Excellent response, Kingsley!
On 11/11/2015 6:37 PM, Kingsley Idehen wrote:
On 11/11/15 3:49 PM, Ruben Verborgh wrote:
Hi Kingsley,
Some valid points. Two quick remarks:
For me, the Semantic Web vision has always been about clients.
I think the "Semantic Web" has always been about "The Web" (clients and
servers) :)
Of course—but the emphasis in the community has mostly been on servers,
whereas the SemWeb vision started from agents (clients) that would do things
(using those servers).
Now, the Semantic Web is mostly a server thing, which the Google/CSE example
also shows.
Okay, I certainly agree with that observation. Too much emphasis on
servers and large datasets has starved the crucial need for
collaboration on the client side.
Areas of starvation include:
1. Bindings various UI/UX frameworks to data access controls capable of
handling JSON-LD, Turtle, RDFa etc..
2. Constructing sophisticated data access controls that simplify Linked
Data exploitation by client-centric developers.
Collaboration taking shape on the Javascript front re. rdflib.js,
rdfstore.js, SoLID, and RWW in general etc.. are great examples of
movement in the right areas (IMHO).
At the moment, consuming seems only within reach of the big players,
who have the capacity to do it otherwise anyway.
No, you can craft a CSE yourself right now and triangulate searches
scoped to specific entity types.
Do you mean making a CSE through the Google interface?
Google offers CSEs as a kind of service. If you leave said service with
Google trimmings there's no cost. If you seek to remove Google trimmings
then they charge a fee. Either way, that's fair enough in my eyes.
But then I'm actually querying the Google servers, not the Web…
Google is a major Web hub, via CSEs you can find pathways to other
places on the Web. What useful about these CSEs is that they return a
boatload of documents that include RDF based structured data [1].
Then intelligence is with a centralized system, not between clients and servers.
Google is just one of many hubs from which RDF documents can be
discovered and access.
Not yet the Semantic Web for me.
I the "Semantic" and "Web" components of the meme breakdown as follows,
in my experience:
1. Semantic -- structured data endowed with machine- and human-readable
relationship type semantics.
2. Web -- hyperlinks functioning dually as mechanism for entity
denotation and connotation (i.e., names resolve to RDF Language based
descriptor documents).
Best,
Ruben
