Edit: obligatory explanation (thanks mods for squaring me away)…

What you see via the UI isn’t “all that exists”. Unlike Reddit, where everything is a black box, there are a lot more eyeballs who can see “under the hood”. Any instance admin, proper or rogue, gets a ton of information that users won’t normally see. The attached example demonstrates that while users will only see upvote/downvote tallies, admins can see who actually performed those actions.

Edit: To clarify, not just YOUR instance admin gets this info. This is ANY instance admin across the Fediverse.

    • RightHandOfIkaros@lemmy.worldEnglish
      51·
      1 year ago

      I mostly mean username and password credentials, primarily the storage and handling of passwords. Most people don’t take cybersecurity seriously and use the same username and passwords for every site they log into. Someone steals your Lemmy data, and they can try it on every social media and gain access to everything. Now, I am not one of those people, and my Lemmy account credentials are unique to Lemmy only, but imagine if someone joined and used the same credentials they always do, including for their bank login. There is where the concern is.

      • VlynEnglish
        1·
        1 year ago

        Very quick excursion on storing passwords correctly (which Lemmy hopefully does):

        You add a “salt” to it. For example my password is hunter2 and my user id 123, so my salted password is hunter2123.

        This gets hashed, a hash is strictly a one-way function that transforms one text into another (unlike encryption with no way back). My salted password with SHA-512:

        b5bb1764b7b475f9e25d8a301944124d64aa6c7a36a78f985847a4de44fc1a8f9aed396dcd8a8379f745e60d4a0b368a1a40a5d4192db02ae8cd561a23d5d515

        So it’s not easily possible to find out what my password is even if you have access to the database.

        Old hashing algorithms like MD5 were beat with rainbow tables (just generate billions of combinations and their results, then compare to the password hashes you stole). The salt mentioned on top already makes this approach useless, because each user password is changed with some user specific info.

        Passwords should be safe then, but of course using a password manager and a random password per service is the way to go.