[AgentME]

Who would decide which BOINC projects work with Bitcoin? What if I made a BOINC project for calculating the number 4 and did it before anyone else noticed it was available? How much bitcoin should that be worth? There have to be ongoing decisions about what BOINC projects are accepted and how much they're worth. Who makes those decisions?

More realistically, what if I privately know a fast algorithm for an obscure type of protein folding, and then I push for a new BOINC project to be accepted that focuses on the specific type of problem? I'll be the only one that can mine it efficiently for a long time. If I calculated my expected profits, I would probably see that it's worth it for me to pay a lot of money to lobby the people in charge of the BOINC-coin decision process to get my project accepted. Unless BOINC-coin only uses a fixed set of projects over all time, or is only worth negligible amounts, then lobbying like this is going to influence what projects it accepts.

Even if BOINC-coin uses a fixed set of projects, then there's a problem of what work units are given out. I imagine many projects work like SETI@Home where some data is given out to users to process. If processing the data is worth money, then it might make sense financially for me to artificially construct datasets that I've already solved, and bribe the SETI@Home administrators to insert my dataset into the worker queue, which I will then quickly "mine" for BOINC-coin.

A BOINC-coin isn't fully decentralized. It requires trusted people to be in control choosing which projects are worth it and to secure the authenticity of the data sets. Bitcoin is about minimizing the need for trust in administrative systems like this.

Even if you solve the above problems, the proof-of-work system needs to be hard to compute but cheap to verify. Many BOINC projects' work units aren't cheap to verify. They just have several users redundantly recalculate the same work units to check that they get the same answers.