Classification of places as cities, towns or villages

Hey guys, is there any consensus regarding the classification of a settlement as a city, town or a village? I did go through the discussion on the mailing list regarding the same and I feel that we should lay down this matter to rest as soon as feasible (because its been pending since long). To summarize, the debate there was to classify human settlements depending on population or on economic importance. Googling around the web, I found out that :

  1. In India, it is the Governor who decides whether a settlement is a village or a metropolitan area (i.e. there is no definition of a city or a town).
  2. The metropolitan areas are governed by municipalities of which there are three types - the choice is again on the Governor - who chooses it by considering if the metropolitan area is an “intermediate area”, “small urban area” or a “large urban area” (Again, no definition of a city or a town).

Since villages are defined by the Governor, we can use the MDDS classification and codes for them. But now it is up to us whether to designate a metropolitan area as a town or a city because there is no definition for either.

Also, I couldn’t find any “city” code in the MDDS classification; all urban settlements were invariably “town”. Maybe this is suggesting that if a non-local mapper comes across a huge urban settlement (which is not a “village” according to MDDS code), by default he should tag it as place=town? Maybe leave it to the local mapper to decide whether it can be upgraded to a “place=city”? If so, what could be the conditions on which he would upgrade?

Hi,

I use only:
http://wiki.openstreetmap.org/wiki/Proposed_features/new_place_values

place= population
city = 100,000+
town= 10000 - 100,000
village= <10,000
hamlet= <100

this is very easy to use.

Best regards

Well, it might be easy to use, but the scheme may feel wrong for some places. For example, my “town” Brahmapur has a population of of about 3.5 lakhs. Yet I cannot give it a tag of “city” because it still has a lot of room to grow. It might be a “minor city”, certainly not a “major city”. The link to the proposal you gave is trying to solve this very problem: sub-classifying towns and cities according to various factors. Not all “cities” are important, some “towns” are very important. Population alone is not enough, presence of civic amenities, economic activities etc. need also be taken into consideration for such classification.

"Not all “cities” are important, some “towns” are very important. "

Would your judgement of importance be the same as mine? Seems to me that is unwise to try to classify according to subjective criteria when there is the reasonably objective criterion of population available. The mapping should show that each settlement is a population centre of a certain size. The current classification may not agree with the government’s but it fits in with general use.

The government designation could however be given to the municipal boundary, or a government classification tag.

Maybe you are right. It is difficult to quantify subjective criteria like economic activities, civic amenities, etc. I was thinking along the lines of Age of Empires game, where you start with a village, upgrade it to a town (by building a town centre) and then upgrade it to a city (by building a city centre, fort, etc.). The population might be large, but unless you have a fort, it is not a city. Again, the city can grow with many amenities with time. Similarly in our case, we can have different level of cities - according to the growth of economic activities, etc. I know it is highly subjective, but we can have a hierarchy of cities like the ones shown below to guide mapping on OSM:

Level 1 - Mumbai, Delhi, Kolkata, Chennai, Secunderabad, Bangalore
Level 2 - Chandigarh, Guwahati, Kochi
Level 3 - Thiruvananthapuram, Mysore, Ranchi, Pune
and so on…

All of the above are cities, but how important is that city with respect to the nation? That is my purpose of starting this discussion.

The scheme you are suggesting is going to be okay only if there is a good correlation between “population size” and “civic amenities, economic activities, etc.” What if there is no good correlation? It would be good if someone does generate that correlation :-/

For the time being, I’m going to continue using the schema you outlined (which is based on population).