Re: Linked Stuff [was Re: RDF's challenge]

On 6/11/13 5:15 PM, Luca Matteis wrote:
> Why are we worried about all of this? "Linked Data" is clearly defined 
> by the four principles of Tim-Berners Lee [1]. RDF is in there. So in 
> order to be Linked Data it has to use RDF.
>
> If you don't want to use RDF, then you're not doing Linked Data. 
> You're just doing something else and you're free to call it whatever 
> you want.
>
> 1. http://www.w3.org/DesignIssues/LinkedData.html

Really?  You are referring to a revision of the original meme [1]. And 
when you digest that meme, please don't come back inferring that TimBL 
must have been thinking about RDF when he produced outlined the four 
points in his original GOLDEN meme.

Look, Data is one thing. Linking is another.

Data is how we express observation.

You express observation via entity relationships. You have a subject 
which characteristics discernible to its observer.

Sets of relationships produce Relations (what the predicate/attribute in 
an RDF or EAV model triple denotes i.e., the thing that has domain and 
range, the thing that facilitates relationships between two things, the 
thing that plays the same role as verb in natural language).

A basic statement (which represents a relationship) of observation is a 
Datum. Each part of the statement (subject, predicate, object) is 
denoted using an identifier.

Identifiers can take the form or literals or references. If you opt for 
literals then your Data (datum collection) is localized to the system 
that understands how to process those literal identifiers. If you opt 
for a Reference, you introduce the possibility of powerful name->locator 
(address) indirection which is still subject to system interpretation 
(e.g., your host OS or an ORDBMS system). If you use HTTP URIs then you 
have name->locator (address) indirection interpretable at Web-scale. 
Basically, you have web-like (or webby) structured data.

So far what do we have?

We simply have web-like structured data representation where the 
components of each Datum are denoted using HTTP URIs. Does that make the 
data meaningful? For instance, could a data consumer injest this 
structured data and perform any kind of reasoning, without human 
intervention or a reference to human-readable document? Could a program 
apply owl:sameAs or owl:InverseFunctional semantics en route to drawing 
inferences from this data?

Now, add RDF to the mix, and all of a sudden you not only have web-like 
structured data, you also have explicit semantics (for the relations) 
baked into the structured data representation.

RDF is about bringing logic to the Web. Basically, Pat Hayes referred to 
this as Blogic [2] in a seminal talk circa., 2009.

Linked Data is fundamentally about making structured data representation 
web-like. RDF is about bringing Logic to web-like structured data. RDF 
and Linked Data are not the same thing, but they work wonderfully together.


Links:

1. 
http://web.archive.org/web/20061121032424/http://www.w3.org/DesignIssues/LinkedData.html 
-- original version of TimBL's Linked Data meme (source of inspiration 
for many early adopters that contributed to the DBpedia and Linked Open 
Data Cloud)

2. http://www.slideshare.net/PatHayes/blogic-iswc-2009-invited-talk -- 
Blogic

2. http://bit.ly/bmdv5N -- Data 3.0 manifesto (I wrote this post in 
response to the retrospective insertion of RDF & SPARQL into TimBL's 
original meme).



Kingsley
>
>
> On Tue, Jun 11, 2013 at 11:04 PM, Kingsley Idehen 
> <kidehen@openlinksw.com <mailto:kidehen@openlinksw.com>> wrote:
>
>     On 6/11/13 4:51 PM, David Booth wrote:
>
>         On 06/11/2013 04:20 PM, Kingsley Idehen wrote:
>
>             On 6/11/13 4:12 PM, Melvin Carvalho wrote:
>
>
>
>                     This is the goal of the Semantic Web: to enable
>                 machines to
>                     usefully and (semi-)automatically, find, share,
>                 combine and
>                     process web data. Because Linked Data is RDF,
>                 Linked Data supports
>                     that goal in a very important way that Linked
>                 Stuff does not.
>
>
>                 We already have the 5 stars of linked data.  If you
>                 use RDF you're
>                 probably 5 star.  If you dont you're probably 4 star
>                 or lower.  That
>                 said, there may be some other linked data system one
>                 day become a 5
>                 star standard.
>
>
>         The stars are to encourage people *toward* Linked Open Data --
>         both Linked Data and fully Open Data.  The stars do *not*
>         indicate that there is such a thing as "one-star Linked Data"
>         or "four-star Linked Data".
>
>
>     That isn't my point.
>
>     My point is that the document provides a nice guideline for moving
>     folks towards Linked Data. It does so without putting RDF at the
>     front-door.
>
>     Again, I am not debating the virtues of RDF. My profound
>     difference with you simply boils down to not seeing the need to
>     inextricably link RDF and Linked Data, at every turn. I have no
>     interest in adding inertia to engagement endeavors when the target
>     audience has no interest in the letters R-D-F. I care much more
>     about the underlying concepts and their utility than I do  labels.
>
>     I am not interested in proving any point or winning any wars
>     around the letters R-D-F. I encourage you to consider doing the
>     same thing. The world fully exploiting the power of the Web is an
>     endeavor achievable without RDF at the front-door.
>
>     Note: not having RDF at the front-door in now way renders it
>     useless or irrelevant.
>
>         Think about it.
>
>
>     I have, for many many years, which is why I am still investing so
>     much time on this subject matter.
>
>
>         Would it make any sense to call a PDF document "Linked Data"
>         just because it is on the web with an open license?
>
>
>     No comment.
>
>         Of course not.
>
>
>     Thank you.
>
>
>         But it would qualify for one star on the path *toward* Linked
>         Open Data.
>
>
>     A Paper Description Format (what PDF is to me) has nothing to do
>     with any kind of openly accessible data, modulo the increasing
>     existence of extractor and conversion tools.
>
>
>     Kingsley
>
>
>
>             Great point!
>
>             The 5-Star Open Data system [1] is a nice approach to
>             framing this most
>             challenging of narratives. It's greatest virtue is not
>             putting RDF at
>             the front-door :-)
>
>
>             Links:
>
>             1. http://5stardata.info/ -- 5-Start Open Data
>
>
>         That is *Open* Data -- not *Linked* Data.  When you reach all
>         five stars it becomes both: Linked Open Data.
>
>         David
>
>
>
>
>     -- 
>
>     Regards,
>
>     Kingsley Idehen
>     Founder & CEO
>     OpenLink Software
>     Company Web: http://www.openlinksw.com
>     Personal Weblog: http://www.openlinksw.com/blog/~kidehen
>     <http://www.openlinksw.com/blog/%7Ekidehen>
>     Twitter/Identi.ca handle: @kidehen
>     Google+ Profile: https://plus.google.com/112399767740508618350/about
>     LinkedIn Profile: http://www.linkedin.com/in/kidehen
>
>
>
>
>
>


-- 

Regards,

Kingsley Idehen	
Founder & CEO
OpenLink Software
Company Web: http://www.openlinksw.com
Personal Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca handle: @kidehen
Google+ Profile: https://plus.google.com/112399767740508618350/about
LinkedIn Profile: http://www.linkedin.com/in/kidehen