Including Metadata in a Web Page

Charlie Morris

June 26, 2000

There are a couple of ways to include Dublin Core elements in a Web document. It can be done within HTML, by using META tags, as in this example:


<META NAME="DC.Title" CONTENT="Are search engines dead?"> 

<META NAME="DC.Creator" CONTENT="Charlie Morris"> 

<META NAME="DC.Subject" CONTENT="The imminent demise of search 

engines has been predicted for years, but they're still kicking. 

Still, the current state of search engine technology leaves a lot 

to be desired. This article explains why search engines so often 

return reams of irrelevant documents, and hints at a possible solution."> 

<META NAME="DC.Publisher" CONTENT="Web Developer's Virtual Library">

Note that a particular document can use any, all or none of the available elements. For the complete list of Dublin Core elements, see http://purl.oclc.org/dc/documents/rec-dces-19990702.htm.

Rather than use HTML META tags, the better way to include these classifying elements in a Web document is to use the Resource Description Framework (RDF), together with XML. RDF is a foundation for processing metadata. XML provides a syntax for RDF, as in the following example:


<rdf:Description about=""

       xmlns:rdf="http://w3.org/TR/1999/PR-rdf-syntax-19990105#"

       xmlns:dc="http://purl.org/metadata/dublin_core#"

       xmlns:ddc="http://purl.org/net/ddc#"

       dc:Title="Are Search Engines Dead?"

       dc:Description="The imminent demise of search engines has been 

	   predicted for years, but they're still kicking. Still, the 

	   current state of search engine technology leaves a lot to be 

	   desired. This article explains why search engines so often 

	   return reams of irrelevant documents, and hints at a 

	   possible solution."

       dc:Publisher="Web Developer's Virtual Library"

       dc:Date="2000-06-24"

       dc:Format="text/html"

       dc:Type="article"

       dc:Language="en">

     <dc:Relation resource="http://wdvl.com">

     <dc:Creator="Charlie Morris">

   </rdf:Description>

Note the "dc" tags, which indicate Dublin Core elements.

The Dublin Core, RDF and XML are not necessarily dependent on one another. The Dublin Core is a system for describing metadata, RDF provides a way to encode it, and XML provides a browser-readable syntax for RDF. As we have seen, Dublin Core elements can be used with HTML or other languages. Also, RDF can theoretically be used with other languages than XML. Confused yet? In any case, the Dublin Core/RDF/XML "stack" seems to be the current state of the art.

And now the big question: Do any existing search engines actually support this wonderful system? According to the Dublin Core folks, several search-engines "will index meta elements with just a little configuration." These are they:

Not exactly a ringing endorsement from the search engine industry, especially as none of the big public search engines are on the list. Of course, it's a chicken-and-egg story - few sites use the Dublin Core elements, so why should the search engines bother with it? Frankly, I'm very surprised that this system isn't a lot better known, considering its capabilities and the strong need for it. Then again, the major search sites are famously secretive about the workings of their search engines, and it may be that they're quietly adding Dublin Core support.

If they're not, by golly, they should be. And all you Web developers ought to be including this valuable metadata in every new site you build. The Internet badly needs a universal, comprehensive classification system, and the tools are at hand to create just that. Let's all get off our bums and spread the word about XML and the Dublin Core. Dublin Core! Dublin Core! There, that ought to earn this page a good ranking in the search engines.

Including Metadata in a Web Page

Further Reading