andy@cayuga.Stanford.EDU (Andy Freeman) (08/17/88)
In article <64445@sun.uucp> guy@gorodish.Sun.COM (Guy Harris) writes: >His comment was that "we stupid Americans" had come up with a mail system that >worked between dissimilar hosts, long before ISO had ever done so - in fact, >the Arpanet supported this *before* the advent of RFC821 and RFC822 - and that >*ISO* had "done it differently". Yup. The bits for Arpanet mail between dissimilar hosts were first transported by NCP, the predecessor to TCP/IP, specified in RFC55, which was published in 1970. The message format was specified in RFC733, which seems to have codified existing practice, dated Nov 77. RFC822, the current Internet message format, dated Aug 82, is very similar to RFC733; it clarified some address format issues. (One of the RFC733 implementations here supported RFC822 without change.) The changeover from NCP to TCP/IP started before RFC822; TCP/IP is RFC791 and RFC793, dated Sept 81, so RFC733 worked on top of both NCP (using an option to its File Transfer Protocol) and TCP/IP. (The implementation mentioned above, which used to use NCP, now uses SMTP, TCP/IP's mail transport protocol, to ship bits around.) X.400 could have been an improvement built on existing practice, that is RFC822; it isn't. If the US had "steam-rollered" Europe on this, they'd have had E-Mail between dissimilar hosts 10 years ago; that's a "small" price to pay for European "pride". (That's "small" for vendors who don't have an 822 implementation; making everyone implement a new mail system lets them catch up. The cost to users is much higher.) So, does anyone know how well the X.400 mail experiment worked? Does anyone have a working implementation of the ISO suite (between dissimilar systems) that does what "obsolete" TCP/IP did 5 years ago? -andy UUCP: {arpa gateways, decwrl, uunet, rutgers}!polya.stanford.edu!andy ARPA: andy@polya.stanford.edu (415) 329-1718/723-3088 home/cubicle