Natural breaking of large subgraphs

[PaulRBerg]

Conjecture

As argued in #114, I am now advocating for a more experimental and pragmatic approach to building our rollup (worrying less about research, and building more).

In that spirit, I would like to ask this question: what if the theoretical problems we have laid out so far are only just that - theoretical?

Specifically, what if UCG will be prevented by the natural propensity of subgraphs to decompose?

It seems quite likely that in a large subgraph of composed streams, there will be one user who will, eventually, break the chain by not topping up on time.

Objections

This conjecture, even if true, won't rule out of the following cases:

1. Purposeful malicious spamming
2. Netflix-level vertices with a gazillion incoming streams

Case no. 2 can be glossed over for now - let's prove that the network works with a low number of users. But case no. 1 cannot be ignored in a blockchain environment. I reckon that the solution is to design a sound economic model for gas prices (e.g. refer to the complexity factor).

RFC

CC @sablier-labs/sabvm