answer to Miltiadis

Miltiadis D. Lytras posted questions about how the Semantic Web did in 2004.

in good spirit of public discussion, i hereby blog my answer to his questions.

1. What was the Key achievement for Semantic Web in 2004?
APIs. Forget the scientific glass-tower. Jena 2.0, Redland, Sesame, … that’s what counts and thats what makes things like Julie work. APIs enable developers to make products for people. Products that let people author RDF, publish RDF and use published RDF.

2. Which was the key technological achievement concerning SW for 2004?

Widespread FOAF usage. Even TimBl mentioned FOAF in interviews. That’s ontology work alive.

3. A title of a research paper published in 2004 with great impact
for SW Research.

hm. Nothing real new this year? Well, I quote myself: “Using Semantic Web technologies to build a Semantic Desktop”

4. An event (during 2004) that you evaluated of great importance for
SW.

FOAF Camp, Enschede, Holland. August.

5. A project / project deliverable that you evaluated as of critical importance for SW, in 2004.
http://www.foaf-project.org thats where we come together.


6. Which is the Key obstacle that you anticipate as the ‘battlefield’ for the evolution of SW?

Hot air coming from people that talk about “inference” while there is no widespread data available they can do their inference on. For me it sounds like “inferiour” than inference. Forget inference and OWL, stick to low tech RDF-S and procedural programming to get the SW lift off the ground.

Ontologies written by people that do not code. They are crap. Without a reference implementation, OWL or RDF-S ontologies are useless. You hit the nerve only when you build a useful application that makes its data available in the ontology. Only an application shows how the ontology should be used correctly. See DubliCoren: half of the terms are vague and are used wildly.

URL discussions. Much hot air on RDF-IG about what a URL denotes and why we should use URIs but no solution. As long as some people wait for the “perfect” system that uses unambigous URIs, nothing will be built at all. Gnowsis on the other hand goes the practical way: use changing URLs to identify stuff and begin coding SW today.

btw: there is no inference engine that can work in a global, distributed, semantic web. i.e.

?x <rdf:type> <foaf:Person> 

would take an hour on my machine….

7. A Business case that SW proved its capacity to solve real world problems.
Vendor independent social networking solved by foaf (isntead of orkut, friendster, etc).
News aggregation by RSS. that was real a kicker.


8. How did you saw the AIS SIGSEMIS Activities? What would you like from us the next year?

Peer review. Part of the articles was very simple. I.e. “component requirements for a universal Semantic Web Framework” – without a reference implementation this is rather speculative. We gnawed on this in practice (www.gnowsis.org). So I would like two things:

  • peer review
  • theories proved by reference implementations