Rules and process repository
This is the view i want to present to the jbpm-rules integration users:
rules and processes sources are managed in cvs or svn.
rule package info files contain references to rule sources on the file system
rule base info files contain references to package info files
an ant script can create a binary (compiled) version of a rule base by pointing it to a set of
rule package info files
rule base info files
optionally that binary rule base can be deployed to the DB-rule repository with the same or a different ant task. when a binary rule base is inserted in the rule repository, an id is assigned (the hibernate id and primary key)
also the eclipse plugins (rules and jpdl) could provide functions similar to the ant script.
the result is that there is a database based rule repository that is just a set of rule bases. rule bases have a name and a version as metadata. during runtime, rule bases are referenced
by name only, in which case the latest version of that named rule base is used
by name and version