« Linked Data Services | Main | HTML5 Classification »

Thursday, July 05, 2012


Feed You can follow this conversation by subscribing to the comment feed for this post.

Mike Kelly

I'm confused. "Linked Data" is a very generic sounding name, but apparently it requires the use of RDF - is that correct?


@mikekelly85, "linked data" sounds generic, but the community defines it to be RDF-only. http://dret.typepad.com/dretblog/2009/11/the-linked-data-police.html happened a while ago when i used the term more loosely in a paper we published, and since then, i have made the distinction between "linked data" (the RDF way of doing things) and "data that's linked" (the REST way of doing things). obviously, it would be really good to bridge these two worlds, which so far are separated, and this blog post is an attempt to look at how those bridges could be built, focusing on what's already out there as established technologies and architectures, and what's missing in the current disconnected landscape.


I don't know how you've arrived at the conclusion that Linked Data is RDF-only. It isn't. Conflating the concept of Linked Data and RDF is the fundamental flaw retrospectively inserted into TimBL's original Linked Data meme. This strange marketing anomaly remains the biggest obstacle to Linked Data comprehension and adoption.

You can give *things* names that resolve to descriptor documents (resources) via indirection without any RDF in the mix. Remember, RDF is about using IRIs for denoting (naming) *things*. It says or claims nothing about use of indirection (implicit or explicit) to identify documents (resources) that describe what denoted.

Linked Data is fundamentally about use of indirection to associate an entity (real-world or otherwise) with a resource that bears its description via resource content.


1. http://bit.ly/M7Rcka -- Web 3.0 Manifesto re. Structured Data

2. http://uriburner.com -- example of a Linked Data service that produces Linked Data from a plethora of data sources while also offering said data in a myriad of structured data formats constrained by the entity-attribute-value model enhanced via incorporation of de-referencable URI based hyperlinks.



we've published about Linked Data Services (LIDS) at ESWC 2011:

Sebastian Speiser, Andreas Harth. "Integrating Linked Data and Services with Linked Data Services". 8th Extended Semantic Web Conference (ESWC 2011), Heraklion, Greece. [1].

The basic idea is to pass parameters in the query part of the URI. LIDS then return RDF. As such, our LIDS look to the outside just like standard Linked Data resources.

LIDS also have descriptions using basic graph patterns. Similar kinds of descriptions of input and output and relation between input and output can be used in services interacting via POST (see [2] for an overview).

Best regards,

[1] http://people.aifb.kit.edu/aha/2012/sms/lids-eswc2011.pdf
[2] http://linkedservices.org/

The comments to this entry are closed.