Frankly, I don't remember what Rob did but you might want to poke around the BBN archives too which Warren should also have. Gurwitz might have written something before the UCB code came about that used ftp. At CMU we used rsync between the UNIX boxes for files like the host table, (that where rsync was originally written for those that's don't know); but how it was kept up to date from the 10's -- I don't remember. As Lars said, MIT would have used ITS as its master before the DNS was birthed, but how the UNIX boxes got it from there is unknown to me [Noel do you know?]. Clem ᐧ ᐧ On Fri, Apr 19, 2019 at 5:53 PM wrote: > Replying to two messages here: > > On Fri, 19 Apr 2019, Lars Brinkhoff wrote: > > Here is a collection, but most of them are from after that date. > > > > https://github.com/ttkzw/hosts.txt > > Thanks. I should have mentioned I looked there. It doesn't have it as > far as I see. It has the same data for the pre-1982 files I already > have (but they are not RFC 608 style). > > But it was interesting to see the three MIT/Stanford versions. I think > they have a different format unless I misunderstand. Since I cannot find > SRI-NIC examples (before RFC 810) I don't know for sure. > https://github.com/ttkzw/hosts.txt/blob/master/pub/hosts/19830119/HOSTS.TXT > > If you can point me to PDP-10 ITS routines, I'd appreciate it. > > On Fri, 19 Apr 2019, Clem Cole wrote: > > 1) I am looking for parsers for ancient (pre mid 1982) > ... > > Got to Warren's archives for BSD 4.2 and look for the htable(8) and > > gettable(8).? I believe the parsing routines will be in htable(8). > > Thanks. I have them already (including 4.1c.2). They are the newer > format. I had looked at htable before. I just looked at my sccs for it: > > MRs: > COMMENTS: > date and time created 82/10/20 21:26:49 by sam > > 4.1 #ifndef lint > 4.1 static char sccsid[] = "@(#)htable.c 4.1 (Berkeley) 10/20/82"; > 4.1 #endif > 4.1 > 4.1 /* > 4.1 * htable - convert NIC host table into a UNIX format. > 4.1 * NIC format is described in RFC 810, 1 March 1982. > 4.1 */ > > I was hoping there was an early revision for previous format there, but > that is the first revision I see in SCCS. > > Thank you for reminding me and thanks for pointing me to gettable. I see > that is a TCP client implementation of the RFC 811 NIC Internet > Hostnames Server. Awesome. I will be writing about it. > > Just noticed the long-obsolete manpage has an error; it says: > > Gettable is a simple program used to obtain the NIC standard host > tables > from a ``nicname'' server. > ... > Gettable operates by opening a TCP connection to the port indicated in > the service specification for ``nicname''. > > It is not a "nicname" client. It is a RFC 811 "hostnames" client. > > The first code before 4.2BSD has: > > sp = getservbyname("nicname", "tcp"); > > The later code has: > sp = getservbyname("hostnames", "tcp"); > > 4.1c.2 /etc/services had: > > nicname 101/tcp hostname # usually from sri-nic > > 4.2BSD /etc/services had: > > whois 43/tcp nicname > > hostnames 101/tcp hostname # usually from sri-nic > > nicname is RFC 812. > > but the manpage never got fixed (well I didn't look after FreeBSD 1.0 > which is the newest version of gettable I have). > > I was hoping that gettable(8) had HNAME/HADDR query commands then I > think it would be a very early network name lookup tool. But it just > had the ALL support to return entire host table. > > Still looking for the pre RFC 810 tables from SRI-NIC ... > >