Re: Schema.org version 6.0 - Schema.org Adapter

Happy new year, folks!

I wanted to share our new npm-library with you: the Schema.org Adapter 
(https://www.npmjs.com/package/schema-org-adapter).

It is a library that takes the schema.org vocabulary as input and offers 
an API for its vocabulary terms. It has built-in reasoning, which means 
that parameters can be passed to the API to enable reasoning on the 
queried vocabulary terms (e.g. resolution of properties, sub-classes, 
ranges, etc.). This library can be used to empower websites/tools that 
work with schema.org, like annotation editors/validators.

I am happy to hear that there will be some attention regarding the data 
model of schema.org for v7.0 (especially enumerations), since a clear 
data model is crucial for us developers to build software around the 
vocabulary 
(https://github.com/semantifyit/schema-org-adapter/blob/master/docu/dataModel.md).

Sinc. Omar

On 22.01.20 12:47, Dan Brickley wrote:
> I agree that it is problematic. I spent some time yesterday digging 
> around and my conclusion is that it would be best to back out the 
> enums change and ship the rest of V6 than to rush a redesign.
>
>  The core problem is that some terms continue to be  effectively 
> operating as types, but that information was removed from the schemas. 
> This in part came from UI assumptions in the schema.org 
> <http://schema.org> site codebase, which wasn't built for terms that 
> play multiple roles.
>
> That situation will need addressing (alongside other aspects of the 
> enum cleanup) but let's get v6.0 out, and give our technical debts 
> some serious attention for v7.0
>
> Dan
>
> On Tue, 21 Jan 2020, 23:51 Eyas Sharaiha, <eyas@google.com 
> <mailto:eyas@google.com>> wrote:
>
>     Is it possible for
>     https://github.com/schemaorg/schemaorg/pull/2443 to be merged
>     before this release is published?
>
>     I commented on there, but it seems like some of the enum
>     rationalizations break some assumptions that tooling currently
>     makes. I.e. DrugClass is now an enum value type only (has property
>     (rdf:type sdo:MedicalEnumeration) and no other rdf:type) but is
>     used as a type name (e.g. sdo:drugClass sdo:domainIncludes
>     sdo:DrugClass). Previously, any object described by domainIncludes
>     would be either an rdfs:Class or a DataType.
>
-- 
Omar J. Holzknecht, MSc
Semantic Technology Institute
University of Innsbruck

ICT - Technologie Park Innsbruck
Technikerstrasse, 21a
6020 Innsbruck
Austria