On Wed, May 28, 2014 at 10:37 AM, Daniel Kinzler <
daniel.kinz...@wikimedia.de> wrote:

> More fundamentally, they are semantically different: an item describes a
> concept
> in "the real world", while a property is a structural component used for
> such a
> description.

As I perceive it, a property is a normal item (concept) imbued with the
option to use it as predicate and allow it to use different datatypes.
There is no property that cannot be expressed as an item, even properties
that represent an identifier, they also could be said that they are a
concept in the real world.
I understand that from the software side you need to make a difference
between "basic concepts" (items) and "concepts that can be used as
predicates" (properties). From the community side we also need to
scrutinize and "rinse" the concepts that hide behind the words before using
them as predicates, but sometimes it is good to stop and consider what are
we really doing.

Yes, properies are simmilar to data items, and in some cases, there may be
> an
> item representing the same concept that is represented by a property
> entity.

I haven't found yet a property that couldn't be expressed as an item.

> I don't see why that is a problem, while I can see a lot of confusion
> arising from
> mixing them.
 It is not a problem now but I considered interesting to analyze what is
the substance of the distinction. If properties and concepts are separate
in the end we will be reproducing their ontological structure when
organizing them. So then it might not make sense to use "subproperty of" to
organize properties, but just "corresponds to item".

Gerard, thanks for bringing the example of OmegaWiki, it is interesting
that two independent communities came to the same thoughts without any
contact between them :)

Wikidata-l mailing list

Reply via email to