I've been discussing this issues: the transition from a 'document oriented' web into a 'data and services oriented web' for a while in the semantic-web related w3c mailing lists. In short it seems we had have to deal with building 'applications' over *documents* for a long time and it seems it does not need to be the the same within a data oriented web (at least for me).

However, applications does not build by themselves. XML and others promised interoperability since a long time before RDF and, although I belive RDF / SW could, there will be no interoperability (as that of RESTFul services) until there is no agreement in 'semiotic' semantics of representations, resources and protocols.

The document I'm attaching here is an extended version of yesterday's abstract but it only states a few concepts regarding integration (alignment and merge of datasources. I'm not a seasoned SW nor integration developer).

On Aug 21, 2017 12:36 PM, "Michel Kern" <echopraxium@gmail.com> wrote:

Hello Sebastian

I share your concerns, even though (or precisely because) I'm not a 'referent' in Application and Data storage interoperability, I'm a seasoned SW engineer (1). My feedback on your article, is that the use case illustrates the challenges and needs for a web of interoperables solutions (maybe you could mention SKOS among the Semweb languages)

My interest for interoperability started in 1997, at this time there was CORBA and Microsoft COM, then regarding RDBMS, there was ODBC/JDBC. There was also all the vendors buzz like SAP pretending to teach esperanto to the babel towers of legacy apps (but in fact replacing backoffice apps by theirs)

Some time XML and its modelling 'applications' (XML Schema and XMI) gave me some hopes, but still no concrete sign of 'user driven apps' built by a dynamic process of looking for and assembling 'autodescribed components' (SOAP and WSDL was then a kind of CORBA successor)

Before the Semantic Web,

I had never really 'bought' the idea that SQL based Data Storage was considered as a 'Silver Bullet', ORM solutions seemed to bring even more entropy.

But with the advent of Graph Databases (e.g. Neo4J) and this brilliant solution (Apache MetaModel) for Data Storage interoperability, I am at last more comfortable and comfortable with the idea that the solutions are available for designing interoperable solutions in an elegant way at last.

I am most fluent in the DataViz and ETL dimensions of Semantic networks. I authored a mindmapping app in 2004: Thinkgraph, now I'm prototyping a solution to provide interoperability between 'knowledge visualization' (e.g. TheBrain, XMind), Data Storage and Curating platforms (e.g. Pinterest, Evernote, Youtube, etc..).

I chose .Net/C# as my development workshop thats why I started a port of Apache MetaModel in C#.

Best Regards

Michel Kern

1. Dassault Systèmes 1991-2000, Nokia 2001-2010