emv@math.lsa.umich.edu (Edward Vielmetti) (08/27/90)
I'm searching for a complete set of RFCs, including the real old ones from the early 1970's. In particular, I'd like to be pointed to a place where I can get copies of all of the RFCs mentioned in Appendix III of RFC 959, "RFCs on FTP". What I'm trying to do is to gather historical material for an RFC on anonymous FTP. RFC 959 doesn't talk about where the convention of specifying "anonymous" as a password comes from, nor was there anything that I could easily dredge up in any other RFC that I have found. Any relevant ephemera which would help me pin down how things have changed and developed over time would be most welcome. --Ed Edward Vielmetti, U of Michigan math dept <emv@math.lsa.umich.edu> moderator, comp.archives
clements@bbn.com (Bob Clements) (08/27/90)
emv@math.lsa.umich.edu (Edward Vielmetti) writes: >What I'm trying to do is to gather historical material for an RFC >on anonymous FTP. RFC 959 doesn't talk about where the convention >of specifying "anonymous" as a password comes from, nor was there >anything that I could easily dredge up in any other RFC that I have >found. To the best of my knowledge, I invented the use of "anonymous" as a username (not as a password). This was in the FTP server for TENEX in the NCP (pre-TCP) days. Probably late in 1972. The implementation was simple. You could read (or write) any file that was world-readable (writable) in the file system. The server logged all transactions, for both anonymous and "real" users. It logged the passwords used by "anonymous", which were supposed to be real user idents so one could contact the user if he was having problems. There was a configuration file and there had to actually be a username "anonymous" in the system password file, otherwise the FTP feature was disabled. So sites that didn't like the idea didn't need to have it turned on. This was largely as a debugging and testing aid. Us early FTP implementors could test against each other's implementations without needing to actually get a password on every machine. (Machines weren't all Unixes in those days, so there were lots of implementations and lots of word lengths and lots of file types.) Other users found it useful, too, so we left it in. Another feature added in that version was a "NUL" device. Reading from it gave you a megabit of data, for speed testing. In those days there were only a half dozen sites on the ARPANET and security was not a big problem. We joked that anyone who could spell "anonymous" was probably OK. A simpler time. >--Ed >Edward Vielmetti, U of Michigan math dept <emv@math.lsa.umich.edu> >moderator, comp.archives Bob Clements, K1BC, clements@bbn.com (w) 617-873-3612
PADLIPSKY@A.ISI.EDU (Michael Padlipsky) (08/29/90)
Bob-- What a relief. Here all these years I'd been feeling vaguely guilty about "Anonymous FTP", having gotten the impression somewhere along the line that one of the TENEX jocks at the NIC had done it as a generalization of the USER NETML, PASS NETML trick I'd invented in RFC 491 to let netmail be "free" but still accounted for, and now I find you willing to take the blame-- er, uh, credit. (I could still swear the line about anybody who could spell "anonymous" was OK was one of mine, though.) One thing does bother me: if you had the anonymous trick in the arsenal at the time, why didn't you make "sndmsg" try it when it got the "You must login first" code from Multics's Server FTP, instead of making me go off and come up with RFC 491? Not that I'm disputing your "priority" to the invention of Anonymous FTP, of course, just wondering. (I think the timing's off for you to have been the one who generalized the canned-account-for-mail gambit if you were using anonymous for EARLY debugging, so presumably it wasn't the case that anonymous wasn't in the arsenal at the time and I'm still off the hook.) cheers, map -------