- Improve supportability
- Separation between normal usage from bleeding edge usage
- public API vs internal packages
- Userguide vs devguide
- More stable database / process definition caching
- Migration becomes feasible
- Automatic JBoss installation
- We ship automatic installer
- No messing with configs and libs
- Tested configurations
- Continuous integration
- Ability to reproduce bugs
- Collect community knowledge in test suite --> Ability to keep improving over longer time --> Longer lifespan of jBPM 4
- Identifying supported configurations. Easy maintenance through config import system
- Grow to next level of adoption
- Lowering the treshold to get started
- Improved docs
- Easier configuration abstraction through imports
- Examples
- API split: normal usage versus bleeding edge
- Automatic installations prevent messing on your own
- API based queries
- Increase 'ability to execute'
- Historical data separated from runtime data
- Improved transaction mappings
- Proper handling of timers and external triggers
- More stable database
- Improved job executor
- Split history vs runtime --> queryable history information
- Implementation improvements
- Improved pluggable architecture. Changed composition to inheritence. Included variables in the PVM. Got rid of the long set of method invocations to update a variable.
- Introduced interfaces for activity implementations. Only exposing the methods that can be invoked.
- Changed recursion with iteration in process interpretation algorithm: dealing with automatic long loops.
Comments