Competition Agoric

Agoric is Mark Miller's take on a smart contract language: E-rights over ocaps over vats over (blockchain, solo, quorum, etc.).

While Agoric's approach is similar to Glow's in some ways (abstraction, layering), in other ways, it is very different in ways that inspire me doubt:

1. Agoric is object-oriented when Glow is functional, which will make it harder to prove correctness of programs;
2. Agoric requires three complex layers of abstraction before they can even start writing DApps, whereas Glow is useful at each simple layer of abstraction, for a growing set of DApps;
3. Agoric is retrofitting a previous model (E) on top of blockchains, rather than trying to seek what the essence of either blockchains or DApps is;
4. their vats provide isolation against process or network errors (and quorum vats provide much needed redundancy that is on our roadmap but that we don't handle yet), but they have no builtin support for incentive alignment in an adversarial environment.

In the end, instead of being able to provide a language that minimizes the complexity of writing DApps, they end up introducing a complex, arbitrary language (E, just with JavaScript syntax) in between the contracts and the blockchain, with its impedance mismatch, with no obvious gain. The vat layer is very relevant; the four dimensions of rights in erights are interesting, but not always relevant.