Dean Allemang gave an extremely informative talk at JavaOne today in the session, “Semantic Web for the Working Ontologist.” I walked away thinking the term Semantic Web is misnamed. And that alone can sum up the major issues surrounding the Semantic Web. Using the term ‘Semantic Web’ implies that the technologies and the concepts can work on the public web. Think for a quick second on how large the web is—it might not be as large as the universe we live in—but it might as well be. Think of all the different ways words are used, think of all the different languages, think of all the misuse of words, think of the general destruction language itself suffers through and then try and build technologies that rely on using language and words to compile information and present it in a usable fashion. You might as well re-invent language itself, teach the entire web-using world how to speak that language, and then insist that everyone uses that language consistently. Then you would have a workable Semantic Web. I don’t easily say that something is impossible, but the very idea of a Semantic Web makes me come very close to saying, this is impossible.
Semantic technologies, on the other hand, are something very different. Forget about compiling public information from all corners of the web into a centralized location. Instead, think about a closed database in which a controlled ontology exists. For semantic technologies to work you need terms clearly defined and used consistently. And when that happens you do have the power to create mash ups and applications that use databases in effective manners. This idea of revitalizing the usage of databases is itself remarkable with what can be accomplished.
If you place these semantic applications on the web, I guess you could call it the Semantic Web, it does sound a little more official that way. But is that accurate? And doesn’t that sum up, once again, the problem with the term, Semantic Web?