we are considering to replace repository with file system to store transformations and jobs in order to:
- To make more efficient the runtime repository loading in memory
- To improve versioning of kettle resources (files .ktr,.kjb)
This is a big change for our projects so we would know if this is a good solution according to the “Kettle RoadMap”.
We mean, for example, using file system instead of repository, in the future, we will loose some kettle’s features, components, compatibility?

Thank you!!