Thanks, I haven't seen that!!!

It seems like it is developing a rating system over the existing PGP Web-of-Trust model, and defines an example contract and some common data format schemas so the data can be kept in the existing PGP web of trust.

I would think that using Repcoin as a base layer instead of PGP would work better, as it offers some additional features beyond PGP:
- Transferal of reputation credits - allows for things like a path forward from a semi-compromised PGP key (i.e., it has been exposed but not used yet); however, if contracts and such are linked to a certain key identity, then maybe it would be better to have an explicit key transfer protocol.
- The central blockchain - instead of requiring a web crawler or http archive to maintain backups of reputation values, instead formalize the storage system into something that all interested parties implement. This (like Bitcoin itself) ensures the strength of the network is proportional to the amount of use it gets, instead of the amount of sysadmins thrown at it. Note that this will probably be unwieldy after a certain size, if you allow arbitrary contract data in there. So it might still be an issue.
- No existing web of trust requirement - how do you know how much to trust a new party? With this there is a way to bootstrap your way to a level of trust without needing cooperation from other users. It's the same issue I have with the PGP web of trust - there is no way to initiate a conversation without having a network in place already.

The other ideas in there, i.e. contracts, the rating system, and trust metrics, could be used on top with minor modifications.

Some really great ideas though!