Choosing a name for your computer
tools.ietf.org
tools.ietf.org
Naming after the purpose of the box also gets messy if you're not strict about renaming when the machine is repurposed. And if you group by "theme groups" as the RFC suggests, you're still stuck renaming a machine if it's no longer one of your 7 dwarfs (database servers, or whatever). Should the name convey meaning at all, or just be a random label?
A developer I know doesn't see the problem with just having the full location (state, data center, rack, rack location), which I'm against, but we haven't been able to agree on a good middle ground approach that scales. Is location data in a name universally bad?
This is more advice for "snowflake servers" and workstations.
As for geographic location I had (viewed) a bad experience college when a server in the main server room started DOSing another box in the same room. The culprit was ruled out for a very long while 'cause it’s name claimed that it was all the way across campus - so it couldn’t be the problem. Because of this I’m with the paper on the no geography in names point. Geography in names causes thinking errors in people.
For servers, I have seen numbered-names work quite well in most of the projects that I've worked on -- www01, www02, app01, app02, db01, mq01, mq02, so on and so forth.
Workstations my university were named by location -- swlab01 or even by the hall numbers, 496-01, 496-02.
Unfortunately, I've never come across more creative themes of hostnames like colors, or planets in Star Wars or football players etc. But I will definitely re-read this RFC next time I get to setup a lab or a series of servers from scratch.
They were never moved? Or did they have to be renamed then?
[1] http://www.reptile-database.org/db-info/SpeciesStat.html
Then owing to a stupid management decision all in my group were required to move to machines of an unfamiliar and inferior sort. I named this one "crickhollow".
I don't think anyone in management got the reference but I felt better.