by Darius Kazemi, September 30 2019
Let's settle this issue
RFC-273 is titled “More on Standard Host Names”. It's authored by Richard Watson of SRI-ARC and dated October 18, 1971.
The technical content
Hooray, we're back to the host name drama! Even though the last few RFCs have been from early 1972, we are now back to mid-October 1971, just six days after RFC-247, the last volley fired in this war, was published. To refresh your memory:
- RFC-226: Standardization of Host Mneumonics [sic], Karp
- RFC-229: Standard Host Names, Postel
- RFC-233: Standardization of Host Call Letters, Bhushan
- RFC-236: Standard Host Names, Postel
- RFC-237: The NIC's View of Standard Host Names, Watson
- RFC-239: Host Mnemonics Proposed in RFC #226, Braden
- RFC-247: Proferred Set of Standard Host Names, Karp
Watson is again representing the Network Information Center and notes that no conclusion has been reached yet. He proposes a new set of considerations for a host naming scheme:
- there may eventually be several hundred sites on the network so the scheme needs to account for this
- abbreviations are necessary to simplify typing
- names should be at least somewhat identifiable with current references used
- and a scheme must recognize “
people's strong identification with historical names associated with their project“
I would like to point out that this is an entirely different set of considerations than the ones he proposed in RFC-237 just a few weeks prior. Those considerations were highly prescriptivist, saying things like a host name should be based on an institutional name rather than a computer name. He has jettisoned these fully and instead embraced this more humanistic set of considerations.
His proposal for a specific format is a hybrid of all suggestions to date, and consists of the following elements, in order:
- an institution mnemonic (4 characters max)
- a dash,
-(decimal ASCII code 45)
- a station mnemonic (no character length maximum but short is good)
- he offers some suggestions as to what you might use but is not prescriptivist, it could be anything from a computer name to a department name to a common project name
These three elements make up the “Formal Name”.
And last there is a “Nickname” which he suggests should be chosen “
to make life easy for people having to learn them”. The idea is that while plenty of programs only use the Formal Name to connect to hosts, Telnet programs, which are user facing, should also be programmed to accept Nicknames and translate them to Formal Names internally.
Watson says the only reasonable answer to the question of “who picks the names” should be that each host gets to pick their own name. He ends the RFC by imploring readers to “
settle this issue as soon as possible, say by November 5”.
How to follow this blog
I'm Darius Kazemi. I'm an independent technologist and artist. I do a lot of work on the decentralized web with ActivityPub, including a Node.js reference implementation, an RSS-to-ActivityPub converter, and a fork of Mastodon, called Hometown. You can support my work via my Patreon.