Use case: RDF data files on a cellphone

A cellphone stores PIM data (names/addresses/phone numbers) as RDF. A
developer of the phone software needs to find the names of everyone with
a phone number in this local store (for display to the user) and then
find all the phone numbers and the roles ("home", "work", etc.) of those
numbers for the person the user selects, and then actually call the
number the user chooses.


This use case also demonstrates programmatic access to RDF data, but
emphasizes the limited environment in which the application operates:
clearly the phone needs to be able to store numbers internally and
doesn't need a networking stack for minimal functionality.
The scenario I envision is that somebody writes a little library for the
phone's underlying OS such that the application developer need only form
a query in our query language and make an API call to that library to
get the data in which they're interested.

Note that there are also hidden "node identifier" constraints in this
use case: choosing the "home" option at the last stage of the interface
will have to produce a query for the number for the particular node
being displayed, not just any number with the identifier "home". Does
this mean that the application has to carry the name data all the way
through? Does it imply that the identifiers to potentially anonymous
nodes must persist across queries? There are certainly lots of ways to
do it; developers will like some more than others.

Received on Tuesday, 6 April 2004 16:49:14 UTC