TimBernersLee'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."
"Hooray! Someone's invented Semantic Networks for the 86th time!" --PeterMerel
The State of the Semantic Web (2008): http://www.w3.org/2008/Talks/0307-Tokyo-IH/
The "definitive" introduction was probably WeavingTheWeb, published in 1999.
another notable intro :-DannyAyers
Very true Danny... the SW activity doesn't even seem to have been released by then. Oh well, here are some more links then:-
Augment that with the eminently practical:SeanPalmer
Here's another : Semantic Web related books
-- DannyAyersAnd a kind of TopicMaps of the Semant(op)ic Universe SemanticWeb mindset - entertaining and very incisive, titled 'MetaCrap?' :) SlipperySlope 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.
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 SemanticWeb 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 SemanticWeb. I imagine someone out there with lots of time and an attention to detail could potentially generate a DublinCore based RDF for each and every page on Wiki. MetaWiki? --StevenNewton
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 MemorableRandomStrings (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."
ERCIM has a really good special issue on this stuff. 26 short topical papers.
See also ClayShirky's essay at http://www.shirky.com/writings/semantic_syllogism.html
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, DannyAyers, at http://dannyayers.com/archives/002017.html - this essay is based on a big StrawMan. -- EarleMartin
But what will we accomplish with it? --MarcThibault
Would you like to read Danny Ayers views (touches on GoogleSearch as well) blog at http://dannyayers.com/archives/2004/11/23/google-web-vs-semantic-web/ 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 http://dannyayers.com/archives/2005/03/29/technology-choices/
The Semantic Knight vs The Hacker (MontyPython parody): http://lists.xml.org/archives/xml-dev/200504/msg00260.html The other two parodies in that thread are also pretty funny.
A No-Nonsense Guide to Semantic Web Specs for XML People at http://www.betaversion.org/~stefano/linotype/news/57/
One possible tool is CubicWeb.
The SemanticWeb is the premier approach toward handling processing and working with RealInformation 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 InformationFriendly 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 EndemeSet. 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 InformationFriendly and DataFriendly?. --JonGrover