As it was recently discussed a lot of the data you generate while using the fediverse is public. If we're looking at the threadiverse even more of it is public including your votes.
I only know the specifics of Mastodon and mbin, so maybe @rimu@piefed.social @nutomic@lemmy.ml @julian@community.nodebb.org and other devs can chime in here.
Voting on Mastodon is a lot more private than voting in the threadiverse. Only the author of a post (and their instance) knows everyone who voted on a post. Everybody else can retrieve the total numbers, but not the individual votes. Of course this comes with the downside that everyone else has to fetch this data and while the instance could send an Update activity - informing other instances that the numbers changed - Mastodon currently does not do that.
In the threadiverse on the other hand, every single vote gets send around the network, including all the details.
I would like the threadiverse software to get a bit better at privacy. Mastodon is often restrictive with activities for that exact reason and while I do not want to completely screw visibility by not sending anything to anyone, I think the visibility of votes can be improved a lot.
So my proposal would be: votes are only sent to the author of a post. The author then sends an Update activity to their followers and the magazine the post belongs to. The magazine then announces this Update activity to all its subscribers. The post object has to contain the relevant numbers of course and Mastodon and PeerTube use shares, likes and dislikes (PeerTube only). These properties then contain a Collection with a property called totalItems and not a list of the people who actually voted, that would defeat the purpose (looking at you PeerTube)
Because nobody wants to break federation with other software, it would be nice if this could be coordinated between all the threadiverse actors
Except that "fake" voting patterns become impossible to spot.
The way they are now, you can look at how an account votes, when it was created, as well as their posts and comments, to determine whether an account is a legitimate user.
Creating fake accounts that look like legit users is a lot of work. Creating just one, then setting a number a really high, is much, much easier.