4 ms·
The problem is that it is trivial to delete anything after (and including) the + from an e-mail address. I've had this happen when trying to use this approach.
by techsupporter 12y ago
The problem is that it is trivial to delete anything after (and including) the + from an e-mail address. I've had this happen when trying to use this approach.
What I do instead is this: I have a subdomain of my regular domain, say "me.example.com." That subdomain is a catch-all; anything sent to it goes to my regular account at "example.com," though I have a Sieve rule to filter those messages to their own folder. Companies and so forth get addresses like companyname@me.example.com or, if I'm giving one in person, myname3@me.example.com so it doesn't seem odd.
At that point, they can't tell the difference between that being a tagged address or not, so the entire address has to be sent untouched. I can then do whatever I like with it, namely trashing the message in a rule. Using a subdomain is preferable for keeping the generic spam down since those aren't as easily discovered.
- wpietri 12y agoThat can definitely be better if you have full control over a domain. But what stopped me from going that far is dictionary attacks. There are plenty of spammers who will just try hundreds of common-ish addresses at a domain. I considered going even further and building something where I'd register each new address as I gave it out, but for now, just adding a tag on to my regular address has been good enough for me.