by Tim Berners-Lee
Progress in communications technology has been characterized by a movement from lower to higher levels of abstraction. When, first, computers were connected by telephone wires, you would have to run a special program to make one connect to another. Then you could make the second connect to a third, but you had to know how to use the second one too. Mail and news was passed around by computers calling each other late at night, and, for a while, email addresses contained a list of computers to pass the message through, such as: timbl@mcvax!cernvax!cernvms.
It's not the wires - it's the computers
The ability to use this communication power between computers wasn't truly useful until the Internet. The Internet allowed one to forget about the individual connections. It was thought of as the 'Internet Cloud'. Messages went in from one computer, and appeared in another, without one having to worry about how they were broken into packets, and how the packets were routed from computer to computer.
When things went wrong, she had to be able to figure out whether it was a problem with her connection to the Internet, with the URI in the link she was following, or an error on the server end. This involved looking under the hood, as it were. But that was, and still is, the exception.
There is another reason to be aware of what is happening. The information one browses comes from a particular server, whose name has been registered to a particular person or organization. The trust you put in that information relates to who that organization is. The serving organization is responsible for keeping the URIs you bookmark today alive tomorrow. So the Web is just a web of documents, except one has to be aware of the social aspects implied by the underlying level.
It's not the documents - it's the things
The power of the Web hadn't reached its full potential until the semantic Web came along. The semantic Web's realization is, "It isn't the documents which are actually interesting, it is the things they are about!". A person who is interested in a Web page on something is usually primarily interested in the thing rather than the document. There are exceptions, of course - documents are certainly interesting in their own right. However, when it comes to the business and science, the customers, the products, or the proteins and the genes, are the things of interest. A good semantic Web browser, then, shows a user information about the thing, which may have been merged from many sources. Primarily, the user is aware of the abstract Web of connections between the things - eg: this person is a customer who made this order which includes this item which is manufactured by this facility ... and so on.
There are again the same two exceptions. When things go wrong, the user must be able to look under the hood to see whether the document was fetched OK but had missing data, or the document was not fetched OK, in which case what was the underlying Web problem. And again, when the user is looking at a bit of data in the data view, perhaps a point on a map or a cell in a table, then she must be able to see easily which document that information came from.
One thing, or person, may have many URIs on the semantic Web. Often such a URI has within it the URI of the document which has some information about the thing: a gene as defined in the Gene Ontology; a protein as defined in this taxonomy; a citizen as defined by the Immigration and Naturalization Service glossary, and so on. Similarly, the URI of the document has within it the DNS name of the computer. So the social structure can be seen within the URI.
The World of Things ... on the Web
The Web of things is built on the Web of documents, which is built on the Web of computers controlled by DNS owners, which itself is build on a set of interconnected cables. This is an architecture which provides a social backing to the names for things. It allows people to find out the social aspects of the things they are dealing with, such as provenance, trust, persistence, licensing and appropriate use as well as the raw data. It allows people to figure out what has gone wrong when things don't work, by making the responsibility clear.
The last level of abstraction is the Web of real things, built on top of the Web of documents, which is in turn built on the network of computers.