On 5/26/2020 5:14 PM, James Cook via agora-discussion wrote: >> Obviously, answering all of these questions is a lot of work and the >> goal of my thesis, but if anything quickly comes to mind, I'd >> appreciate any contributions.
One thing just occurred to me. I've never seen this option used, but if we interpret or legislate so that the archives determine the datestamps or order the messages, it's worth considering what would happen in the ultimate R478 backup situation for "public"; that is, if a message is "sent to all players and containing a clear designation of intent to be public" thereby bypassing the archives entirely? For such a message, all of the recipient datestamps could be slightly different, so only the Sender's headers could provide a specific time. So that edge case (but important backup if all fora break) might be worth discussing, to see if any hypotheses you favor would work for that case. > Blockchains are an elaborate solution to the problem of agreeing on a > set of public events and in what order they happened. This seems a bit > related to Agora's sometimes-problem of agreeing on what happened and > when. I'm not saying it's directly relevant to Agora, though; we > already have reasonably trustworthy mailservers and if we really > wanted an unambiguous standard we could probably just legislate that > the list archives entirely determine the set and ordering of messages. Ordering of two different messages is much rarer of an issue (because messages are often subject to similar delays on a given day, so order doesn't change too often when you change the datestamp standard). It most often comes up when beating set deadlines like voting or bidding, especially if (like in bidding) there's real gameplay value to waiting until the last possible second. The argument against "archive ordering" is that most email clients do not readily display the "list archive times", instead displaying the initial Sender stamp (earliest possible choice). Thus, it's beyond a reasonable burden for officers to have to dig for those alternate stamps for every email they process. If the archives re-wrote the dates the way they now re-write the sender (since the fixes in Dec/Jan) that would fix that complication. But IIRC, the last time this came up we were collectively squeamish about re-dating things like that (though maybe now that we've lived with the sender re-write for a bit, without the sky falling, we'd be more cool with that?) -G.
