Re: Schema.org and OWL

Dear Anthony:

I am not sure where this discussion is heading to. Schema.org is not an attempt to create a logically consistent model of the world, but instead to provide a shared conceptual structure that facilitates data interoperability between Web content and applications that consume this data. This problem has way more dimensions than logical consistency. 

You may want to read the following paper that sketches a few core paradigms of schema.org:

    https://queue.acm.org/detail.cfm?ref=rss&id=2857276

Please do not get me wrong, but trying to propose a big reorganization based on the search for "one true logical model of the world" is not likely to succeed in here. Many people who worked on schema.org have actually been working with Cyc, read the OntoClean works, and been exposed to a lot of the academic debates about ontology engineering.

But schema.org is much more than an ontology in the classical CS sense. Schema.org is surely not perfect, but many of the current parts may seem inconsistent at first but were not designed that way due to lack of expertise or dedication to detail.

For instance, it really matters whether the conceptual distinctions the model provides can be understood by the broad users of schema.org, and that they are easy to populate from existing databases. Optimality for something like schema.org is to be measured by the actual amount and quality / usefulness of data on the Web, from the perspective of its major consumers.

Martin

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




> On 15 Jun 2018, at 00:17, Anthony Moretti <anthony.moretti@gmail.com> wrote:
> 
> Cheers Thad, anything you'd recommend I read to learn more?
> 
> I think Martin's point about passing information from product types to product instances can be addressed higher in the hierarchy than Product actually. I sense people are opposed to shifting properties from more specific types to Thing though (maybe I don't understand something, can someone please explain that to me?) My view is that using overly specific domains for properties causes strange entailment, e.g. in its current form the "height" property entails the subject is either a MediaObject, Person, Product, or VisualArtwork, which doesn't seem right.
> 
> A more general solution, which might benefit other parts of Schema, might look like the following:
> 
> Thing
>     height
>     width
>     depth
>     weight
>     material
>     color
>     predecessorOf
>     successorTo
>     variantOf
>         Product
>             offers
> 
> ThingType
>     heightOfTypicalInstances
>     widthOfTypicalInstances
>     depthOfTypicalInstances
>     weightOfTypicalInstances
>     materialOfTypicalInstances
>     colorOfTypicalInstances
>     predecessorTypeOf
>     successorTypeTo
>     variantTypeOf
>         ProductType
>             offersOfTypicalInstances
> 
> This isn't complete obviously, and I understand how big a change it would be, happy if it just stimulates other ideas I guess.
> 
> On Martin's other point about distinguishing individual products from groups of products, I don't think IndividualProduct is required once SomeProducts is removed from beneath Product (where it incorrectly sits now) - once you do that all Products are individual products unless additionally typed as SomeProducts.
> 
> When I first came across Schema the Products part was the hardest part to grasp because it doesn't follow the rest of Schema, it has its own specialized logic. The following structure could possibly be a remedy:
> 
> Product
> ProductType
> SomeProducts
> 
> I understand the magnitude of the change practically, but from a logical point of view does anybody see something this structure would miss?
> 
> Anthony
> 
> On Thu, Jun 14, 2018 at 2:50 PM Martin Hepp <mfhepp@gmail.com> wrote:
> Not sure I understand what you want to say in here.
> 
> > http://schema.org/model
> 
> is the relationship type / property for materializing the link between a product and its product model. So clearly, this relationship type is a property, not a class / entity type. But the original discussion was referring to schema:ProductModel, which is the class / entity type.
> 
> Martin
> 
> -----------------------------------
> martin hepp  http://www.heppnetz.de
> mhepp@computer.org          @mfhepp
> 
> 
> 
> 
> > On 14 Jun 2018, at 19:03, elf-pavlik@hackers4peace.net wrote:
> > 
> > On 2018-06-14 07:46, Thad Guidry wrote:
> >> Hi Martin !
> >> "no, ProductModel is not an attribute, it is a class in its own
> >> right."
> >> Yes, that's what our current modeling simplifies to since it borrowed
> >> the top level from RDF.
> >> And that's where you have your opinion and others politely disagree
> >> and have their own opinion, that Model is an attribute or trait of the
> >> type Product.  But that is fine !  We have dealt with it and things
> >> are published now. No worries!
> > 
> > http://schema.org/model
> > 
> > "The model of the product. Use with the URL of a ProductModel or a textual representation of the model identifier. The URL of the ProductModel can be from an external source. It is recommended to additionally provide strong product identifiers via the gtin8/gtin13/gtin14 and mpn properties."
> > 
> > Values expected to be one of these types:
> > ProductModel
> > Text
> > 
> > Used on these types:
> > Product
> > 
> 
>

Received on Thursday, 14 June 2018 22:38:32 UTC