Re: relating IndividualProduct to Product

> sounds to me link we go again into
> "why Vehicle subClassOf Product ? (also: Commercial, Economic)"
> https://lists.w3.org/Archives/Public/public-vocabs/2015Mar/0153.html

Note that it is conceptually fine to put a few typical types of goods below schema:Product, because formally, this just means the intersection of the type that describes the nature of the object (e.g. being a Car or Book) and the type that reflects the role of something being object of (or potentially object of) an offer to transfer some rights on this thing for some compensation.

Being a product in schema.org and GoodRelations does not introduce a lot of ontological commitment - just that the entitiy could be the subject of an offer (so it could be called "TradeableObject").

As long as you do not need the type that describes the nature of the object (e.g. being a Car or Book) explicitly without the additional notion of being potentially (!) the object of an offer, there is no real problem.

Thus, it is not really an inconsistency that a schema:Book is not a subtype of schema:Product, whereas schema:Car is - there are many use cases with bibliographical data that are very unrelated to offers - e.g. citation data. Cars are more likely to be modeled in the context of offers that without that context. It would not harm to have book as subtype of schema:Product, but I understand that people in the library community see it differently.

On the contrary, there are two important advantages of the current approach:

1. You save the need for multi-typed entities in all cases where the thing is used as a product. That is a good thing if the majority of information uses those objects in the context of offers.

2. Developers will find the type more easily if they consider the objects as products - i.e. if the schema.org taxonomy matches their mental models.

#2 should not be underestimated - we could run a controlled experiment, and I bet Web developers will find schema:Car much faster below schema:Product than as a direct subtype of schema:Thing (or an "ontologically-shaped" subtype like "schema:TradeableThing".

I would like to stress again that having specific types of things as subtypes of schema:Product is not even a formal / conceptual inconsistency as long as you do not need the type explicitly without the potential product role.

Also note that without a disjointness axiom in the schema or in the data, the difference between the two modeling approaches is meaningless - a schema:Thing could, under open world assumption, always be a schema:Product, so it is a "potential product", same as an object of type schema:Product. 

Best
Martin

-----------------------------------
martin hepp  http://www.heppnetz.de
mhepp@computer.org          @mfhepp







> On 12 Apr 2015, at 16:26, ☮ elf Pavlik ☮ <perpetual-tripper@wwelves.org> wrote:
> 
> On 04/12/2015 04:16 PM, jean-delahousse wrote:
>> Hello,
>> I am still quite confuse. In my use case I have the dealer point of view
>> describing used cars in adds.
>> My idea was to use schema:Car, and hope the extension you proposed at
>> http://sdo-property-value-and-cars.appspot.com/Car could be so I would be
>> able to give more information about the mileage, the engine... of this
>> individual used car.
>> I understand you propose to use IndividualProduct, but then how will I be
>> able to use the very useful properties and class I find in
>> http://sdo-property-value-and-cars.appspot.com/Car ?
> sounds to me link we go again into
> "why Vehicle subClassOf Product ? (also: Commercial, Economic)"
> https://lists.w3.org/Archives/Public/public-vocabs/2015Mar/0153.html
> 
> and gives another example that
> Product/IndividualProduct/ProductModel/SomeProduct might work better
> when used as additional type.
> 
> "@type": ["Car", "ProductModel"]
> 
> "@type": ["Book", "IndividualProduct"]
> 
> etc.
> 
> https://github.com/schemaorg/schemaorg/issues/404
>

Received on Sunday, 12 April 2015 18:23:21 UTC