I posted this thread, without issues. A year later, I finally got a really insightful reply that wholly answered my question well:

https://lemmy.sdf.org/post/39819472/25666072

Now it is “Removed by mod”. But I am the mod… the only mod (wtf?). No, I do not have multiple personality disorder as the modlog shows.

So somewhere, someone booted that user. That had a ripple effect of destroying all that user’s content. So then I found a “restore content” button, used it, and had to give a reason. I thought I would just be sending myself a useless alert, but it actually restored the content.

So it’s fixed to some extent. But I would like to understand more about what happened. I cannot tell from that user’s modlog who banned them, or even which instance did the banishment. It may be true that it was an anti-spam control, and perhaps rightful, but it’s a bit fucked up how it killed all that user’s content when obviously some of it was useful.

I have seen this before in communities I moderate where a comment is gone because a user is banned. I never looked further. But now it’S clear this is not optimum. It was only in the case at hand that a user’s comment was useful enough that I remembered it, and it happened to be recent.

Clearly this is a bug: I am given the option to “restore comment” but I cannot review the comment before clicking that. So how can I first give a reason for restoring the comment if I am restoring it just to see if it should be restored?

  • evenwichtOPEnglish
    4·
    2 days ago

    Banning an account means his whole content is deleted (that’s how it was done since the 90s)

    Lemmy did not exist in the 90s. When a usenet account or email account ceased to exist for any reason, the content on news servers and mailing lists was separate. Software can be crudely designed (web-based in particular), or it can be well designed.

    Saying “(that’s how it was done since the 90s)” implies a conventional wisdom fallacy. It’s irrelevent how it was done in whatever span of time. Tech was often designed better before the 90s. Advancements are obviously made but I see design approaches worsening.

    Most importantly, the tool /can/ be smarter. A good tool might remove all content as a starting point, then a spamassassin-like process could run that works out likely ham, which could then be either reinstated or flagged for human review, depending on resource constraints (both human and machine). A blunt remove all is important to have but crude and reckless if it’s the only option.

    Spammers do write some random content to pretend they are not bots, again a fact.

    And? Intent doesn’t matter. The end game is still about separating spam from ham. Your claim just supports the view that there is some ham to save.