[Talk-ca] Formatting of Municipality Names

Jarek PiĆ³rkowski jarek at piorkowski.ca
Sat Feb 17 12:08:14 UTC 2018

On 17 February 2018 at 00:03, OSM Volunteer stevea
<steveaOSM at softworkers.com> wrote:
> On Feb 16, 2018, at 2:56 PM, Jarek PiĆ³rkowski <jarek at piorkowski.ca> wrote:
>> With "street" in a street name, it's clear to most everyone that Pine St is an abbreviation and Pine Street is the correct unabbreviated Canadian English version. It is not clear to me that "Saint Catharines" is the correct unabbreviated version of the city's name. In fact it looks incorrect to me.
> Thank you, Jarek.  However, I politely disagree that your logic of "it's clear to most everyone that Pine St is an abbreviation..." holds up for software logic (even as it might "to most everyone.")  This is well-established and the reasoning behind OSM's policy I noted previously.

Hi Steve,

Thanks for drawing the distinction.

I think that automatic software manipulation of names (whether in OSM
or elsewhere) is necessarily fraught with problems. It must be done
carefully and would be best done on a limited set of fields. There are
too many exceptions and unusual values to handle fully automatically.
The series of articles "things programmers believe about names",
"things programmers believe about addresses" come to mind.

It makes some sense in Canadian English street names, which are 1)
largely standardized and 2) by convention agglutinative - they have
prefixes and suffixes and so on. Thus it might be desirable to replace
"Pine Street" with "Pine St" or "Saint Clair Avenue West" with "St.
Clair Ave W" for rendering or presentation in short fields.

Even there, great care must be taken, or - to use a common example of
a "weird" street name - "Avenue Road" might be incorrectly reduced to
"Ave Rd". In St. Catharines there is a "South Service Road" which can
be abbreviated to "S Service Rd", but also a "South Drive" which might
make sense as "South Dr" but not as "S Drive". In Niagara-on-the-Lake
there is a "East & West Line", which I invite any algorithm to make
sense of.

(I specify "Canadian English" as I'm not familiar with Canadian French
names or with English conventions elsewhere - and that's before
getting into areas with bilingual names.)

As in the example given by others, the spelling is important to the
cities of Saint John and St. John's. IMHO this should not be
automatically modified by software. Thus we shouldn't change names
that are correct for the sake of software.


More information about the Talk-ca mailing list