I'm really not sure why the author is expending so much energy on the "@". It was the clearly obvious symbol for so-and-so at a given machine. What else could T

I'm really not sure why the author is expending so much energy on the "@". It was the clearly obvious symbol for so-and-so at a given machine. What else could Tomlinson have used from the basic teletype character set? And, as I commented to the author on Twitter, his story of how email came to be glosses over the low-level technical details that would interest at least the HN crowd. (I realize that's not the point of his story. But he touched on it, anyway.) For example, in the pre-TCP era (NCP was the host-to-host protocol), email was delivered by FTP'ing to the given site and appending your message (saved as a file) to the user's publicly-accessible mail box file (which was usually append-only, so you couldn't read it). We were pretty trusting in those days. That meant your email name was your username on the site; for us TOPS-10 users, that means our email address was the usual 18+18 octal project/programmer number. I think mine was 377,6001@harv-10 (no domains in those days). The lucky guys on Tenex systems (the precursor to TOPS-20) and Multics, etc., had real names. I also remember being on the first ARPAnet mailing list (1972?), which was, of course, a discussion group on mail clients' UX and mailing lists, combined. Talk about bike-shedding and flame wars! But of course, none of us had ever experienced anything like a mailing list before, so we had to discover all those issues ab initio.