A main feature of the iServer link model is the distinction between
a resource and the corresponding selector providing the flexibility to address a specific part of a resource. Nevertheless, it is also
possible to support simpler applications where the entire resource is always selected. The augmentation
of arbitrary physical objects with digital information is one possible application of such a simplified link
definition. For example, RFID tags are often used for non-visual object identification in ubiquitous computing.
In the Global Information Systems (GlobIS) research group at ETH Zurich we
have extended the iServer platform with a resource plug-in for object identification based on RFID tags.
This turned out to be a simple undertaking since we only had to model the RFID tags as a specialisation
of the selector which enables us to address parts of an RFID address space. Each RFIDTag stores a unique
identifier. At request time, iServer checks which RFIDTag is selected by a given input identifier and activates
all associated links. Note that there exist various RFID standards, all with different address space ranges..
Our approach for identifying objects based on unique identifiers has many advantages compared to the simple
mapping table approach described in the background chapter. Of course, a simple mapping table could be applied
for resolving the unique identifiers to a Uniform Resource Identifier (URI). However, by implementing an RFID
plug-in for the iServer architecture, we get access to all other iServer features such as, for example,
user management for access control and are not limited in linking physical objects to URIs only, but can
link them to any other iServer resource plug-in. The RFID plug-in for example has been used in the
Lost Cosmonaut application.
