If the network exposed an API to make data uniformly shareable, machine-readable, and searchable, what definitions would it agree on, how would would new definitions be composed, and how would the integrity of the definitions be maintained?
Specialised and tightly-controlled semantic schema may be more about addressing these questions and Uncertainty Reasoning for the World Wide Web.
Reading about Gene Ontology project, which is to
provide a uniform way to describe the functions of gene products from organisms across all kingdoms of life and thereby enable analysis of genomic data.
It makes sense that Gene Ontology is overseen by a trusted council which addresses the above questions and also
colloquialisms, community preferences, abbreviations, legacy names, the multiple ways of referring to chemical elements, capitalization, and all the possible variations that occur in natural language.
It seems to make more sense that each domain and associated community like gene research, astronomy, video media, audio media, etc, is more equipped in their specialised knowledge to oversee the structure and maintenance of their data semantics than the whole of the SAFE or other network.
One option is to align with http://schema.org/, porting the current structures and definitions over to the network, and being involved with the Schem.org Community Group
Although I read in this issue of Scientific American a piece of information that gave me the impression that the semantic web could benefit more from disparate schema, however comparable data, where something called
inference engines could be used to find the keys that connect disparate schema:
Finding relations among different sources is an important step toward revealing the “meaning” of information.
What this might mean for the network is focusing on providing the essentials for building semantic schema but not committing to any in particular.
The last question I’m thinking about is how to scale the composition of RDFS.
For example, the simple act of a person uploading a video. How is the raw data annotated to conform with RDF? Is the user prompted to add standardised tags?
Seaching around Apache Jena and Stanford’s Protégé to find how they may address this problem.
When uploading videos to YoutTube, one manually enters tags to describe the video, but again, at what layer do we concern ourselves with this tagging? Is
Schema.org concerning itself with video descriptions?