I cannot really tell you how to do it as I'm not that fluent in the workings of crypto when looking under the hood. However, I would strongly encourage the devs to really try to lock down ways to mitigate cheaters better than Gridcoin has done.
Collatz had to completely rebuild their site/servers/etc... because at least one Gridcoin user was cheating on a large scale under one of their pools.
https://boinc.thesonntags.com/collatz/forum_thread.php?id=15#190The responses from the Gridcoin team did not sit well with the community as it basically came down to telling the community to fix their projects and to not tell Gridcoin how to run theirs. Science was indeed affected and has to be re-ran. This event was the straw that broke the camels back and essentially caused Gridcoin to be banned from the Annual Pentathlon which is the largest competition in the BOINC world amongst teams.
Gridcoin users basically had coin stolen from them because of the cheater getting rewarded. I do not know if they ever resolved any of it as I don't follow their threads. Collatz lost the old forums that had a lot of the details in them.
And it appears that a possible cheater has appeared yet again at another BOINC project -
http://gene.disi.unitn.it/test/forum_thread.php?id=216#1325On a positive note, have you guys considered possibly adding GPUGrid to your list of projects?
http://www.gpugrid.net/Thanks for the post and the links. So from what I can see, in both instances of the cheating that went on in those projects in that community, the user took advantage of an attack vector in which a vulnerability existed on the specific boinc project's server side, in the validation process, which made it possible to cheat on the actual result submission (IE it fooled the server into accepting the result of the workunit). Another words they studied the traffic that is posted by boinc that allows the work unit to be solved and falsified it. Note that there is no example of cheating outside of those few projects. (Thats partially why we picked Rosetta originally, it was partially discussed in the swongle discussions when we were in testnet with cancer mining- Part of the due diligence was to find a project that would be extremely hard if not impossible to cheat. This is because if you look at the manual from Rosetta commons, the work you are doing is so complicated (it requires a 3mb database on the client side to refer to), with 300 scientists on the back end, I don't think its possible to falsify a low energy 'result' without doing more effort than the client does to compute the low energy result. Also in both cases of the cheating note that the gridcoin user diluted their own project compensation - another words a report could have been written that showed a discrepency in task-turnaround/cobblestone-work, picking up the bad user. In that case if that happened here we would have to disable the project and regroup.
As far as adding other projects, I feel we should stay focused with what we have now - these projects are related to our healing aspect (we picked Rosetta because it was based on potentially curing cancer), and WCG with their AIDS research. Although I'm interested in using clock cycles for the benefit of humanity, I'm not interested in expanding our mining side into a volunteer computing platform, primarily because I would rather support a high quality project and maintain the healing aspect or facet of biblepay. We have a lot of things to do on the roadmap, and one behind the scenes is investigating partners in the financial sector that also have some options for us in replacing proof of work with other cool cutting edge ideas (like analyzing financial markets etc).