What you're suggesting is pretty much what I do now... I put a Data Source with the same name in the project directory. But since the rdl from the server contains a path relative to the root node like, /CustomDatasources/Mydatasource, I still have to go into the rdl code and strip out the /CustomDatasources/ path so that BIDS can find the Mydatasource.rds in the project directory. Then when I deploy, I have to add the /CustomDatasources/ path back in
I was hoping I could create a CustomDatasources folder somewhere that BIDS would find it and put a copy of the Mydatasource.rds in it. That way I could do this one time and not have to create another rds if I create a new project. Maybe this isn't possible in BIDS? Could be that the project folder is the only place BIDS can look for a data source?
As to the report designer. It's not a generic application. It's an app developed by Metavante, our banking system provider, that is tightly integrated with their database schema. Their db contains hundreds of tables and something around 30,000 fields. The report designer provides metadata that helps users locate the fields, perform the necessary joins, select criteria, etc. It allows mere mortals to compose reports and probably addresses 90% of the desired reports. It's the 10% that I have to pull into BIDS and tweak. For example, it can't handle self-joined tables.