Linked Data and RDF, some perspective

All,

I posted this comment to the RDF comments group earlier today, in 
retrospect I should have cc'd this list. Anyway, here is an account of 
the problem and audience profiles re. RDF and Linked Data.

I have but one goal here. Getting opening up civil debate that enables 
you address this nagging problem.

My post (minor edits e.g., fixing the links index at the end);

To qualify a few things about statements such as RDF != Linked Data in 
association with de-reference, I would like to clarify a few things.

Web-like Structured Data Representation:

TimBL's meme outlined a principled approach structured data 
representation that results in a Data Web or Web-like structured data, 
courtesy of HTTP URIs. This approach makes the Web-like structured data 
scale to the expanse of the scale-free Web.

In his original meme [1] he indicated that this principled approach 
enables one to look-up what an HTTP URI denotes, while also indicating 
to publishes that useful information should be accessible from the 
look-up location.

In the revised meme [2] he added "using standards (RDF, SPARQL). This 
introduced the problem of using words that mean different things to 
different audiences which I break down as follows (un-pejoratively):

1. RDFers -- RDF and Linked Data (this thing with appreciative momentum) 
are now inextricably linked, we'll show them now!

2. RDF-Refluxers -- Linked Data is just a re-branding of RDF, they think 
we are stupid!

3. Independents -- WTF! (pardon my French, but I want to signal as 
effectively as possible in this response).

In reality, bearing in mind my proximity to TimBL re. these matters of 
Linked Data, I genuinely believe he meant (but of course I do not speak 
for him):

Use standards such as RDF and SAPRQL as an effective (productive) way to 
provide really useful information when HTTP URIs (as outlined in this 
note) are looked-up. In reality, when you get round to implementing 
Linked Data (as a publisher) all you have to do is redirect user agent 
URI look-up requests to a SPARQL protocol URL which leaves the heavy 
lifting to a SPARQL processor (which may or may not be a full blown DBMS 
engine).

As for this whole JSON-LD and RDF affair, there is one subtle detail 
that makes matters more challenging. JSON-LD is seeking to be published 
as a deliverable from the RDF group. In taking the aforementioned route, 
many of the RDF related push-backs become much more understandable.

Anyway, here are some links for additional context re. my comments re. 
Linked Data as Web-like structured data:

1. http://www.w3.org/DesignIssues/diagrams/history/proposal-fig1.gif -- 
illustration from original Web design document (this is clearly 
depicting a Data Web woven together via the typed relations illustrated 
as connections/connectors)

2. http://www.w3.org/2005/Talks/1110-iswc-tbl/#(4) -- that's Linked Data 101

3. http://www.w3.org/2005/Talks/1110-iswc-tbl/#(7) -- URIs + HTTP (this 
is makes Data Web-like)

4. http://dig.csail.mit.edu/2007/Talks/0511-tab-tbl/#(10) -- Linked Data 
(again, clearly defined and distinct from RDF).

Inserting Logic into Linked Data (i.e., taking it from our "minds eye" 
to a Data Web accessible to humans and machines) is where RDF kicks in, 
with aplomb.

Unfortunately, due to issues associated with OWL misconceptions 
compounded by RDF/XML dominating OWL examples, many RDFers are reluctant 
to utter the words "inference" or "reasoning" since they assume those 
are the issues that scare folks. A current example of that is easily 
discernible form some of the longer threads on the W3C's LDP [3] list.

Pat Hayes gave a very nice presentation on Blogic [4][5] that puts this 
issue of Logic and Data Webs in scope.

Links:

[1] 
<http://web.archive.org/web/20061201121454/http://www.w3.org/DesignIssues/LinkedData.html> 
-- original Linked Data meme (circa. 2006)
[2] <http://www.w3.org/DesignIssues/LinkedData.html> -- revised meme 
(circa. 2009)
[3] 
<http://lists.w3.org/Archives/Public/public-ldp-wg/2013Jun/0132.html> -- 
sampling of push-back to RDF push-back from a recent Linked Data 
Platform thread (i.e., the semantics are what matter and this is what 
RDF is all about)
4. http://slidesha.re/18CtxGK -- Blogic
5. http://videolectures.net/iswc09_hayes_blogic/ -- What's in a Link?

-- 

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

Received on Thursday, 13 June 2013 17:50:41 UTC