DataCube: owl:DisjointWith in ttl file

Hi,

---
before making my comment: I reacted to the email-address  (
public-gld-comments@w3.org) mentioned on the datacube vocabulary and this
forwarding me to this. I suggest that the introduction is adapted to
reflect the right address.
---


I would like to understand why a MeasureProperty is declared disjoint from
an AttributeProperty and DimensionProperty in the turtle file while this is
not mention in the text of *https://www.w3.org/TR/vocab-data-cube/
<https://www.w3.org/TR/vocab-data-cube/>.*

qb:MeasureProperty a rdfs:Class, owl:Class;
    rdfs:label "Measure property"@en;
    rdfs:comment "The class of components which represent the measured
value of the phenomenon being observed"@en;
    rdfs:subClassOf qb:ComponentProperty;
    owl:disjointWith qb:AttributeProperty;
    owl:disjointWith qb:DimensionProperty;
    rdfs:isDefinedBy <http://purl.org/linked-data/cube>;
    .

I believe that this constraint is too strict.
Datasets can categories the same property differently. So a merge of 2
datasets resulting from 2 different origins is very lickely to violate this
constraint.
It means that the merge is inconsistent and hence requires adaptations
which cannot be done via a machine since the choice of the propertycategory
is business decision.

I would propose to remove the disjoint declarations and move them in a
recommendation section stating e.g. "A coherent QB dataset categorizes each
used component property uniquely into one of the categories Attributes,
Dimensions or Measures."

kind regards,

Bert



-- 
Bert Van Nuffelen

*_____________________________________*

Bert Van Nuffelen

Senior Semantic Technologies and Open Data Architect


phone: +32 (0) 16 31 48 60

mobile: +32 (0) 479 06 24 26

email: Bert.Van.Nuffelen@tenforce.com

skype: bert.van.nuffelen


Havenkant 38, 3000 Leuven

http://www.tenforce.com
*_____________________________________*