62 Mail Sendmail: The Vietnam of Berkeley Unix Before Unix, electronic mail simply worked. The administrators at differ- ent network sites agreed on a protocol for sending and receiving mail, and then wrote programs that followed the protocol. Locally, they created sim- ple and intuitive systems for managing mailing lists and mail aliases. Seri- ously: how hard can it be to parse an address, resolve aliases, and either send out or deliver a piece of mail? Quite hard, actually, if your operating system happens to be Unix. Date: Wed, 15 May 1991 14:08-0400 From: Christopher Stacy CStacy@stony-brook.scrc.symbolics.com To: UNIX-HATERS Subject: harder!faster!deeper!unix Remember when things like netmail used to work? With UNIX, peo- ple really don’t expect things to work anymore. I mean, things sorta work, most of the time, and that’s good enough, isn’t it? What’s wrong with a little unreliability with mail? So what if you can’t reply to messages? So what if they get dropped on the floor? The other day, I tried talking to a postmaster at a site running send- mail. You see, whenever I sent mail to people at his site, the headers of the replies I got back from his site came out mangled, and I couldn’t reply to their replies. It looked like maybe the problem was at his end—did he concur? This is what he sent back to me: Date: Mon, 13 May 1991 21:28 EDT From: silv@upton.com (Stephen J. Silver)1 To: mit-eddie!STONY- BROOK.SCRC.Symbolics.COM!CStacy@EDDIE.MIT.EDU 2 Subject: Re: mangled headers No doubt about it. Our system mailer did it. If you got it, fine. If not, how did you know? If you got it, what is wrong? Just does not look nice? I am not a sendmail guru and do not have 1Pseudonym. 2Throughout most of this book, we have edited gross mail headers for clarity. But on this message, we decided to leave this site’s sendmail’s handiwork in all its glory—Eds.
Sendmail: The Vietnam of Berkeley Unix 63 one. Mail sorta works, most of the time, and given the time I have, that is great. Good Luck. Stephen Silver Writing a mail system that reliably follows protocol is just not all that hard. I don’t understand why, in 20 years, nobody in the Unix world has been able to get it right once. A Harrowing History Date: Tue, 12 Oct 93 10:31:48 -0400 From: dm@hri.com To: UNIX-HATERS Subject: sendmail made simple I was at a talk that had something to do with Unix. Fortunately, I’ve succeeded in repressing all but the speaker’s opening remark: I’m rather surprised that the author of sendmail is still walking around alive. The thing that gets me is that one of the arguments that landed Robert Morris, author of “the Internet Worm” in jail was all the sysadmins’ time his prank cost. Yet the author of sendmail is still walking around free without even a U (for Unixery) branded on his forehead. Sendmail is the standard Unix mailer, and it is likely to remain the stan- dard Unix mailer for many, many years. Although other mailers (such as MMDF and smail) have been written, none of them simultaneously enjoy sendmail’s popularity or widespread animosity. Sendmail was written by Eric Allman at the University of Berkeley in 1983 and was included in the Berkeley 4.2 Unix distribution as BSD’s “internetwork mail router.” The program was developed as a single “crossbar” for interconnecting disparate mail networks. In its first incarnation, sendmail interconnected UUCP, BerkNet and ARPANET (the precursor to Internet) networks. Despite its problems, sendmail was better than the Unix mail program that it replaced: delivermail. In his January 1983 USENIX paper, Allman defined eight goals for send- mail: 1. Sendmail had to be compatible with existing mail programs.
Previous Page Next Page