VMS did, in fact have finger back in those days, although it might have not been a stock utility. In Claremont, we had the advantage of having Ned Freed (invent

VMS did, in fact have finger back in those days, although it might have not been a stock utility. In Claremont, we had the advantage of having Ned Freed (inventor of MIME) on the computing staff at HMC, so we might have had a lot of innovative stuff that wasn’t yet universal. JANET did have a gateway to the rest of the world. On interesting thing is that JANET addresses were the inverse of contemporary domain naming, with the highest level part of the domain name appearing first (kind of appropriate for an island where they drive on the “wrong” side of the road). There was also a tendency to use really short names, so a friend/colleague at Imperial College was something like mc @ uk.ac.ic (which was always fun to say out loud). Most of Europe was on a network (EARN) based on the IBM protocols that undergirded BITNET. Switzerland, Austria and Germany followed a stereotypically teutonic and highly rational system of assigning domain names where the first letter indicated the country (D=Germany, A=Austria, C=Switzerland), the next two letters were the region, then two letters for the institution and the last three characters identified the individual machine which was making the best of the situation where there were only eight characters available to identify each machine on the whole network with a flat namespace. BITNET names were a bit more freewheeling, but often were going to be something like institute+machine type, so UICVM for the IBM mainframe at UIC, HMCVAX for the VAX at Harvey Mudd College, etc. There was a Bitnet protocol for real-time chat. Most Unix machines eschewed the Bitnet protocols for UUCP which, at least initially, often required explicit hops from one machine to the next, so you might have to address an email to something like drofnats!caltech!jarthur!dhosek to get an email from Stanford to Harvey Mudd, although by the late 80s, only the most obscure corners of the Unix world required a path with an explicit path. The first generation of the internet did a lot to hide the disparate networking protocols underlying ARPA vs BITNET vs JANET vs UUCP, etc. (EARN and BITNET, using the same protocol effectively acted as a single network at that time), which meant that it was no longer necessary to do things like include gateways in email addresses (WISCVM was one machine that I recall being on both ARPA and BITNET and was a common path to route email between the networks). If you look at old issues of TUGBoat from the 80s, you’ll often see lists of email paths authors would provide to make it easier to reach them including some communication protocols which are long gone like Telex.