Assignment and algorithms can do the job in the data. I partly agree but it seems to me that, given the radical change of substance, the consistency of the information produced before version 7.??? will be lost.

This is confusing the decision. The class is not restricted to a radical change of substance, but reduced in scope. If you disagree, please make constructive arguments to the above. Please consult those who do measurements in their daily practice.

It is simply logically wrong that information produced before version 7.??? will be lost. The migration path provided is definitely loss-free from a technical point of view.

So why then not create a new class, with a new ID and a new substance, restricted in the mentioned sense, and deprecate E16 if wished but leaving it as is for the sake of consistency of legacy information and monotonicity ?

This argument has been understood in the SIG meeting. It would however be a new issue, not 511. I hope you are aware that you require priority of counting coins and words over the whole discipline of conservation technology and archaeometry, which measues as described in the scope note of E16. Is that really what you advocate for?

Concluding, CRM SIG takes the issue of monotonicity *utterly serious.

*You may have missed the argument in the SIG discussion that we have evaluated:

The "con" of a backwards incompatibility before going to ISO, of applications which are *not core*, with a *loss-free* migration path, and together with *another set* of migration instructions.

against to "pro" of a consistent model of measurement, in scope notes examples and properties, which will enable the development of a scientifically correct wider model of measurement in CRMsci.

and against delaying the decision of a monotonicity break for a time when the new ISO standard will be underway.

I hope I could clarify with the above my and CRM-SIG's deep respect for all arguments and absolute sincerity to evaluate all arguments on a rational, comprehensible basis, including yours you described in the meeting and repeated above.

All the best,

Martin

**

 

Francesco Beretta via Crm-sig Wed, 24 Mar 2021 01:12:40 -0700

Dear Martin,

It is to me to thank you so much for taking the time to restate the whole issue from the point of view of the methodology used to develop the CRM, making very clear the reasons for the decision that was taken by the SIG. I understand your arguments and they are certainly robust in the point of view of the scientific method, and more specifically physics.

My concern is about the adoption of the CIDOC CRM by a larger community, making it a widely used ontology in the domain of digital humanities and social sciences, which are interconnected with museums and other GLAM actors, all of them being active in the domain of cultural heritage, for conservation or for scientific exploration.

Restricting the scope of the E16 Measurement class after years of use in a more generic sense, whether or not it is the substance intended by the insiders from the beginning, poses in my opinion a problem which I think should be analysed from this larger point of view. More in general the methodology the SIG adopts to develop the CRM should in my opinion take into account this development towards a larger community of users and the need expressed by more and more people and projects to have an ontology to fill the interoperability gap in cultural heritage domain intended in a broad, multi-disciplinary sense.

These considerations also apply to the articulation between the different extensions in the CRM family, in which the approaches of the different disciplines can be expressed more clearly, and refer to wide spread standards in the different domains, leaving CRMbase with a generic character that makes it the pivot of interoperability. It should be noted that the growing demand is not only for data integration but also, and especially, for new information production.

I personally advocate the adoption of CIDOC CRM in the context of this broader vision, and often seek to make clear to external users methodological choices taken by the CRM SIG, and the meaning of aspects of the ontology that are less directly understandable, without specific training, in communities wishing to adopt it. And I believe, especially when I compare our activity with that of the maintainers of other standards, such as TEI, that a more community driven approach in the development of the CRM would facilitate its understanding and adoption, and thus further enhance the value of all the work that has been done over the years to develop the robust model that we know and cherish.

With all my thanks and best wishes

Francesco

George Bruseker via Crm-sig Thu, 25 Mar 2021 05:47:26 -0700

Dear all,

I also think that the decision was taken too fast and introduced at a final
moment as a fait accompli. Regardless of the ontological purity behind the
decision, there is a community of use to take into account. I indicated
also a no vote to adopting this measure in the meeting for this reason. I
understood that the acceptance of going to an evote would be to try to
build the concensus, not indicate that it is a realized decision and we can
only change the details.

It is no hill I want to die on, but I think given its practical
ramifications to the user community it should have either been given more
air time earlier in order to build concensus and consider alternatives or
have been left as an issue to resolve for another day.

Best,

George

Martin Doerr via Crm-sig Sat, 27 Mar 2021 09:44:55 -0700

Dear All,

With this opportunity, and in response to George's message below, I'd like to repeat the decision taking rules CRM-SIG follows since early days. The early documents regulating our decisions may have been lost in the transition to the new Website. Therefore I raise a new issue to collect or reformulate these rules and publish them on the Website under "Activities" with a title such as "How we operate".

The rule is that an issue to be decided in a SIG meeting, it must have been published before the meeting in a "YES or NO" decidable form.

In the meeting, the justifications are presented and discussed.

If the justifications are not regarded as inconsistent, or insufficient (in particular new counter insight presented), the issue should be decided.

No decision is undone under the same issue. In order to question an issue, a new issue with sufficient further evidence should be raised.

