Tim Berners Lee's idea of
"a Web that includes documents, or portions of documents, describing explicit relationships between things and containing semantic information intended for automated processing by our machines."
See:-
purl.org 
(source of quote above)
-- Sean Palmer
"Hooray! Someone's invented Semantic Networks for the 86th time!" --Peter Merel
The State of the Semantic Web (2008): www.w3.org
The "definitive" introduction was probably Weaving The Web, published in 1999.
another notable intro :-
(you've not been here for a while, have you Sean ;-)
-- Danny Ayers
Very true Danny... the SW activity doesn't even seem to have been released by then. Oh well, here are some more links then:-
And coupled with your lot...
Augment that with the eminently practical:
That should be enough to hold people for now. -- Sean Palmer
Let me inject a detour for the recent Iwanna Learn Xml graduates. See a 2003 "Extreme Extensibility" mechanism at www.xfront.com -- Microsoft Slave
Here's another : Semantic Web related books
www.isacat.net Broken Link 20071212
-- DannyAyers
And a kind of Topic Maps of the Semant(op)ic Universe
-- BernardVatant
A solid argument against the Semantic Web mindset - entertaining and very incisive, titled 'Meta Crap' :)
This is Slippery Slope argumentation, as if the fact that there are imperfections due to human involvement means that it should just all be thrown out. Furthermore, the author only chooses to look at the negative side of the "more-than-one-interpretation-of-data" coin. That's exactly what the semantic web wants to do there, direct people towards the positive side of that coin and flexibly allow but powerfully direct multiple interpretations.
[Tragedy of the commons applies here. People will, and do, lie about their sites. False meta-data, misleading meta-data, broken interpertations... look at the sort SEO attempts Google has to deal with]
However, even in the face of search engine optimization Google somehow manages to provide extremely high quality results based on metadata. In general, I believe that the benefit of common metadata and ontologies far outweighs the cost of the inaccurate or misleading use. Over time, maintainers and users get better at separating the wheat from the chaff.
-- Ryan Hoegg
Google assumes the burden of doing so, and has spent enormous amounts of money, brainpower, and computing power to do so. Requiring every individual user to make the same investment is not practical. A Semantic Web that's tied to a specific vendor to be usable isn't that useful.
The effort to categorize every Wiki page would seem to fall under the aegis of Semantic Web. I imagine someone out there with lots of time and an attention to detail could potentially generate a Dublin Core based RDF for each and every page on Wiki. Meta Wiki?
This is now possible on a page by page basis using Dc Dot. -- John Fletcher
"Semantic Wiki Wiki Web could be a good idea to work on. Don't you think?"
Moved from SemanticWebTips
Write your Semantic Web tips here:-
"If you don't have a domain name that you think you're going to keep..." You need to emphasize that this domain name needs to be kept practically forever, and the URI must not be reused. (Most people don't have that kind of persistance policy.)
"Use URIs that contain as little information as possible." Perhaps point to Memorable Random Strings (MeRS) and other such resources?
"Reuse existing ontologies/namespaces wherever possible, but if your RDF starts looking like Frankenstein's early models then consider defining equivalences/subclasses of classes found in other schema within your own schema."
-- Danny Ayers
Sean Palmer is turning Semantic Web Tips http://infomesh.net/2001/08/swtips/ infomesh.net into a collaborative thing. Feel free to add your contributions above.
ERCIM has a really good special issue on this stuff. 26 short topical papers.
See also Clay Shirky's essay at www.shirky.com
I took out the reference to this at the top of this page. As has been pointed out by more qualified people than me - for example, Danny Ayers, at dannyayers.com - this essay is based on a big Straw Man. -- Earle Martin
Why Semantic Web
But what will we accomplish with it? --Marc Thibault
Would you like to read Danny Ayers views (touches on Google Search as well) blog at dannyayers.com and tell C2 community about its contents and your views of it?
Danny is totally absorbed by the "how" question, clearly confident that if we build it they will come. One key question I haven't seen answered: If all you guys are tagging your data, what incentive do I have to do the extra work of tagging mine? --mt
In Mar05, Danny said at dannyayers.com
...describing the benefits of Semantic Web technologies - until there are more RDF-oriented applications and services around, the benefits of of using a domain-agnostic language above that of XML won't be apparent. What benefits?
Resources
The Semantic Knight vs The Hacker (Monty Python parody): lists.xml.org The other two parodies in that thread are also pretty funny.
A No-Nonsense Guide to Semantic Web Specs for XML People at www.betaversion.org
Books
Explorers Guide to the Semantic Web (see also Semantic Web Layer Cake)
There is an overview of tools available for the Semantic Web at semanticweb.org
One possible tool is Cubic Web.
The Semantic Web is the premier approach toward handling processing and working with Real Information using technology. It consists of entities linked to entities linked to entities using various relationships. The problem with the semantic web is that it is disconnected from data. In order to connect it to data and make it highly useful, the entities themselves need some sort of simple generic Information Friendly technology to define and describe them. I recommend Endemes as a possible technology for bringing together the worlds of data and information. Each endeme is an instance of an Endeme Set. Each endeme has a very simple generic way of representing data. Most data can be stored in endemes, and that which can't can still have its context described by endemes. Endemes are simultaneously Information Friendly and Data Friendly. --Jon Grover
See original on c2.com