What are the weaknesses of the approach to using Swarm as the data layer for the Anarchive compared with using SOLID, or simply IPFS? Could we use OASIS to hedge against dependency issues while growing the project and network over year 1 - so we can rely on SOLID - web 2 data stores, and or IPFS while having an experimental branch for Swarm? How could we structure this so it makes sense and avoids duplication?