This was introduced for good reasons (after a meeting in which the same issue was done and undone several times, and in the end the decision was just the arbitrary iteration when the session ended). It is a major element guaranteeing transparency and stability of the CRM.

The decision criteria are majority, *not 100% consensus*. The latter would leave us in the worst case unable to finish work. The explicit goal is the *widest possible consensus*, taking small majority votes as an indication of an unmature issue to be rejected.

In the past, we had, successfully and without opposition, applied *time-outs* for all discussions (5 to 15 min by clock), because the work-load is immense, and members are expected to read the submissions before the meeting.

The time-out has not been applied since several years, rather because the a-priori estimation of discussion time needed is itself very time-consuming. As a result, our decision processes have been slowed down, creating an *increasing backlog* of unresolved issues, and raising *severe concerns *in the community by which rules the actual issues for discussion are selected. On the other side, managing the harmonization and consistency of an *increasing number of extensions* requires more and *more efficient decision taking.**
*
Therefore I regard the characterization of an issue as "introduced at a final moment as a fait accompli" as inadequate, as long as it has followed the rules.

Specifically, and as a case study, Issue 511 was raised by Robert 9/9/2020 and known, with the addition that: "there is an inconsistency right now (as you recognized long ago!) that should *not persist in version 7.0*".

During the following preparation phase of 7.1, the issue was not resolved because of lack of insight and homework about how this relates to CRMsci, and how to deal with not measured dimensions, such as results of weather forecast etc.

In March 2, I sent a decidable, consistent proposal by e-mail to CRM-SIG, well before the meeting, which was supported by Robert and not commented by anybody else.

As being a "last minute" correction of 7.1, this issue had priority in the meeting. Therefore it was scheduled for the first session on Monday. Instead the discussion time was consumed by issues of lower relevance, even though the session chair might have been aware of the relevance. Against my repeated questions, the issue landed in the end at the end of the agenda.

The issue was decided with a double vote, first if it was ready for being decided, after all pros and cons had been laid out, and secondly by the final decision, orderly following the protocols, as I understand.

Since the objections in the meeting "there is a community of use to take into account" concentrated on monotonicity in general, I would further like to point out, that version 7.1 contains 34 non-monotonic migration instructions, issue 511 being the 35th. All being defined in rules that can be solved algorithmically.

For those 34 decisions, exactly the same community argument holds, if I am not wrong. It was a major effort in a series of issues to base the decision why a concept belongs to CRMbase on a purely rational base, explicitly formulated in a guideline, starting with Issue 260: Review specializations of Appellation. This created *a precedent* that in *well-justified cases*, logical and intellectual consistency has priority over monotonicity, and what the community is used to. Inability to correct severe intellectual errors of the past would be fatal and question renew-ability overall.

Therefore the fact that Issue 511 is non-monotonic did not constitute new, issue specific evidence brought up in the meeting. The SIG decision judged it by majority as a well-justified case to break monotonicity. Details will be in the minutes.

Further, for respecting the community, non-monotonic changes after version 7.1 should be reduced to absolute minimum, but any future correction of a range of E1 CRM Entity is necessarily non-monotonic. The decided range of P39 can now be adjusted by monotonic corrections, if it would be necessary. *I hope we will never again *need non-monotonic changes after 7.1.

The discussion shows the problem of a single violation of the "bottom-up" modelling principle: Excessive generalizations hinder orderly evolution of a model. With issue 511 resolved, CRMbase does no more contain excessive over-generalizations, which I regard as a quality criterion, and it was a criterion to give it priority in the last meeting. It could have been avoided much earlier.

What are the lessons learned? In my opinion:

A) Overgeneralizing the range of properties is not a virtue to be more tolerant for future applications, but blocks proper evolution.

B) *Priorities of issues to be discussed* should be made even more clear. We ask the community for help. Teleconferencing has become mature enough. We need more engagement to participate in meeting preparation!

C) All participants to read carefully the issues and be prepared. All participants to understand discussion time as a rare resource and the rationale of the decision procedures.

D) *Session chairs* to change order of issues if necessary.

Please comment, what can be improved in our procedures!
Please all to make proposals how we can increase throughput, possibly by distribution, and yet maintain methodological and content consistency.

All the best and thank you all for your opinions to this issue!

Martin

Martin Doerr via Crm-sig Thu, 25 Mar 2021 06:08:53 -0700

Dear Francesco,

Thank you for these important considerations. I do support your broader vision. I see several specific challenges that should better be respected on a common path to this vision, which I'd like to explicate and write up the next days, so that we can discuss and effectively address them together.

Kind regards,

Martin

Stefan Gradmann via Crm-sig Tue, 23 Mar 2021 05:52:58 -0700

Yes! -- Stefan Gradmann

Outcome: 

According to the results of the e-vote and following the retraction of the veto by Francesco, the motion to reduce the range of P39 to E18 passed (also changes in affected properties and classes were implemented). A summary of changes can be found here.

The issue closed. 

May 2021

Meetings discussed: