Methodological tips for mappings to CIDOC CRM

Methodological tips for mappings to
CIDOC CRM
Maria Theodoridou, George Bruseker, Maria Daskalaki, Martin Doerr
FORTH - Institute of Computer Science
{bruseker, martin, maria}@ics.forth.gr
daskalakim@hotmail.com

OSLO CAA 2016, March 30, 2016
The CIDOC Conceptual Reference Model (ISO 21127:2006) has been chosen
as the core model for use in several Cultural Heritage projects
2
http://www.ariadne-infrastructure.eu/
http://www.itn-dch.eu/
http://www.researchspace.org/
http://www.parthenos-project.eu/
http://wiss-ki.eu/
http://americanartcollaborative.org/

CIDOC CRM extension suite
3
Few concepts,
high recall
Special concepts,
high precision
CRMinf
CRMsci
CIDOC-CRM
CRMarchaeo
CRMdig
CRMba
Thing
Actor
Event
was present at
happened at

OSLO CAA 2016, March 30, 2016 4
Mapping one schema to another
 Foundational activity
 Convert data stored in existing (source) schemata to an expression in a form (rdf)
compatible with CIDOC CRM and its extensions (target schema)
 It’s not a matching exercise, it’s not matching of terms, but interpreting relations
 It is a consideration of the meaning of the source and a re-expression of that in the
target
 Common point of reference in interpreting source and target as being descriptions
of the same reality
 Goal
 Preserve as much as possible the initial ‘meaning’ of the data
 Enable information exchange and integration between heterogeneous sources of
cultural heritage information

OSLO CAA 2016, March 30, 2016 5
Target Range:
Literal
Target Domain:
E22 Man-Made Object
P43 has dimension
Source Path:
weights
Source Domain:
Coin
Source Range:
WEIGHT
P90 has value
Target Path:
Intermediate Node:
E54 Dimension
Constant Node:
E58 Measurement Unit
P91 has unit
P2 has type
Constant Node:
E55 Type weight
gr
Mapping one schema to another

Mapping & Training Activities
 Several mapping activities were initiated trying to convert existing schemata of
Cultural heritage data to CIDOC CRM
 ARIADNE: diverse archaeological databases
 Research Space: British Museum and Rijksmuseum data
 Training activities
 Research Space Workshops in Yale and in Oxford Universities
 ITN-DCH CIDOC CRM board game
 ARIADNE Summer Schools and Transnational Access
6

General Mapping Principles &
Approach
7

Define primary domain
Define primary domain - typically represents the focus of interest
Identify the CIDOC CRM class that models it best
 Is there a need for a new Class?
 Do we need properties that are not available in CIDOC CRM?
The introduction of a new class should comply with the “Minimality” modelling
principle of CIDOC CRM (CIDOC CRM, 2015):
“A class is not declared unless it is required as the domain or range of a
property not appropriate to its superclass, or it is a key concept in the
practical scope”
8

Example: Two approaches for defining Coin
 Introduce a specialization of E22 Man-Made Object:
Exx Coin subclass of E22 Man-Made Object
 Define the Type of E22 Man-Made Object:
E22 Man-Made Object. P2 has type: E55 Type = “Coin”
To choose we need to answer the question:
 Does the new class Coin have new properties that are
not available in E22?
Create a new Class, work with it and if it comes out that it has no new
properties, revert to its CRM superclass
9
Define primary domain

Practical Mapping Tips
10

We map local identifiers in relational database tables explicitly only if these
identifiers are visible in the user interface and used in other documents as
well. Alternatively, we use the local database identifiers only for generating
URIs for the record instance, if the thing identified belongs to our holdings or
was created/invented by us. Otherwise we better use a global authority.
Golden rules for identifiers:
 they should be widely known for the item they identify
 they should be related to those who can confirm the relation
between the identifier and the identified (museum object identifiers)
11
Common database fields:
Local identifiers

only a local URI:
or
a local URI and an Identifier:
12
Local identifiers
<COIN>
<ID>627</ID>
<COUNTRY_ID>1</COUNTRY_ID>
<FIND_SPOT_ID>242</FIND_SPOT_ID>
<FIND_MANNER_ID>2</FIND_MANNER_ID>
………………..
</COIN>
E22 Man-Made Object
http://www.oeaw.ac.at/coins/627
E22 Man-Made Object
http://www.oeaw.ac.at/coins/627
E42 Identifier
http://www.oeaw.ac.at/id/627
P1 is identified by

Whenever possible use a global authority
El Greco (Dominico Theotokopoulos)
VIAF:
or
GETTY ULAN:
13
Local identifiers
E39 Actor
http://viaf.org/viaf/100215785/
E39 Actor
http://vocab.getty.edu/ulan/500010916

Label and Appellation are alternative implementations in rdf
For contemporary names (not historical) the use of rdfs:label + language
attribute is sufficient
We use the E41 Appellation class only if there is need to assign some
property to the Appellation
14
Appellations

Yes/No fields of relational databases are typically mapped to types.
Table name: Gefaesse (vessels) (E22 Man-Made Object)
Field name: GefFragmen: yes/no field
If value of field is yes
then Gefaesse.GefFragment maps to E55 Type fragment
15
Yes/No fields
E22 Man-Made Object
Vessel 1
E55 Type
fragment
P2 has type

All the hidden constants, the common context information of a db, can be
added as constant node information at a suitable object type, typically in
the domain node.
Quite often semantics are built into
 the User Interface
 the queries
 values of the identifiers
16
Implicit and contextual information

OSLO CAA 2016, March 30, 2016 17
Target Range:
Literal
Target Domain:
E22 Man-Made Object
P43 has dimension
Source Path:
weights
Source Domain:
Coin
Source Range:
WEIGHT
P90 has value
Target Path:
Intermediate Node:
E54 Dimension
Constant Node:
E58 Measurement Unit
P91 has unit
P2 has type
Constant Node:
E55 Type weight
gr
Implicit and contextual information
Constant Node:
E40 Legal Body
OEAW
P50 has current keeper

 Need to separate categorical and factual data
Inconsistent information:
 Find spot -> for a specific coin
 Historical facts -> for a category of coins
18
Categorical vs factual info

Categorical vs factual info
19
P108i was
produced by
P137 exemplifies
PC1 produced
things of type
 Need to extend the model in order to
support categorical production
(similar to FRBRoo
R26 produced things of type)
E12 Production
production of coins AU
E55 Type
AU from Rome
E22 Man-Made Object
MyCoin

Nationality can be modeled as:
 membership in an E74 Group = citizenship, social participation
an actor may join and/or leave the group (former, current member)
 E67 Birth = provenance
 E55 Type = behavior
suitable for ambiguous, fuzzy characterizations
20
People:
Groups, Nationality, Origins

OSLO CAA 2016, March 30, 2016 21
People:
Art Institute Chicago, Photo M. Doerr, 2010
E55 Type
Spanish
E39 Actor
http://viaf.org/viaf/100215785/
P2 has type
E74 Group
Cretan
P107i is current or former member of
E67 Birth
The birth of El Greco
E53 Place
Greece
P7 took place at
P98i was born

Art Institute Chicago
http://www.artic.edu/aic/collections/artwork/87479
22
People:

Accidental roles are roles that do not characterize an actor independently
from a particular context of activity. These roles are properties of the P14
carried out property but this cannot be represented in rdf.
They should NOT be assigned to the actor
23
People:
Accidental roles
E39 Actor
E55 Type
E7 Activity
P14 carried out by (performed) P14.1 in the role of

Solution 1: Specializing P14 carried out by for each role
suitable if we know a priory all the possible values
P14.1.1 carried out by as researcher P14.1.2 carried out by as farmer
People:
Accidental roles
E7 Activity
http://www.ics.forth.gr/Activity/2013HeraklionCRM-SIGmeeting
E7 Activity
http://www.ics.forth.gr/Activity/2013MaladesPlantingOliveTrees
E39 Actor
http://www.ics.forth.gr/Actor/Doerr

Solution 2: Introducing class PC0 Typed CRM_Property and its subclass PC14 carried out by
and properties P01 has domain and P02 has range
PC14 carried out by
http://www.ics.forth.gr/Actor/Doerr_as_Researcher
PC14 carried out by
http://www.ics.forth.gr/Actor/Doerr_as_Farmer
E7 Activity
http://www.ics.forth.gr/Activity/2013Hera
klionCRM-SIGmeeting
P01 has domain
E55 Type
http://www.ics.forth.gr/Role/Researcher
E55 Type
http://www.ics.forth.gr/Role/Farmer
P14.1 in the role of
E7 Activity
http://www.ics.forth.gr/Activity/2013
MaladesPlantingOliveTrees
E39 Actor
P02 has range
P01 has domain
P02 has range
PC0 Typed CRM
Property
E1 CRM Entity
P01 has domain (P01i is domain of)P02 has range (P02i is range of)
People:
Accidental roles

Solution 2: equivalence with CIDOC CRM definition
PC14 carried out by
http://www.ics.forth.gr/Actor/Doerr_as_Research
er
E7 Activity
http://www.ics.forth.gr/Activity/2013Hera
klionCRM-SIGmeeting
P01 has domain
E55 Type
E39 Actor
P02 has range
P14 carried out by (performed)
PC0 Typed CRM
Property
E7 Activity
http://www.ics.forth.gr/Activity/2013
HeraklionCIDOC-CRM-SIGmeeting
E39 Actor
E55 Type
equivalent to P14.1 in the role of
People:
Accidental roles

Solution 3: Add a note to each Activity with the Person’s role
does not require any extension of the model
Literal
The Activity was carried out by Doerr in the role
of Researcher
Literal
The Activity was carried out by Doerr in the role
of Farmer
E7 Activity
P2 has note
E7 Activity
E39 Actor
P14 carried out by
P2 has note
People:
Accidental roles

Solution 4: Assign a type to each Activity
does not require any extension of the model
E55 Type
http://www.ics.forth.gr/Role/Research
E55 Type
http://www.ics.forth.gr/Role/Farming
E7 Activity
E7 Activity
E39 Actor
P14 carried out by
People:
Accidental roles
P2 has type P2 has type

People:
Accidental roles - Issuer
29
Target Range:
E39 Actor
Target Domain:
E22 Man-Made Object
P108i was produced by
Source Path:
ISSUER_ID == PR_ID
Source Domain:
//COIN
Source Range:
//ISSUER
P14 carried out by
E12 Production
p1
"Issuer" is an accidental role, does not characterize
an actor independently from particular contexts of
activity. Therefore the Actor does not have the type
"Issuer" but the activity only has the type "Issuing“
P17 was motivated by
E7 Activity
ia1 E55 Type
Issuing
“gave order”

Political office is modeled as an E74 Group with one member
an actor may join and/or leave the group (former, current member)
https://en.wikipedia.org/wiki/Alexis_Tsipras
Alexis Tsipras is the 185th and current Prime Minister of Greece, having been sworn in on 21
September 2015. He previously served as the 183rd Prime Minister of Greece from 26 January
2015 to 27 August 2015.
E74 Group The Prime Minister of Greece
P144i gained member by E85 Joining P143 joined E39 Actor Alexis Tsipras
P4 has time-span E52 Time-Span 26 January 2015
P146i lost member by E86 Leaving P145 separated E39 Actor Alexis Tsipras
P4 has time-span E52 Time-Span 27 August 2015
P144i gained member by E85 Joining P143 joined E39 Actor Alexis Tsipras
P4 has time-span E52 Time-Span 21 September 2015
30
People:
Political office

 Geopolitical units map to a Period with type “State”, etc.
 Map them as Place but the id has the name of the country plus a date,
when this name of country was assigned.
 It is a snapshot that allows to be always in the right time space context.
always refer to a global authority e.g. TGN (falls within, overlaps, is
identified by)
31
Places:
Countries

“Why hell is not a good place”
 E53 Place represents always a real extent in space independent of time
 An imaginary place can be represented as an E89 Propositional Object
The ambiguity between real and imaginary is an ambiguity of the phrase and
not an ambiguity in the mind of the author of the phrase (as long as he is
sane)
Alternatives should be made explicit: principle of recall over precision
32
Places:
Imaginary Places

How many interpretations of hell?
33
Places:
Imaginary Places
Taoist Hell
Catholic Hell,
Pisa 13th century
Byzantine-Orthodox
Hell
Crete

William Blake's depiction of "The Vestibule of Hell and the Souls Mustering to Cross
the Acheron" in his Illustrations to Dante's "Divine Comedy“ object 5 c. 1824-27. The
original for the work is held by the National Gallery of Victoria
Example taken from https://en.wikipedia.org/wiki/Acheron
34
Places:
Imaginary Places
E73 Information Object
The Vestibule of Hell and the Souls Mustering
to Cross the Acheron
E89 Propositional Object
Hell
P67 refers to E53 Place
Acheron

Now Italian archaeologists working at the Greco-Roman site of ancient Hierapolis
(modern-day Pamukkale) in Turkey have uncovered that city's gate to the underworld.
Example taken from http://news.nationalgeographic.com/news/2013/04/130414-hell-
underworld-archaeology-mount-olympus--greece/
35
Places:
Imaginary Places
E89 Propositional Object
Entrance to HellP67 refers to
E53 Place
ancient Hierapolis
E13 Attribute Assignement
possibly refers to
E73 Information Object
P140 was attributed by
P141 assigned
Archaeologists Find a Classic
Entrance to Hell

Pieces of a broken object are mapped as products of an event that destroyed
the object and not as part_of.
 For provenance reasons
 For semantic clarity
As long as the object existed the respective pieces of matter were not distinct,
had no identity
36
Objects:
Pieces of a broken object
E22 Man-Made Object
Piece 1
E81 Transformation
destruction of object
E55 Type
fragment
E22 Man-Made Object
broken object
P124 transformed
P2 has type
P123 resulted in

Pieces of a set are mapped as part of
 Chess set has parts: Chess board and chess pieces
Collective objects in the general sense, like a tomb full of gifts, a folder with stamps or
a set of chessmen, should be documented as instances of E19 Physical Object, and
not as instances of E78 Collection. This is because they form wholes either because
they are physically bound together or because they are kept together for their
functionality.
37
Objects:
Pieces of a set
E22 Man-Made Object
Chess set
E22 Man-Made Object
Chess pieces
E22 Man-Made Object
Chess boardP46 is composed of

OSLO CAA 2016, March 30, 2016 38
Conclusions
Always consult the reference document
Definition of the CIDOC Conceptual Reference Model
http://www.cidoc-crm.org/comprehensive_intro.html
Read carefully the scope note of the class and property that you intend to use.
Learning to use a formal ontology requires understanding the basic principles
expressed in its scope, classes and relations
http://www.cidoc-crm.org/index.html
New site (available for testing)
http://new.cidoc-crm.org/

OSLO CAA 2016, March 30, 2016 39
Conclusions
IF ASK!

Thank you!
Maria Theodoridou
maria@ics.forth.gr
40

Methodological tips for mappings to CIDOC CRM

More Related Content

Viewers also liked (18)

Similar to Methodological tips for mappings to CIDOC CRM (20)

More from ariadnenetwork (20)

Recently uploaded (20)

Methodological tips for mappings to CIDOC CRM