Path: utzoo!yunexus!geac!lethe!dptcdc!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!decwrl!labrea!polya!weening From: weening@Gang-of-Four.Stanford.EDU (Joe Weening) Newsgroups: news.sysadmin Subject: Re: mail headers Message-ID: Date: 20 Apr 89 01:50:59 GMT Article-I.D.: Gang-of-.WEENING.89Apr19185059 References: <248.244422A4@mudos.ann-arbor.mi.us> <636@dtscp1.UUCP> <1628@ccnysci.UUCP> <641@dtscp1.UUCP> Organization: Computer Science Department, Stanford University Lines: 64 In-reply-to: scott@dtscp1.UUCP's message of 19 Apr 89 19:44:34 GMT In article <641@dtscp1.UUCP> scott@dtscp1.UUCP (Scott Barman) writes: OK. Now that everyone has publicly and privately flamed me (and that most of you have said the same thing--basically), I would like to formally state my position and drop this. Hopefully, those who decide will just consider my side. One has to recognize that the Internet and its protocols have become dominant, and are supplanting any "standards" that ever existed in the UUCP world. Since I come from an Internet-based environment, I'm obviously biased in favor of this. The reference to "those who decide" is pretty meaningless, since there is no serious proposal to change anything here. But I'd like to reply to a few of Mr. Barman's comments. Now I don't understand this "official" bit. Where is it written that "this is the way?" I check through all the Usenet documentation including what can be found in *.newusers and I do not see this as official. I've been also "guided" to RFC 822. Well, the last I heard, RFC stands for Request For Comment and not "this is the standard!" RFCs are used to present Internet standards as well as experimental protocols. They are usually clear about when they do so, but RFC 822 came slightly before the rest of the Internet protocols were in wide use, so it may not say that it is a standard. Later RFCs, however, do say that RFC 822 is a standard for Internet mail transmission. I send mail to a!b!c!user then the From lines would look like From mail-agent date >From mail-agent date remote from b >From mail-agent date remote from a >From my-userid date remote from my-machine It has almost the same information as the Received lines and each machine would produce the "postmark" and append the mail to it before passing it along without having to go in and edit the headers. Also, from this I can write a function that will recreate the return path since this is consistant and many of the Received lines I have seen are not. It is also less verbose and less cumbersome. Now would someone please tell me (without flames) why this is no good? Unix was not a predominant operating system when the Internet stand- ards were designed. The "!" character acts as a comment character on several systems. This may seem like a minor point, but the protocol designers needed to come up with a syntax that would work across diverse systems. It necessarily had to be a compromise of some sort. I'm sorry I don't have time to respond to the many other points in the original message. But please consider the following: 1. Between Internet hosts, there is generally no problem delivering mail. 2. Between UUCP hosts using a path that contains no Internet links, there are also not many problems. So your solution is simple: if you don't want to join the Internet, just find paths that stay away from it! Unfortunately this is no longer practical. So in return for all of the extra connectivity you now have compared to the "good old days" of V7, you are going to have to accept some of the things that make life easier for us Internet folks. -- Joe Weening Computer Science Dept. weening@Gang-of-Four.Stanford.EDU Stanford University