Michael Vizard writes in: Linked Data to Take Programmable Web to a Higher Level:
The whole concept of a programmable Web may just be too important to rely solely on APIs. That’s the thinking behind a Linked Data Working Group initiative led by the W3C that expects to create a standard for embedding URLs directly within application code to more naturally integrate applications. Backed by vendors such as IBM and EMC, the core idea is to create more reliable method for integrating applications that more easily scales by not creating unnecessary dependencies of APIs and middleware.
…
At the moment most of the hopes for a truly programmable Web are tied to an API model that is inherently flawed. That doesn’t necessarily mean that Linked Data approaches will eliminate the need for APIs. But in terms of making the Web a programmable resource, Linked Data represents a significant advance in terms of both simplifying the process of actually integrating data while simultaneously reducing dependencies on cumbersome middleware technologies that are expensive to deploy and manage.
Conceptually, linked data is obvious idea. But getting everybody to agree on an actual standard is another matter. At the very least, however, a generally accepted approach to linking data within applications that make the whole programmable Web concept more accessible to developers of almost every skill level should not be all that far off from here. (emphasis added)
I am often critical of Linked Data efforts so let’s be clear:
Linked Data, as a semantic identification method, has strengths and weaknesses, just like any other semantic identification method. If it works for your particular application, great!
One of my objections to Linked Data is its near religious promotion as a remedy for semantic diversity. I don’t think a remedy for semantic diversity is possible, nor is is desirable.
The semantic diversity in IT is like the genetic diversity in the plant and animal kingdoms. It is responsible for robustness and innovation.
Not the fault of Linked Data but it is often paired with explanations for the failure of the Semantic Web to thrive.
The first Scientific American “puff piece” on the semantic was more than a decade ago now. We suddenly learn that it hasn’t been a failure of user interest, adoption, etc., that have defeated the Semantic Web, but a flawed web API model. Cure that and semantic nirvana is just around the corner.
The Semantic Web has failed to thrive because the forces of semantic diversity are more powerful than any effort at semantic sameness.
The history of natural languages and near daily appearance of new programming languages, to say nothing of the changing semantics of both, are evidence for “forces of semantic diversity.”
To paraphrase Johnny Cash, “do we kick against the pricks (semantic diversity)” or build systems that take it into account?
[…] Another Word For It Patrick Durusau on Topic Maps and Semantic Diversity « Linked Data: Esperanto for APIs? […]
Pingback by Open Services for Lifecycle Collaboration (OSLC) « Another Word For It — July 29, 2012 @ 9:55 am