Re: ISSUE-66: LinkedDataT

Hello everybody,

My two cents on RDF vs. Linked Data
-------------------------------------------------------

I have a dream and this is the best community to create a new MO when
writing specs for this dream to come true: formal defined unambiguous RFCs.
Parenthesis mean nothing.

Ok, Linked Data (and Semantic Web) is more of a vision/buzzword than a RFC,
but it is interesting how loose definitions go a long way, confusing
possible adopters.

I completely agree with Erik about merging communities rather than pitching
technology.

I believe listening to REST community will be nice for SemWeb people, a
community that already has some kind of conflict between "AI/Logics people"
vs "Web people".

Survey on Hypermedia APIs
-----------------------------------------

I will try to inspect several specs for hypermedia/HATEOAS, such as HAL,
JSON schema and JSON-LD and try to define some dimensions to evaluate them:
ease of use, expressiveness vs orthogonality, serialization (clearly a big
win for JSON here), level of adoption, and so on.

This is an ongoing work in its *first baby steps*.
BTW, I will love to hear some feedback about it.
Maybe there are some specs I'm missing, for example.

*Disclaimer*
I'm trying to focus on recent specs based on JSON and HATEOAS.
I believe that's what Markus calls as 3rd generation Web APIs in his
dissertation [1].

[1]
http://www.markus-lanthaler.com/research/third-generation-web-apis-bridging-the-gap-between-rest-and-linked-data.pdf


On Tue, Aug 5, 2014 at 5:54 PM, Markus Lanthaler <markus.lanthaler@gmx.net>
wrote:

> On 5 Aug 2014 at 17:34, Ruben Verborgh wrote:
>
> > So I tried to start the rewrite in another branch, but I got stuck.
> > (https://github.com/HydraCG/Specifications/commit/1d22c5360)
> >
> > I tried to follow this:
> >
> >> Maybe it would be more straightforward to explain it the other way
> round:
> >>  - documents are in natural language
> >>  - machines are bad in understanding natural language
> >>  - machines prefer structured data using unambiguous identifiers
> >>  - the Web uses URLs* as identifiers
> >>  - RDF allows data to be expressed in a machine-processable way by
> >> leveraging URLs
> >>  (- RDF expresses data in the form of triples) -- could be omitted
> >>  - RDF can be serialized in various formats such as JSON-LD, HTML+RDFa,
> >> orTurtle
> >
> > I got stuck, because I wanted to arrive at "Linked Data"
> > but I end up at "RDF".
> > At the moment, it seems the old direction was clearer.
>
> What about
>
>    "Machines prefer structured data using unambiguous identifiers,
>     and on the Web, these identifiers are URLs."
>
> ... Linked Data combines these two things* to make it easier for machines
> to
> process and integrate data from different sources. RDF formalizes a
> triple-based data model that can be used to publish Linked Data in various
> formats such as ...
>
> * using a better term for "things"
>
>
> >>>>>   By publishing Linked Data [LINKED-DATA],
> >>>>>   we enable automated clients to consume information.
> >>
> >> [.]
> >>
> >> If it doesn't add anything, just leave it out.
> >
> > This sentence (at the moment) introduces Linked Data,
> > so it is quite crucial. How about:
> >
> >   By publishing Linked Data [LINKED-DATA],
> >   enable automated clients can interpret information.
> > maybe augmented with
> >
> > , by relying on URLs as universal identifiers.
>
> See my suggestion above. I'm also fine with your proposal here.
>
>
> HTH,
> Markus
>
>
> --
> Markus Lanthaler
> @markuslanthaler
>
>
>


-- 
Atenciosamente,
Ícaro Medeiros.

--------------------------------
(+55 21 2483 6928 / +55 21 98327 0069)
Senior Software Developer @ Globo.com (Semantics Team)
PhD Candidate @ *Pontifical Catholic University* of Rio de Janeiro (
www.inf.puc-rio.br)
MSc @ Federal University of Pernambuco (www.cin.ufpe.br)
BSc @ Federal University of Alagoas (www.ic.ufal.br <http://www.ufal.br/>)
-| @icaromedeiros <http://twitter.com/icaromedeiros> |
github.com/icaromedeiros |-

== Scientia Vincere Tenebras ==