The Naked RDF King
I kind of like the unintended pun in the tagline of Planet RDF:
it’s triples all the way down
But I’d rather go back and work on my own model and its implementation than poke fun at a naked king.
URI do not qualify as a Semantic Model
URI are not appropriate for the Semantic Web, Tim Berner Lee wrote that himself back in 2001:
http://www.w3.org/DesignIssues/HTTP-URI.html
section 2.5 - Extra info with URI:
Effectively, the URI scheme has now failed to identify anything by itself.
but nobody seemed to care then.
RDF is out of context
Nothing new either. In the same paper about design issues cited above, the father of the Semantic Web allready concluded that triples could not carry a context, that they could not express contextual dissent.
2.6 - Different meaning in different context:
It doesn’t work in the long term because it breaks the axiom that a URI must identify one thing.
Two years later, on Thu, 18 Dec 2003, Chris Bizer extended the Semantic Web data model:
in a quest for the still missing trust and proof for the Semantic Web.
Reality Checks
Since then, quatuors with context have been forced into many RDF store:
Perfect Index Structures for Storing Semantic Web Data With Contexts
http://sw.deri.org/2004/06/yars/doc/summary
and it not triples all the way down after all.
Finally, Named Graphs had to added to the pile of RDF/XML standards by the same Chris Bizer and patched on RDF/XML and URIs to have some form of semantic model for articulated text at last.
Real RDF Applications
Developpers have applied RDF, but not to make a semantic web. In a response to growing dissent about the W3C standards, Shelley Powers wrote that critics Ignore the fact that it’s working and she gave a few examples:
Mozilla/Firefox has been quietly using RDF for much of its underlying menu structure and other uses for six years or so now.
Is Mozilla a semantic application? Does it comes with its own RDF search engine for the page browsed? No it does not. RDF is used by Mozilla for its own internal resource description, not for semantic web inference.
RDF/RSS, known as RSS 1.0, has been providing syndication feeds for years.
But it requires a Universal Feed Parser because there are more than ten incompatible kind of RSS dialects and because the whole stack of XML/RSS/HTTP on which the Blogosphere was built cannot produce consistent character encoding.
FOAF is used to drive out networking in various environments.
And how many people actually use it?
Two Japanese student started a project to build a “Personal Publishing Platform with RSS and FOAF” in 2001, using RDF (Semblog: http://www.semblog.org). It practically got nowhere, like the rest of the W3C semantic web.
So What?
As a data model for the Semantic Web, RDF triple is broken.
Forcing quators in its store without changing de data model is like patching text articulation in RDF/XML triples with URIs as subject and object because there is no context available (which is exactly what Named Graph do).
It does not work.
RDF stores boast millions of triples, but how many of operations does a complex SPARQL request entail? And for what result? Syllogisms.
What can an application developper do with an RDF store that cannot be achieved in a simpler way with a relational database or an object database?
If you have tried to actually write an application with RDF/XML for the real world, I suppose you known what I mean. I did not take the time, pain and suffering to even test one of those Java monsters.
I don’t like Java.
Making it Simple
I like Python, less code and simplicity.
It’s easy to poke fun at serious and smart people trying to save a bad idea from a reality check, but there are better things to do than denigrate their sincere efforts
I wanted to make that Semantic Web work as Joshua Allen outlined in four years ago:
And it is getting real. Real soon now.
October 25th, 2005 at 2:20 pm
[…] Viene poi accennata la questione del SynWeb e della interessante iniziativa di Laurent che trovo perfettamente riassunta in questo post, che commentero’ a breve, ma che e’ un chiaro riassunto delle questioni spinose di RDF… -> The Naked RDF King […]