From: Mike Dean (mdean@bbn.com)
Date: 02/04/03
Here's a candidate use case for DAML Rules:
Use Case: Ontology Transformation
The representation most appropriate for storing
information may not be the most appropriate for
manipulating it. Rules can help bridge this gap.
Several examples:
A genealogy knowledge base [1] may represent data as
childIn and spouseIn relationships between Individuals
and Families to avoid duplication of information.
Some applications may be better suited to work with
direct relationships between Individuals (e.g. parent,
mother, sister, etc. [2]).
A knowledge base may contain a property birthDate, but
the application needs to work with currentAge.
A knowledge base may contain individual sales
transactions, but some applications need to deal with
aggregations (e.g. total sales by department).
Several issues came up in thinking about this:
Do we want to distinguish between transformations that add
statements to the current KB and those that normally are
used to create a second KB? See [3] for an example
application using such a filter?
Do we want to deal with aggregations separately?
Both filters and aggregation could relate to DAML Query.
Would it be reasonable to includes rules in the query
premise?
Mike
[1] http://www.daml.org/2001/01/gedcom/
[2] http://www.daml.org/2001/02/gedcom-ruleml/
[3] http://www.w3.org/2001/02pd/
This archive was generated by hypermail 2.1.4 : 02/04/03 EST