Nano Protocol

The difference between a Principal Representative and a regular Representative with a voting weight of < 0.1% is that the votes of the Principal Representative are re-broadcasted by other nodes, and the votes of the regular Representatives are only communicated directly by the regular Representative.

What this means is that, if we had a network where every node holds less than 0.1% of the voting weight, then the practical difference is that a given node would need have direct communication with each voting node to collect their vote until they have enough votes to exceed the threshold - so we can't take advantage of the shortcut of re-broadcasting.

Is my understanding correct? Or is there something more to the distinction between the types?

This is a question I asked on the discord chat, but I'm also placing it here in case anyone else can contribute.

I was hoping someone could help me understand something.

One of the common 'criticisms' of the nano protocol and the network is that the transactions-per-second (TPS) are small. A common response that I have read is that with the growth of the network comes an improvement of the hardware and more development resources go into improving TPS. I agree completely with the development portion of the response, but I am having a hard time wrapping my head around why an improvement of part of the infrastructure would lead to higher TPS.

The way I see it is:

• With wider adoption, the total number of nodes increases, and voting power becomes more distributed

• With more distributed voting power, more nodes will need to vote on a block before consensus is reached

• The rate at which a transaction is cemented depends on the slowest voter that contributes to reaching the quorum

With increased adoption there will be more "high quality" nodes, but the rate bottleneck is set by the slowest voting members. I think that increasing the number of nodes in the network decreases rather than increases the TPS, and that the effect is unlikely to be offset by the presence of some stronger nodes.

Is this view accurate? Or am I missing some important bottleneck in the process that is overcome by having some very strong nodes?