Re: The Battle for Linked Data

On 3/26/12 11:49 AM, Hugh Glaser wrote:
> So What is Linked Data?
> And relatedly, Who Owns the Term "Linked Data"?
> (If we used a URI for Linked Data, it might or might not be clearer.)
>
> Of course most people think that "What *I* think is Linked Data is Linked Data".
> And by construction, if it is different it is not Linked Data.
> Kingsley views the stuff people are talking about that does not, for example, conform to a policy that includes Range-14 as "Structured Data" - naming things is important, as we well know, and can serve to separate communities..

Yes!

Schema.org, webdatacommons.org etc.. are examples of Structured Data. A 
good thing for the Web circa. 2012.

When structured data is published to the Web in a manner that's 
oblivious to Name/Address (Reference/Access) disambiguation we shouldn't 
refer to that as Linked Data. Doing so simply compromises the Linked 
Data meme and its associated best practices.

>
> There are clearly quite a few people who would like to relax things, and even go so far as to drop the IR thing completely, but still want to have the Linked Data badge on the resultant Project.

Yes!

And that's were many of us have a problem and draw a serious line in the 
sand.

The World Wide Web is a dexterous layer of abstraction over the 
Internet. It's always been a collection of structured data bearing 
resources. What's changing over time is the granularity of its structure 
and the semantics that underpin its *relations* based tapestry. 
Basically, hyperlink based webby-ness is an evolving thing, so to speak.

> There are others for whom that is anathema.
Yes!

>
> I actually think that what we are watching is the attempt of the Linked Data child to fly the nest from the Semantic Web.

Hmm.  Linked Data is (IMHO) ground zero for the Semantic Web. That said, 
you need structured data appreciation and propagation to create full 
context for this mercurial journey. The allure of TimBL's meme shouldn't 
detract from the virtues of Web scale structured data where relation 
semantics serve needs beyond distinguishing Presentation and Content, 
which is where Web 2.0 basically stops. We now have Presentation, Coarse 
Grained Structured Data, and Fine Grained Structured Data (aka. Linked 
Data) emerging from what used to simply be referred to as Content.

> Can it develop on its own, and possibly have different views to the Semantic Web, or must it always be obedient to the objectives of its parent?

Its a critical conduit to the Semantic Web.

The journey (as I see it) looks something like this:

1. Web of Documents - where relation semantics focused primarily on 
presentation
2. Web of Documents - where relation semantics distinguish between 
presentation and content (the old XML value proposition in a nutshell)
3. Web of Data -- where relation semantics enable emergence of 
coarse-grained structured data
4. Web of Linked Data -- where relation semantics enable emergence 
fine-grained structured data (aka. Linked Data)
5. Web of Semantically Linked Data - where relation semantics enable 
emergence of structured knowledge (some might say this is the reasoning 
friendly Web).

Each level provides appreciation context for what follows. Thus, they 
are all really valuable components of a virtuous evolving system if we 
could just resist the tendency to conflate matters simply because a 
specific meme has demonstrated momentum.

>
> Often the objectives of Linked Data engineers are very different to the objectives of Semantic Web engineers.
> (A Data Integration technology or a global AI system.)

Yes.

> So it is not surprising that the technologies they want might be different, and even incompatible.

Yes.

>
> If I push the parent/child analogy beyond its limit, I can see the forthcoming TAG meeting as the moment at which the child proposes to reason with the parent to try to reach a compromise.
> The TAG seems to be part of the ownership of the term "Linked Data", because the Linked Data people (whoever they are) so agree at the moment - but this is not a God-given right - I don't think there is any trade- or copy-right on the term.

But we need to know what we are doing. What rules apply to what system. 
The WWW is superbly multidimensional due to its underlying architecture.
> A failure to arrive at something that the child finds acceptable can often lead to a complete rift, where the child leaves home entirely and even changes its name.

Sometimes you have to help the child understand sibling-hood :-)
>
> And of course, after such a separation, exactly who would be using the term "Linked Data" to badge their activities?

Thing is that "Linked Data" is just another term for the Web as 
originally envisioned [1]. The Web is here forever so monikers such as  
"Linked Data" or "The Semantic Web" don't ultimately matter that much. 
Their prime purposes (IMHO) are really about aiding exposure of the 
Web's multidimensional nature.

>
> Like others in this discussion I am typing one-handed, after earlier biting my arm off in preference to entering the Range-14 discussion again.
> But I do think this is an important moment for the Linked Data world.

Amen!

Links:

1. http://www.w3.org/Talks/2001/12-semweb-offices/timbl-1989-image.gif 
-- the vision diagram
2. http://www.w3.org/Talks/2001/12-semweb-offices/slide8-0.html -- early 
steps toward vision
3. http://www.w3.org/Talks/2001/12-semweb-offices/slide13-0.html -- its 
all about hyperlinking.


>
> Best
> Hugh


-- 

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