Peter Norvig's view published on AlwaysOn seems to be colored by a decidedly web search engine perspective. If we start looking at Enterprise Semantic Applications (semantic applications developed for targeted enterprise/corporate/scientific/engineering user base, whether the data comes from the enterprises, or is a syndicated/licensed content, or is a open Web content), you can start to see some exciting alternative perspectives and realties. Let us review Peter's belief that the point that there is not enough RDF and SW content. This is growing at an extremely rapid pace (I am sure others will put out numbers; and there are specialized search engines such as Swoggle that focus on Semantic Web content with rapidly growing size of indexed documents: check ). More importantly, the promise of Semantic Web is closely tied to having the tools for semantic annotations of heterogeneous content, i.e., create semantic metadata automatically. This is much easier to do when you have high quality domain ontologies that bound the scope of automatic extraction. And I really do not see content suppliers putting the metadata in (as we do not see Web page authors using metatags), or at least this will be optional and just one form of input. Instead, metadata will be created with respect to potential use (e.g., there are some definite concepts when we deal with WorldNews, USNews, TechnologyNews, and so on). Commercial technologies (example) can process millions of pages per day and extract semantic metadata, and all these can be represented as RDF (and that is a good idea because of the benefits esp. for high end semantic applications such as analytics). Granted this is hard to do on a panWeb scale where you have no single domain or even a limited set of domains, and huge diversity of users who may need to see the content from different perspectives. Even here, I believe much can be done, but it will take a little more time - maybe 3 years.

Let me next respond to the comment about ontologies. There are many cases within Enterprises and even for consumer applications (e.g., see examples; also expect to see use of Ontologies soon by Amazon and this types of companies). Rather than focusing on common sense or general purpose ontologies, the immediate future is with domain and task/application specific ontologies (latter are appropriate for enterprises, eg., SOX or Anti-Money Laundering (AML) applications). This is done now successfully, with line of business applications (e.g., AML application deployed at one of the largest banks). These types of ontologies routinely have millions of instances (look at SWETO, NCI ontology, GlycO, as very different types of examples: think of SWETO with about a million instance as a poor sample of what is being done for real world Enterprise class Semantic Applications representing only a 10th of population and with diffused focus that a typical enterprise deployment; with 767 classes, think of GlycO has well-focused scientific ontology developed by domain experts; NCI ontology involves more community involvement). Additional thoughts on SW adoption are in my earlier item on this blog.

Postscript: Just before posting this, I stumbled across Danny Ayer's views in response to Peter Norvig's item; I agree with those view almost completely. After my initial posting, I came across another set of comments by Tim Finin.