Joho the Blog
An Entry from the Archives

the best semantic web I know is pandora, and is not by chanche that it's about a non semantic item like music, we defenetly have a long way to go....

I know that Mashups arent strictly reckoned to be instances of the Semantic Web, but since its only a matter of semantics (small 's') I reckon that the Tunisian prisons mashup in your next post can be regarded as an instance of the semantic web waiting to be recognised as such.

Mashups dont use the classic Semantic Web roadmap (RDF etc) but they do capture the basic idea, that automated processes can be applied to well structured information on third party sites which is left open to manipultation through the web.

I know that Mashups arent strictly reckoned to be instances of the Semantic Web, but since its only a matter of semantics (small 's') I reckon that the Tunisian prisons mashup in your next post can be regarded as an instance of the semantic web waiting to be recognised as such.

Mashups dont use the classic Semantic Web roadmap (RDF etc) but they do capture the basic idea, that automated processes can be applied to well structured information on third party sites which is left open to manipultation through the web.

The Semantic Technology Conference is all about "successful application of the Semantic Web" in the enterprise, etc. (I've never attended, so I only know about it through advertising.)

One thing to consider about your request: there is "the Semantic Web", which is the semantically rich World Wide Web; and there are applications that use semantic web technology (e.g., Firefox uses RDF); and there are applications / systems that use "semantic technologies" that support ontologies (e.g., any code or relational database, potentially), but don't use semantic web technologies.

For example, one could say that Flickr uses an ontology, and it's encoded in their code. But they haven't abstracted that ontology into a machine readable ontology document using semantic web tech (OWL), or exposed that document on the World Wide Web.

Likewise, applications that interact with Flickr through it's api have to "know" something about Flickr's ontology. But, they "know" that because someone hard codes "knowledge" into an application, not because the application can generally find ontological mappings--and say, treat a Flickr photo and a Shutterfly photo as the same kind of thing.

So, one thing to consider: are you interested in applications that do semantic mapping using "semantic technologies", using "Semantic Web" technologies, or just using computer code?

The last case isn't too interesting in terms of the Semantic Web, in the sense that it's about one-off cases, e.g., as if people kept creating new markup languages for their blogs, instead of using HTML. Nevertheless, a lot of tricky issues are solved, one at a time, in this way!

I mean Semantic Web applications that use Semantic Web technology, which means AFAIK that at a minimum they use triples, probably encoded in RDF.

If all people mean by "the Semantic Web" is that it'd be good to agree on metadata, that'd be sort of a let-down, wouldn't it?

"If all people mean by 'the Semantic Web' is that it'd be good to agree on metadata, that'd be sort of a let-down, wouldn't it?"

Yeah, totally!

It's the difference between structures that are re-usable for practically any kind of semantic info, and more one-off specific structures tied to specific semantic info.

But, in the enterprise setting, when issues are addressed via semantic tech approaches, even in the one-off specific structure sense, the results can be interesting. I mean, it's still using computers to connect info / knowledge in ways beyond what the individual systems "know" on their own.

I think reversibility should be the priority, as much a creating flexible meaning borders, that can adjust during socialization

regarding flickr thought as an ontology, and I really have low knoledge of codes of anykind, I don t know how you could do this with their database:
http://color.slightlyblue.com/
it seams they do have seamless stucture within their space. and it will never be possible to have one single internal language(point of view), that would be endless by definition.

but again may be I am mixing planes of meaning

You might be interested in some of the presentations from the Semantic Web for the Military User conference in 2003. It's been a while now, but several pilot projects in the govt and military arena were discussed there; some have continued, and others have arisen since.

(disclaimer: my colleagues at BBN Technologies are deeply involved in Semantic Web activity, and for the most part i like them, so this isn't entirely objective :-))

How about this?

SIOC Ontology is an interesting example you may use.

It provides a common vocabulary to express information in online community sites (forums, weblogs, ...) in a machine readable form.

SIOC information can be automatically exported from blog engines such as WordPress (using WordPress SIOC Plugin). It opens up community sites to the Semantic Web and helps decouple consumers of data from the specific APIs of each particular blog or CMS engine. Current size of the SIOC-o-sphere that we are aware of is ~150k triples.

While this is a not an industrial or company wide project, it is used by promising projects such as OpenLink Data Spaces.

More information can be found at: http://sioc-project.org/
If you have additional questions about SIOC please ask them on the SIOC-Dev mailing list.

dmoz.org - the open directory project. Still used by Google and Alexa, and the largest RDF repository in the world.

Also, the output of topix.net -- automatically tagged information from 50k blogs and news sources, with RDF feeds for 360k topics, all driven via a semantic extraction from unfielded data...