…deleting records could cause big problems. Referential integrity across database tables simply wouldn’t allow it… it would cause a resonance cascade.
So, to get around the problem, a lot of places simply “overwrite” records when they are deleting them. They replace certain fields with garbage so the structure of the data remains, but the human elements are no longer present. At the heart of those “certain fields” are email addresses, the most widely used identifiers on all the web. And that, dear reader, is how we got to this cursed discovery.
I saw a discussion on the internet where someone mentioned that they deleted users in their app by overwriting their email addresses with
$somethingRandom.com. Mmm, I thought - I wonder how common of a thought process that is? I bet whoever ownsdeleteduser.comgets loads of emails!I decided to check it out, but to my genuine surprise - no one owned
deleteduser.com, so now I do.Source [web-archive]
These companies sure have an interesting interpretation of what "deleting” means.
This is a common practice in databases, you just mark record as deleted and make data unidentifiable.
Idk man in our databases delete is literally an SQL delete most of the time (or becomes that at some point). Though I guess we’re not the data-hoarder type company…
We aren’t the data hoarder type either and doing this would break half the functionality. It’s old systems not built with data deletion in mind that rely on there for example always being a customer associated with a transaction. That customer might not have anything other than an ID anymore, but it needs the ID. And you’re usually not legally allowed to just delete financial transactions.
Let’s say your deleted customer paid for some goods and you delete this data, how do you know what was sold, to what number of customers and for how much etc?
It’s true! We have to keep massive amounts of sensitive customer data that keeps getting leaked! Commerce is impossible without it! SQL predates money!
Once more, the customer identification data is striped out, there is nothing to leak, unless you have retain mandatory data. And yes, you have to have financial records for your company or how do you think it works?
Once more, the customer identification data is striped out, there is nothing to leak
Well, at least you solved leaks by pretending they never happen.
And yes, you have to have financial records for your company or how do you think it works?
How do you suppose it worked before databases?
Do you think that ledgers stored customer data for everyone who bought anything?
Considering he’s getting emails with full addresses and other PII, I wouldn’t say they’re making it “unidentifiable”.
We’re not talking this specific case, we are talking in general. In that case the data was obviously not replaced properly.
.invalidis a TLD which isn’t publicly routable.I bought a similar domain for setting up wildcard forwarding addresses. Let’s just call it
funnyemail.comI have Simple Login set up so I can create (and destroy) simple throwaway email addresses using
*spam@funnyemail.com, likelemmy.spam@funnyemail.comwhen I need a quick account.What I wasn’t expecting is that a couple of times per year I get something unexpected to
nospam@funnyemail.comorihatespam@funnyemail.com. Most recent one was an e-vite to a child’s birthday party.I’m kinda curious to set up a folder for
*@funnyemail.comjust to see who else is tapping into it for more than justaddresses
