[Talk-in] Classifying places - cities, towns and villages

[I Chengappa]

I agree with much of what you say. The current standard is as arbitrary as
anything else; what I don't agree with is that to partially follow a new
government system that is intended for a different purpose is the correct
way for the OSM. Without the population=tag, you will never get all the
200000+ cities, and this is useful information that should be added when
available. My argument is with the idea that the place=city tag should be
defined by a standard other than the OSM standard, because it is equally
arbitrary. On a worldwide comparison, different countries will have
different usages, and ultimately only the population can be the
determinant. Back to Mandya, which is a town according to
http://wiki.openstreetmap.org/wiki/India:Places, but is governed by a City
Municipal Council; this has obvious potential for issues and problems.
Change the definitions on this page to make CMC's cities as well and other
similar issues will remain.

The MDDS standard is not intended to deal with mapping issues; it may be
useful in this regard but it cannot be definitive. I have said before that
we should first plot and record what is physically on the ground; that is
the right thing for us. . If common usage says that a settlement is a
village, then we should record it as such, not exclude it because it is not
a panchayat.

As I see it the Mandya node could be tagged something like this;

place=city
population=131000
census2011:code=29 573 (or mdds:code)

and the city boundary can be tagged thus;
boundary=adminstrative
admin-level=7 ( or whatever is appropriate)

The consistency with the mdds is thus derived from the mdds:code tag. If
other designations are needed, such as the level of the local government,
again this can be done by using appropriate tags.

We going round in circles a bit. How about we put a summary of our
respective arguments on the talk page of
http://wiki.openstreetmap.org/wiki/India:Places and leave it to settle a
while and perhaps also get comments from people who may not get these
mails.

Thanks, indigomc


Thank you, indigomc.

On 17 May 2012 09:43, Arun Ganesh <arun.planemad at gmail.com> wrote:

> and still mean the same thing. When looking at a border area, say India
>> and Nepal, it would be useful not to have the same tag mean different
>> things on different sides of the border.
>
>
> Let me cut my reply short and put forward a geographical query: Give me a
> list of all cities in the world that has a population of over 200,000
>
> We have two options:
> 1) Use the current tag definitions and continue to record metadata in an
> inconsistent manner. Because place=city just means any place with more than
> 100,00 people, i cant get an answer to my query
>
> 2) Rethink the tag definitions so that metadata is recorded in a
> consistent manner
> If place=city actually meant that, a place with a city government, and
> there is a population= tag which had the population value. Then I could
> query the data in infinitely more ways and the dataset is more useful than
> before.
>
> Let me reiterate, the current standard, whatever it is or whoever defined
> it, is arbitrary and it is that which shoehorns data into lumps of not very
> useful clusters.
>
> I believe in doing things right when we have the option. I'm a user of the
> osm data and my interests lie in extending the utility of the dataset. If
> you have a suggestion that could could answer my query, i'd like to hear it.
>
> --
> j.mp/ArunGanesh <http://j.mp/ArunGanesh>
>
> _______________________________________________
> Talk-in mailing list
> Talk-in at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk-in
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/talk-in/attachments/20120519/a4a1aaf3/attachment-0001.html>