untitled

78 4 The Semantic Web and Bioinformatics Applications

<association>

<instanceOf>

<topicRef xlink:href="#citation-attributes"/>

</instanceOf>

<member>

<roleSpec><topicRef xlink:href="#owner"/></roleSpec>

<topicRef xlink:href="#NLM"/>

</member>

<member>

<roleSpec><topicRef xlink:href="#status"/></roleSpec>

<topicRef xlink:href="#completed"/>

</member>

<member>

<roleSpec><topicRef xlink:href="#date-created"/></roleSpec>

<topicRef xlink:href="#date991021"/>

</member>

</association>

<topic id="date991021">

<baseName>

<scope>

<topicRef

xlink:href="http://kmi.open.ac.uk/psi/datatypes.xtm#date"/>

</scope>

<baseNameString>1999-10-21</baseNameString>

</baseName>

</topic>

Except for some syntactic details such as striping and built-in attributes in

the RDF namespace, RDF documents can be very similar to general XML

documents. As the example above illustrates, XTM documents have no such

advantage.

XTM is a graph-based language that has much in common with RDF. Both

of them are intended to be a mechanism for annotating web resources. The

web resources that are being annotated occur within documents which fur-

nish the “primary structure” defining the resources. The annotations are a

“secondary structure” known asmetadataor “data about data.”

Although XTM and RDF have many similarities, they also differ in some

important respects:

• XTM relationships (calledassociations) can have any number of roles. By
  contrast, the RDF languages only support binary relationships.


• XTM has a notion of scope or context that the RDF languages lack.