Issue 270: A hoard as crm:E78_Collection ?

Starting Date: 
2015-11-29
Working Group: 
3
Status: 
Done
Closing Date: 
2016-07-13
Background: 

Posted by dan Matei on   29/11/20144

Friends

I have to say that a particular coin is a member of a particular hoard. I googled to see how others are dealing with that, but...

I'm tempted to:

<coin> <crm:P46i_forms_part_of> <hoard>
<hoard> <rdf:type> <crm:E78_Collection>

abusing a bit the E78 scope note:

"This class comprises aggregations of instances of E18 Physical Thing that are assembled and maintained (“curated” and “preserved,” in museological terminology) by one or more instances of E39 Actor over time for a specific purpose and audience, and according to a particular collection development plan."

The "... according to a particular collection development plan." troubles me. Can we say that the guy burying a hoard had a "collection development plan" ?

There is a better practice for modelling that ?

Posted Christian-Emil on 29/11/2014

In the case a curator in a museum buries his collection, a hoard may be considered as as a collection. The collection class is intended for museum collections, see the examples in the scope note.

Posted by Martin on 29/11/2014

Continuing:

I don't know, why E78 Collection attracts so much attention. The scope note of
E19 Physical Object says:

"The class also includes all aggregates of objects made for functional purposes of whatever kind, independent of physical coherence, such as a set of chessmen. Typically, instances of E19 Physical Object can be moved (if not too heavy)."

The CRM is not a terminological system to classify things. It is made to provide relevant properties. We should only use a more specific class, if we expect the respective additional  properties to be relevant for
querying. To say that an E19 "has type: Hoard" should be enough. Only if we want to specify a curator and an E87 Curation Activity with a curation plan, using E78 Collection would be adequate. The less classes we use, the more effective the queries.
 

Posted by Stephen Stead 29/11/2014

Martin

The problem probably lies in the word “Collection”! Everyone reads that and thinks that the defining characteristic is the act of collecting rather than the true differentiator which is the curation.

Perhaps changing the name to “E78 Curated Set” would solve the problem. Nobody would know what it meant and so would read the scope note!!!!

Posted by Simon Spero on 29/11/2014

The definition of Collection possibly overly restrictive, even from an archival point of view.

An collection of records will probably have  been assembled by a different agent to the curator; materials may be discarded as "not archival", but if the fonds gets respect, the curator is not free to do much assembling.

It is possible to finesse this by a sufficiently broad reading of "plan", but it seems as if the roles of assembling curating/preserving are intrinsically linked in the CRM (I ought to take a look at the old EAD mapping) . 

Posted by Martin on 30/11/2014

Dear Simon,

This is an interesting discussion. Preserving things others have collected has been described as
"SECONDARY COLLECTOR CONTEXT" and is well distinguished in archival practice to my knowledge.
See also: https://www.ideals.illinois.edu/bitstream/handle/2142/45860/EDM-DCC_Whit...
The case had been discussed when we defined E78. The "collection in the collection" can either be
seen as one object in  a collection, or, if incompletely acquired , the secondary collection plan can be to acquire  the missing parts to the original collection, or, to continue the primary collector's plan. In any case, the original collector was a curator to the collection.

Posted by Dan Matei on 1/12/2014

Hi

Well, crm:E78_Collection attracts so much attention either because it is important or it has problems, or both :-)

(Second thought: we can say that the guy burying a hoard had a "collection development plan": he planed to "develop", the hoard, retrieving it after some time and spending the coins :-)

True that crm:E19_Physical_Object "includes all aggregates of objects", but I feel the need of something like crm:Fx_Assemblage between crm:E19_Physical_Object and E78_Collection, because I would like to make a clear (conceptual) distinction between:

<helmet>  <crm:Py_is_component_of> <armor>

or

<coin> < crm:Py_is_component_of > <hoard>

and

<wheel>  <crm:P46i_forms_part_of> <car>

So, I would like a crm:Fx_Assemblage exactly for the reason you invoke: „We should only use a more specific class, if we expect the respective additional  properties to be relevant for querying.".

Posted by Dan Matei on 1/12/2014

Hi, dear Vladimir

Here is what I replyed to Martin (or you see the list ?):

-------------------------------------

Well, crm:E78_Collection attracts so much attention either because it is important or it has problems, or both :-)

(Second thought: we can say that the guy burying a hoard had a "collection development plan": he planed to "develop", the hoard, retrieving it after some time and spending the coins :-)

True that crm:E19_Physical_Object "includes all aggregates of objects", but I feel the need of something like crm:Fx_Assemblage between crm:E19_Physical_Object and E78_Collection, because I would like to make a clear (conceptual) distinction between:

<helmet>  <crm:Py_is_component_of> <armor>

or

<coin> < crm:Py_is_component_of > <hoard>

and

<wheel>  <crm:P46i_forms_part_of> <car>

So, I would like a crm:Fx_Assemblage exactly for the reason you invoke: „We should only use a more specific class, if we expect the respective additional  properties to be relevant for querying.".

-------------------------------------

I'm still in doubt...

Posted by Martin  on 1/12/2014

Dear Dan,

This is a deep methodological question:
"True that crm:E19_Physical_Object "includes all aggregates of objects", but I feel the need of something like crm:Fx_Assemblage between crm:E19_Physical_Object and E78_Collection, because I would like to make a clear (conceptual) distinction between:

<helmet>  <crm:Py_is_component_of> <armor> or <coin> < crm:Py_is_component_of > <hoard>

and
<wheel>  <crm:P46i_forms_part_of> <car>"

This is not what we have made the CRM for. You SHOULD not make such a distinction,
if there is no use case that would create query ambiguity. CRM is not a language to describe the nuances of cultural heritage objects. Since a hoard cannot have parts like a car,  and a car not parts like a hoard, the distinction does not help in any query.

We all must be clear that the CRM is made ONLY for searching possibly related things across disciplines in a global network of knowledge. People don't use the CRM because it has already 150 properties. It is absolutely counterproductive to introduce more. Already,
it is impossible to write queries using all 150 properties without long preparation.

My position is, in order to capture the nuances of cultural heritage, we should write good scholarly texts, and not admire the brave new world of formal ontologies as the future scholarly language.

We must be absolutely clear that the wish of each scholar to make a clear (conceptual) distinction is not an argument for the CRM. If it would be, we would now, after 18 years,
struggle with millions of distinctions, and no chance people to understand the whole.
Only clear functional requirements that a search statement would become ambiguous or return too much noise is an argument.
 

Posted by Dan Matei on 1/12/2014

Dear Martin

On 1 December 2014 at 17:26, martin <martin@ics.forth.gr> wrote:

    This is a deep methodological question:
    "True that crm:E19_Physical_Object "includes all aggregates of objects", but I feel the need of something like crm:Fx_Assemblage between crm:E19_Physical_Object and E78_Collection, because I would like to make a clear (conceptual) distinction between:

    <helmet>  <crm:Py_is_component_of> <armor> or <coin> < crm:Py_is_component_of > <hoard>

     and
    <wheel>  <crm:P46i_forms_part_of> <car>"

    This is not what we have made the CRM for. You SHOULD not make such a distinction,
    if there is no use case that would create query ambiguity. CRM is not a language to describe the nuances of cultural heritage objects. Since a hoard cannot have parts like a car,  and a car not parts like a hoard, the distinction does not help in any query.

My main criterion (not always explicit !) to use one property or other is not the query (dis)ambiguity, but the facility of the semantic reasoning, i.e. inference. Am I wrong ?

So, my feeling (not more than that) is that the reasoner is helped if crm:Py_is_component_of suggests that the subject of the assertion is a thing "functional" in itself, vs. crm:P46i_forms_part_of that suggests otherwise.

We have built the CRM because we had such principles. It is overdue to write them down.

Good idea !

Posted by Simon Spero on 1/12/2014

I'm not entirely sure that this is the right reading of Wickett et. al.; the issue seems that in  the CRM, roles are conflated, or at least fused.

It is a plausible reading that  donating a private collection to an archive must cause the identity of the collection to change (with the archival collection being a derivative, (possibly improper) sub-collection of the donated collection.

This could be the case if the intensional definition of the collection as received by the archive necessarily includes the identity of the donor (though presumably the indexing, once resolved, would not require the identity of the collection to change if transferred to another institution).

It could also be the case if the  collection policy /plan is an identity criterion, and necessarily changes when the curator changes.

However this reading would seem to require any change in policy or plan to form a new intensional collection, even if the change in policy is one that a priori cannot change the contents or description of the former collection (e.g an institution wide policy requiring that "all unicorns will be stored in vibranium edged boxes").

Since I am rarely on the "same" page as Karen wrt identity I'm checking with the editor :-) 

Posted by Martin on 1/12/2014

Dear Dan,

On 1/12/2014 7:26 μμ, Dan Matei wrote:
> My main criterion (not always explicit !) to use one property or other is not the query (dis)ambiguity, but the facility of the semantic reasoning, i.e. inference. Am I wrong ?
I'd regard this as secondary application. Many  reasoning applications need to narrow the world down, which often comes in direct conflict conflict with the recall of a global query.
If that happens, search has preference over expressive power. Of course, reasoning can also increase recall. Then it is preferable.

In other words, imagine a global network of knowledge. The idea is, if I cannot get all facts possibly relevant to my research question, I need to worry about drawing deep conclusions from the rest. Once we have solved that, then we can go into the next round of reasoning services. That may be a CRM extension, but I'd argue not the CRM itself.
>
> So, my feeling (not more than that) is that the reasoner is helped if crm:Py_is_component_of suggests that the subject of the assertion is a thing "functional" in itself, vs. crm:P46i_forms_part_of that suggests otherwise.
If that would be the case, we would need an example in which the same thing can be
integrated in one thing as functional component, and in another as non-functional, and the distinction is so frequent, that we would be confused by too many answers otherwise.
But, feeling is not enough for the CRM . We always require enough real data to demonstrate the need.

Of course, we are all free to make extensions for particular applications.
So, what is reasonable for an application, may not be relevant enough to increase
an ISO standard . Anyway always good to think things through...
 

posted by Martin on 1/12/2014

On 1/12/2014 9:47 μμ, Simon Spero wrote:
> I'm not entirely sure that this is the right reading of Wickett et. al.; the issue seems that in  the CRM, roles are conflated, or at least fused.
That is the intention. The CRM should cover things by reasonable generalizations. It is not directly an ontology for the archival world. Specializations should be possible, that should not require to invalidate "fused" roles that provide better recall. Would that make sense?

Current Proposal: 

Should we rename the  name of the class E78 Collection ?


In the 32nd joined meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9 and the 25th FRBR - CIDOC CRM Harmonization meeting, the sig agreed on renaming the name of E78 Collection. A proposal is to rename E78 to  "Curated Holdings". The sig have assigned to Dominic of thinking a better name. Martin will send a paper from Europeana to Dominic

Oxford, February 2015


In the 34th joined meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9 and the 27th FRBR - CIDOC CRM Harmonization meeting, the crm-sig  decided to vote for the name of E78 Collection (issue 270). The proposal is to rename the E78 Collection to E78 Curated Holdings

October, Crete, 2015

Posted by Carlo meghini, George Bruseker, Martin Doerr, Guenther Goerz     on 5/2/2016

 "YES"


Posted by Wolfgang Schmidle on 6/2/2016

Is the s in "holdings" necessary? http://www.merriam-webster.com/thesaurus/holdings suggests using the plural, but to me as a non-native speaker it seems that "holdings" can also mean, to a varying degree, the items in a collection rather than the collection as a whole.

Google says:

1. "curated holdings"
* Of the >3 million curated holdings, approximately 75% of are identified to genus level or better.
* Curated holdings have more than doubled (factor of 2.5) over the last 20 years.
* Total fully curated holdings comprise approximately 320,000 specimens in 1,380 drawers [as part of the Diptera collection having approximately 897,000 specimens]
* “Wine Wednesdays” provide half-off savings on any bottle of wine from the restaurant's well-curated holdings, which boast more than 1,500 ...
--> at least in the first two examples: the items in a collection rather than the collection as a whole?

* Curated holdings should be visible, as appropriate.
* There, forty exhibitors straight from Bruno's enviable Rolodex will showcase their curated holdings of twentieth-century fine and decorative arts.
--> more than one collection

2. "curated holding"
* the Lexington Collection, a select and carefully curated holding of gorgeous Greek and Roman coins
* As the largest, most comprehensive, and best-curated holding of arthropods, predominately insects, ...
* "Unhinged," is a curated group exhibition of nearly 225 artists from Pierogi's Flat Files ..., itself a curated holding of over 800 artists' works.
--> a single collection

With E78 Curated Holding, on the other hand, it may be confusing if one talks about two E78's, i.e. Curated Holdings.

What was the argument against Steve Stead's suggestion "curated set" (apart from the fact that "nobody would know what it meant")? What about "curated collection"?
 


Posted by Martin on 6/2/2016

 

Good point! "Curated Holding" hence... 


Posted by Franco Niccolucci    on 6/2/2016

Not a vote, but an amateur comment.

-> Collection vs Curated Holdings
Of course. There are “collections" of things which are not curated, just an assemblage of stuff. Like some old objects I have which I call my collection of old computational instruments, but according to the qualified opinion of my wife is just a dust attractor.

But, what about a collection (oops, curated holdings) of videogames like the one of the National Videogame Museum of Frisco, TX, USA (https://en.wikipedia.org/wiki/Videogame_History_Museum)? It seems that what they collect (oops, “curate and hold”) forms a collection of E28 Conceptual Object rather than E18 Physical Thing.

