I very much agree that SaaS is yet another strategic approach to controlling information and the software used to gather it. But it's hard to completely throw away such a useful abstraction. From a pragmatic view, SaaS is a convenient separation of concerns applied to both infrastructure and software.
Perhaps I missed it, but does RMS actually supply a solution to problems solved by SaaS? I noticed a few already in the threads here, but this basically characterizes most choices:
1. A completely decentralized approach, where everyone shares the software and information equally.
2. Every SaaS must run and share opensource code, as well as somehow opening sourcing the content (safely) as well.
3. Assume the worst about all endpoints, eventually empirically and/or contractually trust certain ones in a white list.