Eric Allman: Hello. I have man­aged to live an extra­or­di­nar­i­ly lucky life­time. One of my first strokes of luck was hap­pen­ing to arrive at UC Berkeley the same year that Unix did, 1973. And so I got in on the ground floor there. And there were some extra­or­di­nary peo­ple at Berkeley as well, which made Berkeley Unix into some­thing that was close­ly tied to the growth of the Internet.

When I start­ed, of course, there was no capital‑I Internet. There were, how­ev­er, even­tu­al­ly a bunch of small­er net­works that did­n’t talk to each oth­er, and I decid­ed that they need­ed to talk to each oth­er at least for email. And we had ARPANET, we had some­thing called Berknet, we had UUCP, and lat­er PurdueNet and CSNET and so forth. And so that was the lowercase‑I inter­net. That’s what John Quarterman calls The Matrix.” And that was deliv­er­mail, that was to make those things work together.

Bill Joy was one of the prin­ci­pals on the Berkeley Unix ver­sion that had the first TCP/IP stack, and he need­ed some­body to write the mail serv­er. And some­how he con­vinced me to do this, which if I had real­ized at the time I nev­er would have done it but that’s true of many things. So that was the con­ver­sion from deliv­er­mail to send­mail. The goal of send­mail was still to try and uni­fy things, pull things togeth­er, and not nec­es­sar­i­ly to do every­thing, which many sys­tems these days appar­ent­ly feel the need to reim­ple­ment absolute­ly every­thing, which seems waste­ful to me.

I also end­ed up work­ing on some­thing called sys­log, which is the basic sys­tem log­ging facil­i­ty. I did that as part of the send­mail project, but inten­tion­al­ly to be gener­ic. And that in some sense has been more suc­cess­ful than send­mail even though nobody seems to know that I wrote it. It’s in pret­ty much every print­er, wire­less access point, you name it. It’s…it’s just there. And so I’m actu­al­ly quite proud of that, even if I am anony­mous on that side.

There are of course way too many peo­ple to thank, so I’m not going to try and thank all of them but I’m going to call out just a cou­ple of names. Bill Joy, of course, who if he had­n’t talked me into this I would­n’t be here today. Michael Stonebraker and Susan Graham, who were both at var­i­ous times employ­ing me to do some­thing else and yet still man­aged to allow me to work on send­mail. Bryan Costales, who lit­er­al­ly wrote the book on send­mail and made it much more popular—documentation is impor­tant, I’m here to tell you. And final­ly my hus­band Marshall Kirk McKusick, who sup­port­ed me through a lot of this thing, includ­ing all the times I spent way more time with my com­put­er than I did with him. Thank you. 

Help Support Open Transcripts

If you found this useful or interesting, please consider supporting the project monthly at Patreon or once via Cash App, or even just sharing the link. Thanks.