Microdata provides a wide range of data type definition that can be implemented in a cross-platform, cross-architecture and cross-pretty-much-everything fashion. Since the type definitions are data oriented and give semantics the priority, the current collection (and since is an ongoing effort, the growing collection) of types provided by microdata cover the needed definitions for standard property names and property types for much of the social networking universe entities.
My proposal is to modelate the system by implementing in each tear standard types from the microdata vocabulary. It addresses a series of everyday issues when dealing, as we are, with abstractions of entities that come from quite different sources and provide an also different share of attributes, while maintaining a set of abstract types such as Article, Image, etc:
- Interoperability: schema.org microdata provides a unified standard for property names.
- Simplicity: since the schema.org microdata was designed for the semantic web, both types and property names address conveniently most common entities we need to deal with.
- Future-friendliness: schema.org and microdata in general are ongoing efforts; implementing its standard practices opens the door to yet-to-be-known implementations of the architecture.
- Extensibility: it is perfectly possible to define our own vocabulary whenever microdata comes short, while reusing the exact same share of tools.