I anticipate that the usage of certain applications could be hindered by transaction costs on mainnet. For the mvp list of apps tipping and voting I think would fit in this category. The hamburger app (buying/editing the banner) would be sufficiently high-value i think, as would registering and claiming distributions (though perhaps not so much for low value claims).

Voting

Actual sentiment and governance voting will primarily remain on Reddit, with Reddit reading weighting values for voters from Ethereum. The exception is voting for acceptance of new merkle roots for registration periods and distributions. My hope here is that we can request a volunteer group of participants to promise to participate in these. Could call them guardians, they would be people from the community with good repute who volunteer for this probably 1/month task.

Tipping

Once tokens are moved on-chain then we need to decide how tipping should work. Currently the app I have built has some nice features (can tip to un-registered username and they can claim once they register) but does not try to mitigate tx cost in any way. Centralised, custodian-type tipping is one option. Another is to explore further certain payment/state channel options (so far I have found these complicated to decipher how daonuts would use them but I would welcome a second opinion). As well, the mvp could be left without any tx cost mitigation and leave that for a followup upgrade (gas cost atm is quite low and many tips may still make sense value-wise).

A Comprehensive approach

The one solution I can think of that helps solve scaling in a more general way is a bridged side-chain. If enough validators would volunteer I find this a compelling option because multiple types of apps could be scaled this way more generically and easily. Some docs and tools I find relevant here:

• Parity POA chains
• Parity token bridge
• POANetwork token bridge

Future applications like curation (high volume, but v. low value) would likely depend on implementing tx cost mitigation. If other subreddits chose to use daonuts then they might also piggyback on this side-chain. I also don't think it's crazy that certain other dapps use such a side-chain to scale, either. The problem, though, is that at least currently I think setting this up technically takes the project too far beyond what I can reasonably support myself. I would need at least some technical assistance from people who could sketch out how it should be implemented and who I could rely on to answer questions.