Peter Van Dijck has composed an interesting summary of the themes that arise in debates about the Semantic Web, including the recent discussion of Shirky's essay. I follow this discussion with some bemusement and distance. I don't think it's an exaggeration to claim that I'm the first RDF/XML pancake: experiencing the crush of the "RDF versus XML" debate, back in 1999 as a Chair of the XML Signature Working Group. Since then, I've quietly commented in many places on why I came to think that the RDF/XML syntax was inappropriate to that work, and yet how the work was improved by thinking about the data model. The political tensions are something best forgotten. The result of this crush is almost comical: the XML Signature Recommendation includes a GIF of an RDF model that basically models the XML syntax, quite different from the original semantic model.
While the public has rehashed many of the technical debates I had four years ago, this is the least interesting topic for me. I'm most interested in, though disappointed by, the discussion around historical innovation, which brings me back to Van Dijck's document. The first theme he identifies is the following claim, "If the Semantic web was anywhere even close to half-as-good as the claims that are being made for it, it would ALREADY have gained massive widespread acceptance. Meaning VAST implementation." I've heard that before, and my answer has always been: there's about three million Web sites out there!
Some will complain that my response is not fair, the Web is completely different, but they aren't up on their history. If you look at the original 1989 proposal to CERN for what would become the "World Wide Web", it's actually a proposal for what people now consider the "Semantic Web." The proposal is for an architecture that can be represented with a directed label graph, "We can call the circles nodes, and the arrows links. . . . Ideally, it represents or describes one particular person or object. Examples of nodes can be: People, Software Modules, Groups of People, Projects, Concepts, Types of hardware, Specific hardware objects. The arrows which links circle A to circle B can mean, for example, that A: depends on B, is part of B, made B, refers to B, uses B, is an example of B." This short description has got it all, including the existential quandary about representing a concept of a thing, and the thing itself.
A theme that Van Dijck hasn't identified is a common response by the advocates, "People originally resisted the Web and said it was stupid, but now look at it!" Yes, according to the original conception, about three million half baked Web sites. And I quietly think to myself, "Past performance is not an indicator of future results." Or at least that's the case in the stock market, I don't know if that's the case for innovation.
The interesting thing about both sides of the argument is that they both rely upon an incomplete articulation of history for their argument. And while the technical substance of the debate doesn't change with this realization, our understanding of innovation and the rhetoric surrounding this technology might. What are other cases of half-way innovations? In what ways does history cleave the innovation: what aspects are quickly deployed, and what is their relationship to those pieces still waiting in the wings?
Simply, when you must ask, "is it done yet?", what does an adoption curve look like?
Trackback from Raw on 2004-04-09
A mini-series at xml.com from Rich Salz, the latest being :webservices.xml.com: XKMS Messages in Detail. I keep forgetting to check...