I searched for files on saildart that include the text SRI-ARC . The oldest one I can find is file NET.SYS dated 1973-04-08: https://www.saildart.org/NAMES[NET,SYS]4 It appears to be an older version of the 1974-04-04 file I mentioned in my previous message. My apologies for top-posting. I'll stop posting now as I don’t have any ideas for finding hosts files older than that one. -----Original Message----- From: amp1ron at gmail.com Sent: Thursday, December 5, 2019 11:19 PM To: 'Rob Gingell' ; 'Larry McVoy' Cc: coff at minnie.tuhs.org Subject: RE: [COFF] ARPAnet now 4 nodes I see there's a LOT of hosts files in at least several different formats archived on saildart. A duck duck go search for "site:saildart.org hst.net filedate 197" found me lots of files dated in the 1970s. In the first few pages of search results there's this 1974-04-04 file https://www.saildart.org/NAMES%5BNET,SYS%5D3 . At least that's in the mid-70s date range you're looking for. And maybe there are even older files on saildart. I've pulled out the entries from that file with the lowest 20 numbers and ordered them by number. Between number 1 and 23 in that 1974-04-04 file hosts 17, 20, and 22 are missing. So even if hosts hadn't yet been renumbered, by 1974-04-04 it appears some hosts were dropped from at least this file NET.SYS . ; Here we define the names (long and short) of the sites and their numbers X ,NMC,1,0 X ,NIC,2,noslf!noelf X UCSB,UCSB,3,0 X UTAH,UTAH,4,noslf!noelf X ,NCC,5,noslf!noelf X ,MLTX,6,0 X ,RAND,7,0 X SDC,SDC,8,0 X ,HARV,9,0 X ,LL67,10,0 X ,SAIL,11,noelf X ,ILL11,12,0 X CASE,CASE,13,0 X ,CMU,14,0 X ,AMES,15,noslf!noelf X ,AME67,16,0 X RADC,RADC,18,0 X NBS,NBS,19,0 X TINKER,OCAF,21,0 X USC,USC,23,0 -----Original Message----- From: Rob Gingell On Behalf Of Rob Gingell Sent: Thursday, December 5, 2019 10:26 PM To: amp1ron at gmail.com; 'Larry McVoy' Cc: coff at minnie.tuhs.org Subject: Re: [COFF] ARPAnet now 4 nodes On 12/5/19 5:19 PM, amp1ron at gmail.com wrote: > Maybe some of these hosts files that Lars Brinkhoff gathered together will help: > > https://github.com/ttkzw/hosts.txt Thanks for the pointer. I had come across those. The trouble is, for the NCP era, there's only one host file in the collection. The rest are all from the post-Internet transition (and thus the numbers can't be inferred to convey a probable chronological sequence.) And for the one from the NCP era, it's the one that only has the first page and so it's missing a bunch of stuff. (It's not really a HOSTS.TXT file but a prettified annotated edition with other information, and so the file in the repository is a PDF of a scanned physical printout.) Still even that one page adds some information. From the information exchanged previously we had hosts 1 through 4, and then host 13. And the likely matches for about two dozen numbers. And the fragment from the one page in the repository adds 5 (though it's clear from the comments that it was a recycled number), 9, 12, confirms 13, 14, 15, 16, and then a smattering of others up to 232. Some of the liaison names are tickling memories of long ago acquaintances! I had thought that once upon a time there was an archive of a mid-1970s TENEX distribution, like 1.33 or 1.34. The distribution might have embedded a stale HOSTS.TXT file that would have been complete for the time. But I haven't managed to find it again. Still none of this really answers Larry's query in a satisfying way. I imagine somewhere there's just a ledger that has the answer to the question I thought he posed about who showed up when with what on the ARPAnet. The collected papers of someone like Jon Postel might have something of that nature (but a brief search doesn't reveal an archive literally like that) but then substantial body of his work lives on in the RFC library.