Re: The Battle for Linked Data

On 26 March 2012 16:49, Hugh Glaser <hg@ecs.soton.ac.uk> wrote:
> So What is Linked Data?

I think this can be defused:

"'Linked Data' is the use of the Web standards to share documents that
encode structured data, typically but not necessarily using a graph
data model."

Considerations --- It's important to be open and inclusive. It's
important to mention the webby graph data model without getting bogged
down in the detail (RDF? which version? which format? OWL too?). It's
important to mention standards. Sharing (intranets included!) is more
important than 'publishing', or 'public', though the latter should be
alluded to. If we stray too far from the graph data model and Web
standards like URIs, we lose interop; if we stray too far into nerdy
semweb rdf detail, we lose mainstream audience. It's a balance and
it's for the market not us to say where the sweet spot lies. And if we
start religiously starting to force modeling idioms on the world, we
lose credibility; no anti-bnode laws, or strictures about
http-range-14. Some things are best left unspoken! Fashions will come
and go; look at HTML frames and Flash splash screens. Good taste will
triumph, without the "Linked Data" slogan needing to encode all its
aspects.

'Linked Information', from a FOAFy perspective, is then the larger
"let's share what we know" perspective
(http://www.flickr.com/photos/danbri/4030764915/ etc) in which we
apply equal passion to the sharing of information that is in non-graph
data formats, or in people's heads. Doing so brings the graph data
model into a distinctively central role, since it can describe other
data formats (GML files, spreadsheets, MP3s, videos, mysql dumps...
RDF's original use case as metadata), and it can describe people and
their characteristics. So we can be pro-RDF here without forcing it
down people's throats... and we can be pro-data while admitting that
there's vastly more to Web-based information sharing than triples, and
more to 'sharing what we know' than sharing data.

cheers,

Dan