[tz] "non-standard" timezones

[Paul Eggert]

On 01/04/2013 02:19 PM, Shawn Steele wrote:

> Are there cases where people are using their own Timezone data?
> (I'm thinking like a company or research station or something that
> doesn't snap to some "normal" tz for some reason)

For a research station, as far as I know common practice is to
set the TZ environment variable a la POSIX.  So, for example,
if the folks in the new Bharathi research station in Antarctica
decided on their own time zone, they could set TZ='BHAT-5', say.
That way, they don't need to futz with the time zone data.
(This is hypothetical -- I expect that they actually just
use Indian time, with TZ='Asia/Kolkata').

> For non-standard timezones is there a recommendation for naming conventions?

There isn't, but if there's a need we should probably establish one.
"Etc" is probably not a good prefix, since it's in use in the
standard list.  "Extension", maybe?