The Semantic Web is a set of triples where one of the elements of the triple maps onto a node in a word-wide network of properties and the other two* maps to different resources**. (might sound a bit odd)
*There might be cases where only one of the elements are mapped, but that makes that statement "less valuable".
**Of course, the nod in the property-network is also a resource but it is simpler to understand this way.
I mean that the spaces and its elements combined create the SW.
Read the 20010330 The "core" architecture of the Semantic Web ... first perhaps.
This is what I believe 20010404 16:08 ... it might be really stupid ;-)
All this is illustrated in the picture below...this is actually the first picture of the SW (although it is only a small part!).
Extremely easy to misinterpret this picture. The top and the bottom space are the same (i.e. not disjoint) (this cold be figured out by the URI mapping). They are only separate so that the roles of the resources (in a triple) are more apparent. That makes the spaces "role-spaces". It is NOT an RDF graph. Of course, a node (property) in the property network could also be the object etc. but then it does not have the role as the property in that triple.
What the image is supposed to illustrate is that the Semantic Web is the triple space, i.e. I think that the SW only will be usable by machines (to help us in the end). A machine will use the triple space and the property network (yes it is a set of resources too, but they have another role there) to produce useful knowledge to us. Of course the machines will extract triples from Web pages, and databases.
Thus, humans will browse the bottom set, and machines will extract triples from that space and use it and the property network to derive stuff. (these might be posted as simple HTML docs) Thing will be added and changed soon, I'm a bit confused.
Note! I will write a small article explaining this picture...and throw in some machines too.
Please, send me a comment!