Q: What is it?
Semantic Web helps us to find what we need on the Web (and more).
Q: Doesn't Google, AltaVista or other search engines already do that?
By classic search engines you search for appearance of a word or a combination of words. Of course, these engines consider how the words appear in a document, the links to these documents and so on. However, very often we look for things more specialized or complicated, that cannot be easily described by mere combination of words. Suppose you look for a document about Semantic Web. Easy?! Type in "Semantic Web" in Google!
Now suppose you look for a document about semantic web for non-professionals that describe it in simple words. So type "Semantic Web in simple words" or "Semantic Web for non-professionals" in Google. What about "semantic web for dummies"? No matter which one of these combinations you try, there will be plenty of irrelevant documents in the result of your search. Besides, there will be plenty of documents of your interest on the Web that do not appear in the result. These combinations of words only estimate what you really need.
It can be even worse! Suppose we look for some documents about the use of semantic web in business for managers. Can we search for such documents by a few keywords on Google? There are things in our minds that their descriptions take more than a few keywords.
Q: So, I think of something and then Semantic Web will help me to find it! Are we talking about magic?
At the beginning of
The
Science of Discworld,
Terry
Pratchett has a few nice quotes, including:
"Any sufficiently advanced technology is indistinguishable
from magic" from
Arthur
C. Clarke
"Any technology distinguishable from magic is insufficiently
advanced" from
Gregory
Benford
...
The moment we know how magic works it changes to science. So,
let's turn this magic to science:
Using special languages we describe the content of a document
that we publish on the Web. The same special languages may be
used to describe what we look for. You want to know about some
of these languages:
RDF or
OWL.
These are languages recommended by World Wide Web Consortium
(W3C) for this purpose (and
more).
Q: I knew there would be a trick, somewhere! These languages don't really help; they make our life even more miserable! Can't we just use a simpler language?
Don't panic! The inventor of the Web perceives Semantic Web as machine-understandable Web. These languages are for machines. What you call a "simpler language" is indeed more difficult for computers to process, because they have much ambiguity. The purpose of these languages is their easy processing by machines and their precision. If people can manage to describe the content of their documents on the Web in such language and formulate their query in such language then we will get much better result when looking for a document about "semantic web described in simple words".
Q: If I am not dealing with those languages, I may consider that. Who is the poor bastard who does that, anyway?
Programs! Just like today when you use Web, you don't need to know HTML. In future when you use Semantic Web, you will not need to know RDF or OWL. The first step of the challenge of Semantic Web has been to define languages mentioned above (and they are still at their early stages). The other challenge is developing software programs that help people to describe their documents and/or describe what they look for on the Semantic Web. Finally, there will be reasoning systems matching descriptions of required documents and descriptions of contents, then find what is needed. This is more complicated, but more promising compared to keyword-based search.
Q: That sounds like Artificial Intelligence. Hasn't AI already failed?
If you mean AI couldn't produce the type of agents in Matrix, you are probably right. Only I am not sure if that is where AI was heading. There is no mystery or magic in AI - it's science! A program plays chess against you and beats you. Is that AI achievement or is it not? One more thing, these reasoning systems provide the capability of doing what is needed for Semantic Web, you don't like calling them artificial intelligence, call them whatever you like! They do the job!
Q: So, all Semantic Web does is finding exactly what I need?
That is close, but there is more into it. True, it is about improving precision. Even if what you need does not appear on the Web, Semantic Web can still find documents close to what you need. It can also tell you that they are different, how different they are and how you can use them for your purpose. This is actually where Semantic Web is heading.
If you have any comments to improve the above text or any question,
let me know.