Possibly defining E78 as a curated assemblage of instances of E72 Legal Object or, maybe even better, of E70 Thing (in some cases rights may be difficult to ascertain as in the examples below), would reconcile the CRM with the curators of the Frisco museum; and also simplify life to the curators of the Museo Officina Profumo from the Old Pharmacy of Santa Maria Novella in Florence, which I hope you will be able to visit when you come here for the next CRM meeting, as well as to their colleagues of the Musèe International du Perfum of Grasse, France. Is fragrance a Physical Thing? What about the exhibitions for visually impaired people consisting in a garden where they smell the scent of different flowers along the visit?

-> Holding vs holdings
My knowledge of English is even more superficial than my knowledge of the CRM, so my opinion is not so qualified. It seems to me that “holdings” is used here in the plural as a collective noun, and not as the result of putting together one individual “holding" with another one and another one and so on, as a genuine plural; it is unfortunately written in the plural, what makes pluralizing it a bit awkward if one wants to keep the “holdings" distinct from each other, as Werner has pointed out. Not only: it makes difficult to express indefiniteness as when using the indefinite article “a” to indicate “one of a series", like in the second sentence of this email where I was in trouble being unable to call it “a curated holdings” as “a” cannot go with “holdings”. So what would be the correct way of expressing the equivalent of “a collection” i.e. one instance whatever of the class “curated holdings”? Wouldn't the sentence "Is_A curated_holdings” describing a future subclass of E78 sound strange?

Finally, holdings is a synonym of property, according to my Oxford dictionary, which is not the case if the objects forming the collection (oops, holdings) are just deposited, on lean, or illegally/controversially detained. There are famous examples of the latter.

So I would prefer “curated set” or better “curated assemblage”, being “set" a very generic term not incorporating the concept of intentionality.

Anyway, this concern about a name is not so important: that which we call a rose, by any other name would smell as sweet. More important is, in my opinion, the issue concerning E18 vs E28 => E72/E70 to characterize the components of E78, as noted above. 


Posted by Martin on 6/2/2016

Dear Franco,

As an explanation what has been discussed in the meeting:
Librarians talk about "holdings" wrt to library contents.
 It means the physical copies. Therefore they are physical. So, there is a good practice of the term there,
which actually motivated the proposal.

The argument why we have modelled collections as physical things, regardless the intentional content:
The video games, in order to be in a collection, must be represented by physical copies.
If we would disregard the physical nature of the copy, we could not talk about location and destruction.
This is also the sense how libraries distinguish holdings from content.
 

Would that make sense? 


Posted by Franco Niccolucci     on 7/2/2016

 Dear Martin,

very clear, thanks. Just a bit nineteen-century-ish: what about digital libraries and digital curation, which does not concern curating the servers (or the Cloud) on which their instances of E28 reside?

It seems that they are beyond the (current) scope of the CRM, what sounds a bit paradoxical, but nevertheless perfectly logical. 


Posted by Martin on 7/2/2016

Dear Franco,

May be I confused you: by physical copies in the CRM we do not mean on paper. We mean any material
carrier, which implies computer discs, albeit somewhere in the Cloud only the provider knows (but someone
must know it).

Badly enough, this is not in the collection scope note. It is only implicit in the information carrier.

I propose to adapt the scope note of E78. The issue was modelled in more detail in FRBRoo, but not transferred
back into the CRM.

Would we agree now ?

Martin

E84 Information Carrier

Subclass of:         E22 Man-Made Object

Scope note:         This class comprises all instances of E22 Man-Made Object that are explicitly designed to act as persistent physical carriers for instances of E73 Information Object.

An E84 Information Carrier may or may not contain information, e.g., a diskette. Note that any E18 Physical Thing may carry information, such as an E34 Inscription. However, unless it was specifically designed for this purpose, it is not an Information Carrier. Therefore the property P128 carries (is carried by) applies to E18 Physical Thing in general.

Examples:          

§  the Rosetta Stone

§  my paperback copy of Crime & Punishment

§  the computer disk at ICS-FORTH that stores the canonical Definition of the CIDOC CRM

The 34nd joined meeting of the CIDOC CRM SIG and ISO/TC46/SC4/WG9 and the 27th FRBR - CIDOC CRM, A proposal was to be “E78 to "Curated Holdings” and it is decided that the members should vote by email

Heraklion, October 2015

Outcome: 

In the 35th joined meeting of the CIDOC CRM SIG and   28th FRBR - CIDOC CRM Harmonization meeting, the crm-sig changed the name of the E78 Collection to E78 Curated Holding and modified the scope note as well as the scope note of  E84 Information Carrier. The changes have been incorporated into CIDOC CRM version 6.2.2.

The issue is closed

Prato, February 2016

Available Documents: