In the HEED project we encountered problems in finding the proper ClassOfIndirectProperty for the domain of Pumps. The project participants work from a kind of RDL defined by the Hydraulic Institute and by NIST. These are good definitions, meaningful for that domain, that, however, not always fit in the generic mould of the RDL.
Where an indirect property, in Part 2, "is derived or deemed from the properties of the possible_individual", that derivation can be, and often is, a complex matter, that very well may differ from domain to domain.
That is why I toss up the idea to classify all instances of ClassOfIndirectProperty with a ClassOfClassOfRelationship that indicate the domain or domains in which that ClassOfIndirectProperty is meaningful. By doing this we shift the responsibility for the definitions to where that belongs: the (international) representatives of the applicable domain.
Of course it will happen that instances of ClassOfIndirectProperty, defined in different domains, will actually appear to be identical. If the domain representatives involved agree on that, these instances can be declared equivalentClass.
Please voice your opinion on this.
Domain-specific instances of ClassOfIndirectProperty in RDL
-
- Posts: 283
- Joined: Sun Jan 22, 2012 10:02 pm
-
- Posts: 77
- Joined: Tue May 15, 2012 8:48 am
Re: Domain-specific instances of ClassOfIndirectProperty in
While this is a valid approach where such domains are clearly defined I would be againstHansTeijgeler wrote:In the HEED project we encountered problems in finding the proper ClassOfIndirectProperty for the domain of Pumps. The project participants work from a kind of RDL defined by the Hydraulic Institute and by NIST. These are good definitions, meaningful for that domain, that, however, not always fit in the generic mould of the RDL.
Where an indirect property, in Part 2, "is derived or deemed from the properties of the possible_individual", that derivation can be, and often is, a complex matter, that very well may differ from domain to domain.
That is why I toss up the idea to classify all instances of ClassOfIndirectProperty with a ClassOfClassOfRelationship that indicate the domain or domains in which that ClassOfIndirectProperty is meaningful. By doing this we shift the responsibility for the definitions to where that belongs: the (international) representatives of the applicable domain.
Of course it will happen that instances of ClassOfIndirectProperty, defined in different domains, will actually appear to be identical. If the domain representatives involved agree on that, these instances can be declared equivalentClass.
Please voice your opinion on this.
making this a mandatory requirement. As you suggest this could lead to a large increase in
the number of seemingly identical Class Of Property Instances with unpredictable results.
Think of the number of Classes and Individuals that could be assigned the property
APPROXIMATE INSTALLATION WEIGHT
I am in favour of being able to Classify sunch instances BUT feel strongly that it should only be
used where appropriate and necessary.
What concerns me rather more is the confusion between ClassOfIndirectProperty and SinglePropertyDimension
They appear to have been used in the RDL in an essentially arbitrary manner. For example
CAPILLARY TUBE LENGTH - ClassOfIndirectProperty
TUBE LENGTH - SinglePropertyDimension
The classic example is of course APPROXIMATE INSTALLATION WEIGHT for which there are TWO entries
BOTH of which are of entity type ClassOfIndirectProperty
APPROXIMATE INSTALLATION WEIGHT(Retired) - RDS1467625211
and
APPROXIMATE INSTALLATION WEIGHT - RDS1661758501
The only difference appears to be that the newer definition has the PropertySpace classified as WEIGHT
WEIGHT is of course a defined as entity type SinglePropertyDimension
as is ESTIMATED WEIGHT
Regards
Keith
Re: Domain-specific instances of ClassOfIndirectProperty in
If I understand problem correctly, some naming conventions should be agreed first.
IndirectPropretyX specific for domain A and its namesake IndirectPropertyX for domain B can not have the same label. Probable domain reference should be included in the name.
Additional classification by domain COCOR can only make automated reasoning easier.
IndirectPropretyX specific for domain A and its namesake IndirectPropertyX for domain B can not have the same label. Probable domain reference should be included in the name.
Additional classification by domain COCOR can only make automated reasoning easier.
-
- Posts: 283
- Joined: Sun Jan 22, 2012 10:02 pm
Re: Domain-specific instances of ClassOfIndirectProperty in
Hi folks,
I overlooked something: any ClassOfIndirectProperty in the RDL shall be defined as class of relationship between ClassOfIndividual and a PropertySpace.
In case a particular ClassOfIndirectProperty would apply to, for example, Pumps then that instance of ClassOfIndividual shall be PUMP, which means that all specializations inherit that ClassOfIndirectProperty or have a specialization of the latter.
See for instance a reworked diagram for one of the templates (the parts in grey are shown for reference only): The ClassOfIndividual in the red ellipse should be, for the example, http://posccaesar.org/rdl/page/RDS427229 (PRESSURE VESSEL). The SME's for Pressure Vessels then should agree on the applicable instances of ClassOfIndirectProperty for that domain.
I have not seen the classes of indirect property being defined that (part 2 compliant) way in the RDL. So this is a nice opportunity to do it right, preferrably under the auspicies of ISO, so that all those SME's can be rounded up.
So I withdraw my suggestion for a special classification.
I overlooked something: any ClassOfIndirectProperty in the RDL shall be defined as class of relationship between ClassOfIndividual and a PropertySpace.
In case a particular ClassOfIndirectProperty would apply to, for example, Pumps then that instance of ClassOfIndividual shall be PUMP, which means that all specializations inherit that ClassOfIndirectProperty or have a specialization of the latter.
See for instance a reworked diagram for one of the templates (the parts in grey are shown for reference only): The ClassOfIndividual in the red ellipse should be, for the example, http://posccaesar.org/rdl/page/RDS427229 (PRESSURE VESSEL). The SME's for Pressure Vessels then should agree on the applicable instances of ClassOfIndirectProperty for that domain.
I have not seen the classes of indirect property being defined that (part 2 compliant) way in the RDL. So this is a nice opportunity to do it right, preferrably under the auspicies of ISO, so that all those SME's can be rounded up.
So I withdraw my suggestion for a special classification.
-
- Posts: 283
- Joined: Sun Jan 22, 2012 10:02 pm
Re: Domain-specific instances of ClassOfIndirectProperty in
@Keith - I agree with your observation that there is something to be desired with respect to the accuracy of the RDL data. Just at random selected: UPPER LIMIT MAINTENANCE WEIGHT is declared an instance of SinglePropertyDimension!
Let's talk about that in another thread.
Let's talk about that in another thread.
Re: Domain-specific instances of ClassOfIndirectProperty in
Hans, it is not so bad.I have not seen the classes of indirect property being defined that (part 2 compliant) way in the RDL.
Out of 2913 instances of COIP only 81 have no mandatory roles. Remaining 2832 have both hasClassOfPosessor and hasPropertySpace roles occupied by properly typed entities.
-
- Posts: 283
- Joined: Sun Jan 22, 2012 10:02 pm
Re: Domain-specific instances of ClassOfIndirectProperty in
Victor, please give me some examples, that saves me the search. Thanks!
Re: Domain-specific instances of ClassOfIndirectProperty in
Just a small part of an output - COI and Property Space (or subtype) assigned
Or a start of the list of 2832 URIs of COIP's (can send you a whole list):
Code: Select all
grouped by 12 CYLINDER ENGINE : ClassOfInanimatePhysicalObject (1)
found NUMBER OF CYLINDERS: 12 1 : PropertyRange
grouped by 1-HEXENE : ClassOfMolecule (4)
found EVAPORATION POINT: 1.2 degF : PropertyRange
found TEMPERATURE RANGE 145 degF : PropertyRange
found MOLECULAR WEIGHT: 84 u : PropertyRange
found MELTING TEMPERATURE: -218 degF : PropertyRange
grouped by ABB AMG 400 S4 : ClassOfInanimatePhysicalObject (14)
found LOWER LIMIT AMBIENT OPERATING TEMPERATURE: 40 degC : PropertyRange
found OVERALL WIDTH: 860 mm : PropertyRange
found OVERALL HEIGHT: 850 mm : PropertyRange
found NET TOTAL WEIGHT: 2980 kg : PropertyRange
found TRANSPORT COST: 10000 kr : PropertyRange
found LEAD TIME: 365 : PropertyRange
found GROSS LONGITUDINAL CENTRE OF GRAVITY DISTANCE: 910 mm : PropertyRange
found UPPER LIMIT AMBIENT OPERATING TEMPERATURE: 100 degC : PropertyRange
found OVERALL LENGTH: 1936 mm : PropertyRange
found LOWER LIMIT AMBIENT STORAGE TEMPERATURE: -25 degC : PropertyRange
found WARRANTY PERIOD: 365 : PropertyRange
found UPPER LIMIT AMBIENT STORAGE TEMPERATURE: 60 degC : PropertyRange
found PRICE PER UNIT: 500000 kr : PropertyRange
found WEIGHT RANGE 3100 kg : PropertyRange
grouped by ACCELEROMETER/ACCELERATION PICK-UP : ClassOfInanimatePhysicalObject (1)
found ISO2041 ACCELERATION : SinglePropertyDimension
grouped by ACETALDEHYDE : ClassOfMolecule (5)
found TEMPERATURE RANGE 69 degF : PropertyRange
found MOLECULAR WEIGHT: 44 u : PropertyRange
found MELTING TEMPERATURE: -193 degF : PropertyRange
found VAPOUR PRESSURE: 750 mmHg(0C) : PropertyRange
found EVAPORATION POINT: 4 degF : PropertyRange
grouped by ACETIC ACID : ClassOfMolecule (5)
