what's metadata, RDF and the Semantic Web?

Metadata is data about data. For example, the title, author, creation data, and licensing information of a file are all examples of metadata about that file.

RDF is a framework for metadata. The basic structure of RDF is very simple. There are three parts:

• the subject: the file we're talking about, identified by its URL
• the predicate: the type of metadata (like title, creator, etc.), also identified by a URL
• the object: the value of this type of metadata (like "Gnomovision" or a person name "Yo-Yo Dyne")

Together, these make what RDF calls statements. These RDF statements are saved in a special form of XML (a more generic version of HTML, the language that web pages are written in) called RDF/XML.

The Semantic Web is the plan for adding RDF to the Web. (When you get metadata on our site, we have simple instructions on how to add it to your web page.) The idea is that when enough pages carry this machine-processable metadata, we can build more exciting tools that take advantage of it.

RDF can be used to get different programs to talk to each other with less manual cutting-and-pasting. You could drag your flight itinerary onto your calendar program, which would read the embedded RDF metadata and add it to your schedule. You could drag a friend's top ten songs list onto your music player and it could try and obtain the songs for you automatically.

RDF can be used to create more powerful search engines. Right now the only question you can ask a search engine is "What pages have these words in them?" but if pages included RDF metadata, you could ask more advanced questions like "what's the current temperature in California?" or "find me pictures of cats that allow commercial redistribution". Then you could build programs that used this information, like an alarm clock program that also displayed the curren weather or a collage-making program that only used photos with permission.

RDF can be used to get aggregate information about the whole Web. We could also build tools that took advantage of the aggregate information. A crawler that downloaded everyone's top ten lists from their homepages could build a list of the most popular songs. A program could cross-reference this with a pricing catalog to see how much it would cost to buy all of them.