New Zealand/Place names

From OpenStreetMap Wiki
Jump to navigation Jump to search

New Zealand Place Names

This page is concerned with establishing consensus on how to deal with New Zealand place names.

It is for general use but is being developed in support of the LINZ data import.

There are two main issues:

  • Localization (l10n) and language translations
  • Establishing town size to tag name. (e.g. what population size before a village becomes a town?)


See the main OSM page: Key:name#Localization

  • The ISO 639-1 two letter code for Māori is "mi".
  • The ISO 639-1 two letter code for English is "en".

Together with the country code, we have mi_NZ and en_NZ locales.

An introduction to UTF-8 and Unicode

Non-latin letters, including macron variants of latin letters such as used by te reo Māori are stored in what is known as UTF-8.

Some introduction to the nitty gritty of how that all works:

Unicode Macrons, mainly for Te Reo Māori
Character Encodings
Māori spell checker for OpenOffice howto

How to tag

name=the default name, used locally
name:mi=the name in Māori
name:en=the name in English

If a localised name such as name:mi is added, always add the other one (name:en) as well, even when it's the same as name. The reason is that if a user has set their locale to Māori (preferred) and English (fallback), a map app can't determine the language of name and will always prefer name:mi and name:en to name, and if name:mi is missing, it will display name:en, even when the Māori name (that the user prefers) is in name.

If all three names are the same, only name is needed as there is no ambiguity.

name should ideally be the name as spelled on the ground, mainly on traffic signs (these are the most useful for navigation). However, sometimes it can be ambiguous, in that case, it's better to prefer the name gazetted by NZGB. If the name on the ground is different to the name gazetted by NZGB, the official name can be put in official_name=*; some mappers also change name in that case since new traffic sign will use that spelling while others prefer to keep the names on the ground until new signs start appearing.

For names checked with the NZGB Gazetteer, add ref:linz:place_id=* using the number in the last segment of the URL, e.g. if the URL is, the ID is 54403 and the tag should be ref:linz:place_id=54403.

TODO: Open question: when to use which? for example:

name:en=Mount Egmont
name=Kaikōura or Kaikoura ?

Converting text

If you need to transform from ISO-8859-1 (basic latin) encoding to UTF-8 (supports advanced macrons) encoding environment, you can use the GNU iconv utility: (at least on Linux)

iconv -f ISO_8859-1 -t UTF-8 iso_file > utf_file

Interesting stuff to know

  • Components of Māori place names translated

Sizes of populated places

A fair amount of work and decisions about matching NZ's ideas of what these are versus OSM's global conventions on what to call these has already been worked out as part of the LINZ data import & set up in Rob C's LINZ2OSM web app. TODO: summarize+copy that here so that everyone is working from the same page, and not trying to reinvent a wheel which already exists.