found MOLECULAR WEIGHT: 60 u : PropertyRange
found MELTING TEMPERATURE: 62 degF : PropertyRange
found VAPOUR PRESSURE: 11 mmHg(0C) : PropertyRange
found EVAPORATION POINT: 5.4 degF : PropertyRange
found TEMPERATURE RANGE 244 degF : PropertyRange
http://posccaesar.org/rdl/RDS3788119, http://posccaesar.org/rdl/RDS13019023, http://posccaesar.org/rdl/RDS3797401, http://posccaesar.org/rdl/RDS5739613, http://posccaesar.org/rdl/RDS5726045, http://posccaesar.org/rdl/RDS5719884, http://posccaesar.org/rdl/RDS2215916651, http://posccaesar.org/rdl/RDS14617295, http://posccaesar.org/rdl/RDS14305502, http://posccaesar.org/rdl/RDS14137482, http://posccaesar.org/rdl/RDS14140145, http://posccaesar.org/rdl/RDS14329499, http://posccaesar.org/rdl/RDS7476740, http://posccaesar.org/rdl/RDS14057053, http://posccaesar.org/rdl/RDS17623506, http://posccaesar.org/rdl/RDS14615709, http://posccaesar.org/rdl/RDS7922351, http://posccaesar.org/rdl/RDS8651365, http://posccaesar.org/rdl/RDS5724829, http://posccaesar.org/rdl/RDS8651760, http://posccaesar.org/rdl/RDS17621013, http://posccaesar.org/rdl/RDS7601973, http://posccaesar.org/rdl/RDS17890965, http://posccaesar.org/rdl/RDS16213684, http://posccaesar.org/rdl/RDS3741168861, http://posccaesar.org/rdl/RDS14303921, http://posccaesar.org/rdl/RDS5721676, http://posccaesar.org/rdl/RDS14395134, http://posccaesar.org/rdl/RDS14621748, http://posccaesar.org/rdl/RDS5726685, http://posccaesar.org/rdl/RDS14617966, http://posccaesar.org/rdl/RDS5720076, http://posccaesar.org/rdl/RDS14390376, http://posccaesar.org/rdl/RDS14294779, http://posccaesar.org/rdl/RDS14623212, http://posccaesar.org/rdl/RDS5739805, http://posccaesar.org/rdl/RDS13697738, http://posccaesar.org/rdl/RDS14389888, http://posccaesar.org/rdl/RDS3797248, http://posccaesar.org/rdl/RDS3790941, http://posccaesar.org/rdl/RDS1020853651, http://posccaesar.org/rdl/RDS3797537, http://posccaesar.org/rdl/RDS16744269, http://posccaesar.org/rdl/RDS16697534, http://posccaesar.org/rdl/RDS13268931, http://posccaesar.org/rdl/RDS17623727, http://posccaesar.org/rdl/RDS5715852, http://posccaesar.org/rdl/RDS3787541, http://posccaesar.org/rdl/RDS1724788201, http://posccaesar.org/rdl/RDS13282117, http://posccaesar.org/rdl/RDS14391230, http://posccaesar.org/rdl/RDS1020232051, http://posccaesar.org/rdl/RDS3781591, http://posccaesar.org/rdl/RDS14632057, http://posccaesar.org/rdl/RDS14631081, http://posccaesar.org/rdl/RDS3795871, http://posccaesar.org/rdl/RDS14056321, http://posccaesar.org/rdl/RDS14057236, http://posccaesar.org/rdl/RDS1658280701, http://posccaesar.org/rdl/RDS6786794, http://posccaesar.org/rdl/RDS14296190, http://posccaesar.org/rdl/RDS3788442, http://posccaesar.org/rdl/RDS11471870, http://posccaesar.org/rdl/RDS3797333, http://posccaesar.org/rdl/RDS17545859, http://posccaesar.org/rdl/RDS3796755, http://posccaesar.org/rdl/RDS13266772, http://posccaesar.org/rdl/RDS14304006, http://posccaesar.org/rdl/RDS5724573, http://posccaesar.org/rdl/RDS14633277, http://posccaesar.org/rdl/RDS5713676, http://posccaesar.org/rdl/RDS14284371, http://posccaesar.org/rdl/RDS5725085, http://posccaesar.org/rdl/RDS5731741, http://posccaesar.org/rdl/RDS8640722, http://posccaesar.org/rdl/RDS14619796, http://posccaesar.org/rdl/RDS5718220, http://posccaesar.org/rdl/RDS16698396, http://posccaesar.org/rdl/RDS14631264, http://posccaesar.org/rdl/RDS5740509, http://posccaesar.org/rdl/RDS13651284, http://posccaesar.org/rdl/RDS16698622, http://posccaesar.org/rdl/RDS14389461, http://posccaesar.org/rdl/RDS16212885, http://posccaesar.org/rdl/RDS13267911, http://posccaesar.org/rdl/RDS14618393, http://posccaesar.org/rdl/RDS16213939, http://posccaesar.org/rdl/RDS16213429, http://posccaesar.org/rdl/RDS14634619, http://posccaesar.org/rdl/RDS5715980, http://posccaesar.org/rdl/RDS1469240071, http://posccaesar.org/rdl/RDS16194671, http://posccaesar.org/rdl/RDS14390742, http://posccaesar.org/rdl/RDS13268115, http://posccaesar.org/rdl/RDS16697842