Semantics

[almereyda]

It seems prevailent that the Semantic sujet arrives within an uncertain timeframe.

Wiki already has elements with implied semantics:

• The journal by its model.
• The story by its serialization S( S( S(protocol) - V(://) - O(FQDN) ) - V(view) - O(welcome-visitors) ) - V(FQDN) - O(welcome-visitors)
• REST-API? I have to admit, I don't even know if it exists.
• Authentication : OAuth / Persona

Which other ontology domains could you already identify?

[WardCunningham]

Tim Berners-Lee, the inventor of the Web and Linked Data initiator, suggested a 5 star deployment scheme for Open Data. http://5stardata.info/

Early on I self-assessed decisions I had made in federated wiki and gave myself three stars.

Later I discussed my interpretation of the five stars with a hangout visitor that was steeped in the theory behind the semantic web. He argued that I really did deserve all five stars. I thought his standards for stars were pretty low. But then I assumed I would be judged by the schema-heavy existing work in the semantic web.

I'm not quite getting the notation you're using in this issue. Let's continue the discussion in this issue and see if we can tweak federated wiki into a new path to semantic data.

[almereyda]

schema-heavy Yes, let's start there. Slowly. Because from my view on the Semantic Web, it's ontology design is still quite an arguably criticisable experts system.

Nice anecdote, thanks for providing the 5 star context to TBL and sorry again for the improvised triple notation. Sure Turtle and N3 would make that more clear (to the experts :wink: ).

What is a non-experts web anyway? The Social Web(TM*)? I doubt the term will not be coopted again, like so many before.

* not to be taken seriously; no pun intended.

[almereyda]

@coevolving is talking of Graphs for knowledge representation, which makes me think about a thin JSON-LD layer for federated wiki again: As its data model is already purely JSON-based, it should be simple to write a very minimal vocabulary, in JSON-LD's language a @context, which explains the keys of a given dataset, here a fedwiki page, a sitemap or an export. Further, a dedicated plugin could take care of injecting neccessary links, like @ids and @contexts, if a site owner decides to let it do so.

http://search.fed.wiki.org:3030/ gives us an idea about which items we may want to identify and publish (as seperate/# URLs) in a linked data format and thus have to be covered by a vocabulary definition of their meaning, itself expressed in RDF. Because then, by reusing existing terms and expressions, we understand wiki is not the place to publish everything from anew, but aggregating into the federation from a bigger web of data, in the same time contributing back to it without violating its autonomous architecture.

In a future with practical uses of the linked data publishing format we can imagine site owners to publish very diverse data via their loosely coupled wiki interfaces. Then we are all ready to contribute our part to the Linked Open Data cloud.