I mentioned some of this in our discussion but will relay here for others reading:
TDTL: I am only concerned about all solutions being enterprise level.
- The second layer solution is an excellent approach and would add versatility without adding bloat to the blockchain. On the other hand, it may not inspire as much confidence with external investors as an on-chain solution would.
- The second layer solution must be enterprise level from the start. It cannot be a Hivemind which has various format issues and was for some reason unknow to me written in Python (and now requires optimization). The new layer should be written in C++ for ultimate scaling and if it is written in Python it must be structured in a professional and scalable manner. Everything we do going forward must be aimed at being accepted as enterprise level and not be reminiscent of some half-baked startup. We are building Hive to last the test of time.
- We do need a flexible approach to smart contracts, NFTs and FTs. We're not married to the "SMTs" and we don't need to be. We're all aware that Steemit Inc was ran in a way that stifled the creativity and capability of the development team. We have use cases for all of the above. Smart contracts: structured payment systems. NFTs: collectibles. FTs: community tokens.
- A layered approach in general makes for a friendlier ecosystem towards development and onboarding of partially completed or unstarted projects. It allows for a phased set of milestones among the developers; ie. incorporate with select on-chain functions, then separate a side function to let's say create and utilize a token, then separate another server for a Hivemind node if needed. Your reference to Arcange's microservice is spot on.
There's a few conflicting points here but these are generally my thoughts.