CaCore

From Informatics

See also: CaCore In Geworkbench

"A data management framework". The caCore docs state you need this to develop a "caCore-like" system. Basically, what we need is an xmi file that can be uploaded to the caDSR (well, there is a little more to it than that). To get this xmi file, we can develop a uml diagram and "export to xmi".

caCore provides:

• EVS - supplies controlled vocabularies
• caDSR - a metadata registry
• caBIO
• CSM

Documents of importance:

• caCore Programmers Guide
• caCore_SDK Basic Install Guide - the sdk gives us a way to generate schemas for datatypes

jdk 1.5_06 (cagrid 0.5.3 must use j2sdk 1.4.2_04)

• edit in $caCORESDK_HOME/deploy.properties
  • paths should use /, not \
  • run ant build-system in $caCORESDK_HOME.
  • run ant deployWS in $caCORESDK_HOME.
  • shutdown tomcat (ant shutdown.bat from $CATALINA_HOME)
  • add a log4j.jar to $CATALINA_HOME/common/lib
  • start tomcat (ant startup.bat from $CATALINA_HOME)
  • you should be able to ping http://localhost:8080/example/Happy.jsp

The MetaData Issue

The example given is: class Agent (from caBIO) has the attributes:

• nSCNumber: Long
• name: String

This is how NCI models agent in the caBIO data model. The CIA could also model an Agent with the same attributes. To avoid have duplicate objects that are essentially the same, let there be "data about data" (metadata) to resolve which object is to be used. This metadata must be available at runtime ... enter caDSR.