The state of the semantic web

Lookin' good!

The W3C's Ivan Herman recently gave a talk on the State of the Semantic Web in Bangalore, and he's made the slides available online. Anyone remotely interested in the semantic web or RDF should look through the presentation; it may seem esoteric in places, with its talk of Horn rules and F-logic, but in general it's a clear, up-to-date summary of the important current issues.


One of my favorite slides is titled A major problem: messaging (which at first I thought referred to message passing, making me think "huh?"). It's a good summary of misconceptions about the semantic web, which are addressed on subsequent slides: AI reincarnated, merely ugly XML, top-down ontologies trying to dictate everything...: none of the above! The slide ends with the message "Some simple messages should come to the fore!", but I would quibble that one of these messages isn't so simple: "People should 'think' in terms of graphs [his italics and underscore], the rest is syntactic sugar!"

When most people think of graphs, they think of pictorial representations of data. Because I have a computer science degree, I know that by "graphs" he's referring to a particular data structure for which pictorial representations are a nice option but not the data structure's reason for being, and I worry that the use of this technical term will confuse the wider audience to whom we're trying to evangelize the semantic web. No punchy, correct, superior alternative springs to mind, though—collections of relationship descriptions and attribute/value pairs?

It was also nice to see the industries and famous brand names listed on the slides Some RDF deployment areas, The "corporate" landscape is moving, There has been lots of R&D, and Portals. This will help dispel one of the biggest misconceptions of all about the semantic web: that it's limited to academic, ivory tower research projects. Now if we can just cut back on the use of technical computer science terms...


Why not "People should just 'think' in terms of webs"?

A web of resources is how I think of RDF graphs, the visual image is closer to the actual data structures, and its a more positively associated with web/web2.0, simplicity, etc.

People should just think in terms of: Web Data or Web Data Sources or Data Sources :-)

At the end of the day, URIs are pointers to Data :-) This is why dereferencing URIs is feasible and such a powerful concept once broadly understood :-)

All this brings me to a slide in Ivan Herman’s talk about “The state of Semantic Web” in Bangalore, highlighted by Bob DuCharme. The slide highlights, I think correctly...

In Dutch it’s easier: a ‘grafiek’ is a pictorial representation of data, and a ‘graaf’ is well, the mathematical type of graph :). Although it’s also the word for count in Dutch (as-in count Dracula, not count numbers :)), but I don’t think that’s likely to cause confusion.

Unfortunately, graph probably still isn’t going to be understood by people without a slightly mathematical background. They might have had it mentioned in high school, but long forgotten.

I think if you want to explain the ‘graphical’ (ehehe) nature of RDF, it is best done with an accompanying picture, and as little text and URIs as possible.


But what does "snee" mean? (Just kidding.)